IBM TotalStorage Enterprise Storage Server: Implementing the ESS in Your Environment

An IBM Redbook Publication
IBM Redbook Form Number: SG24-5420-01
ISBN: 0738424390
ISBN: 9780738424392
Publication Date: 28-Mar-2002
Find Similar Download

Related People

Gustavo Castets - Author [+3] [-3]
Donald (Chuck) Laing - Author
Jukka Myyryläinen - Author
Diane Williams - Author

Abstract

Since the initial availability of the IBM TotalStorage Enterprise Storage Server (ESS) there have been many modifications and additional functionality. The original ESS Implementation Guide was written in November 1999. The 2002 version is an updated guide that includes all of the changes since then.

This IBM Redbooks publication is a guide for the installation, implementation, and administration activities of the ESS in both the S/390 and open systems environments. It will help you plan and accomplish the installation, tailoring, and configuration of the ESS in your environment. It explains how you can use the functions available for the ESS to efficiently manage your disk storage data as well as the ESS, once it is operative.

We cover the latest announcements on the ESS: disk capacity intermix; 72.8 GB capacity disk drive; flexible configurations; Control Unit Initiated Reconfiguration (CUIR) support; large volume support (LVS); read from secondary; ESS Master Console; Subsystem Device Driver (SDD) and the Command Line Interface (CLI) support for additional operating systems; INRANGE Channel Extender support; TPF support for PPRC and FlashCopy.

We also provide information on the new models F10 and F20; FICON native host attachment; new Fibre Channel/FICON host adapters (short wave and long wave); Linux support for Intel-based servers and zSeries servers; iSeries and AS/400 support for copy services; and new cache options.

Language

English

Table of Content

Part 1. Characteristics, planning, and installation
Chapter 1. Introduction and positioning
Chapter 2. Configuration planning
Chapter 3. Physical installation planning
Chapter 4. Availability
Chapter 5. IBM TotalStorage Enterprise Storage Server Specialist
Chapter 6. IBM TotalStorage Enterprise Storage Server Expert
Part 2. Implementation in the zSeries environment
Chapter 7. zSeries systems support
Chapter 8. CKD storage configuration
Chapter 9. S/390 and zSeries host setup tasks
Chapter 10. Data migration in the zSeries environment
Chapter 11. Managing and monitoring the ESS
Chapter 12. ESS Copy Services for S/390
Part 3. Implementation in the open systems environment
Chapter 13. Open systems support
Chapter 14. Open systems host setup tasks
Chapter 15. ESS configuration for open systems fixed block storage
Chapter 16. Data migration in the open systems environment
Chapter 17. Managing and monitoring the ESS
Chapter 18. ESS Copy Services for open systems
Appendix A. ESS configuration planning process
ibm.com/redbooks
Front cover
IBM TotalStorage
Enterprise Storage Server
Implementing the ESS in Your Environment
Gustavo Castets
Donald (Chuck) Laing
Jukka Myyryläinen
Diane Williams
Learn how to install, tailor, and
configure the ESS
Plan for migrating data and
changing configurations
Know the tools for monitoring
and managing the ESS


International Technical Support Organization
IBM TotalStorage Enterprise Storage Server:
Implementing the ESS in Your Environment
March 2002
SG24-5420-01

© Copyright International Business Machines Corporation 2001, 2002. All rights reserved.
Note to U.S Government Users - Documentation related to restricted rights - Use, duplication or disclosure is subject to restrictions set
forth in GSA ADP Schedule Contract with IBM Corp.
Second Edition (March 2002)
This edition applies to the IBM 2105 Enterprise Storage Server model F. For other information, please see the
Publications section of the IBM Announcement letter for the IBM Enterprise Storage Server. You can also visit
our Web site at:
http://www.storage.ibm.com
Comments may be addressed to:
IBM Corporation, International Technical Support Organization
Dept. QXXE Building 80-E2
650 Harry Road
San Jose, California 95120-6099
When you send information to IBM, you grant IBM a non-exclusive right to use or distribute the information in
any way it believes appropriate without incurring any obligation to you.
Take Note! Before using this information and the product it supports, be sure to read the general
information in “Special notices” on page 299.

© Copyright IBM Corp. 2002
iii
Summary of changes
This section describes the technical changes made in this edition of the book and in previous
editions. This edition may also include minor corrections and editorial changes that are not
identified.
Summary of Changes
for SG24-5420-01
for IBM TotalStorage Enterprise Storage Server
as created or updated on March 28, 2002.
March 2002, Second Edition
This revision reflects the addition, deletion, or modification of new and changed information
described below.
New information
This 2002 version of the redbook is an update of the November 1999 original document, and
includes the latest announcements on the IBM TotalStorage Enterprise Storage Server:
IBM TotalStorage Enterprise Storage Server models F10 and F20
New Fibre Channel/FICON host adapters (short wave and long wave)
Cache options: 8, 16, 24 and 32GB
FICON native host attachment
ESS Master Console
Flexible configurations
Disk drive capacity intermix
Support for 72.8GB disk drives, scaling the ESS total capacity up to 22TB
Control Unit Initiated Reconfiguration (CUIR) support
32K cylinder Large volume support (LVS)
Linux support for zSeries servers and Intel-based servers
PPRC and FlashCopy support for iSeries and AS/400
ESS Copy Services Command Line Interface (CLI) supporting additional platforms
Subsystem Device Driver (SDD) supporting additional platforms
Read from secondary
INRANGE Channel Extender support
TPF support for PPRC and FlashCopy

iv
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment

© Copyright IBM Corp. 2002
v
Contents
Summary of changes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
March 2002, Second Edition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .v
Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xvii
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix
The team that wrote this redbook. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix
Notice. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi
IBM trademarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xxii
Comments welcome. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xxii
Part 1. Characteristics, planning, and installation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Chapter 1. Introduction and positioning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1 Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2 Functions and features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Explanation of terms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.4 Platform support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.5 Getting ESS-related information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.5.1 ESS Web site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.5.2 List of supported servers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.5.3 ESS documentation Web site. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.5.4 Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Chapter 2. Configuration planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2 Capacity planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2.1 Capacity and configuration rules. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2.2 Storage capacity options. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.2.3 Capacity planning by storage format. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.2.4 ESS Copy Services. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.2.5 Flexible configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.3 Logical configuration planning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.3.1 Logical subsystem mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.4 Planning for performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Chapter 3. Physical installation planning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.1 General considerations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.2 Host attachments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.2.1 Daisy chaining SCSI adapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.2.2 Unique target IDs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.2.3 Fibre Channel host attachment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.3 S/390 ESCON/FICON cables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.4 Electrical power. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.4.1 Dual line cords . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.4.2 Single phase power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

vi
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
3.4.3 Three phase power. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.4.4 Other considerations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.5 Physical location and floor loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.5.1 Transit. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.5.2 Physical placement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.5.3 Floor loading. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.6 Cooling and airflow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.6.1 Operating temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.6.2 Airflow. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.7 Call Home and Remote Support. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.7.1 IBM TotalStorage Enterprise Storage Server Master Console. . . . . . . . . . . . . . . 39
3.7.2 ESS Master Console differences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.8 Web client and Ethernet connection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.8.1 ESS Specialist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.8.2 Minimal connection to the Web client . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.8.3 Connection to Local TCP/IP Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.8.4 ESS Copy Services behind the firewall. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Chapter 4. Availability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.1 ESS internal recovery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4.1.1 Host adapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4.1.2 Clusters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.1.3 Disk drives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4.1.4 Power and cooling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
4.1.5 Ethernet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
4.2 Maintenance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
4.2.1 Maintenance strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
4.2.2 Upgrades . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.2.3 Non-disruptive service actions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.2.4 Disruptive service actions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
4.3 Error notification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
4.3.1 Call Home feature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.3.2 Simple network management protocol (SNMP) . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.3.3 E-mail . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.3.4 Pager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.3.5 S/390 notification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.4 Configuration for availability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4.4.1 Maintaining connectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4.4.2 Access to data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
4.4.3 Power and cooling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Chapter 5. IBM TotalStorage Enterprise Storage Server Specialist. . . . . . . . . . . . . . . 79
5.1 ESS Specialist. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
5.2 ESS Specialist prerequisites. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
5.3 ESS Specialist components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
5.4 Connecting to your ESS Specialist. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
5.4.1 ESS Master Console. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
5.4.2 ESS Specialist panels. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
5.4.3 Welcome panel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
5.4.4 Help. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
5.4.5 Security. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
Chapter 6. IBM StorWatch Enterprise Storage Server Expert. . . . . . . . . . . . . . . . . . . . 87
6.1 Using ESS Expert. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

Contents
vii
6.2 Accessing ESS Expert . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
6.3 Navigating through ESS Expert . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
6.3.1 Additional documentation on ESS Expert. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Part 2. Implementation in the zSeries environment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Chapter 7. zSeries systems support. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
7.1 Basic support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
7.1.1 OS/390 and z/OS support. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
7.1.2 VM/ESA support. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
7.1.3 VSE/ESA support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
7.1.4 TPF support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
7.1.5 Linux. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
7.2 FlashCopy support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
7.3 FICON support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
7.4 Large Volume Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
7.5 CUIR support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
Chapter 8. CKD storage configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
8.1 Basic concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
8.1.1 ESS Specialist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
8.1.2 Logical Control Units. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
8.1.3 Disk groups and ranks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
8.1.4 RAID arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
8.1.5 CKD logical volumes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
8.1.6 Parallel Access Volumes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
8.2 Designing the LCU address layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
8.3 Configuration process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
8.3.1 Selecting the configuration method. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
8.3.2 Configuration worksheets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
8.3.3 Communication Resources worksheet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
8.4 ESS custom configuration for CKD storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
8.4.1 Tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
8.4.2 ESS Specialist panels for CKD storage allocation . . . . . . . . . . . . . . . . . . . . . . . 118
8.4.3 ESS Specialist messages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
8.4.4 Storage Allocation panel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
8.4.5 S/390 Storage panel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
8.4.6 Configuring logical control units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
8.4.7 Configure disk groups for an LCU. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
8.4.8 Add volumes to an LCU . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
8.4.9 Modify PAV assignments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
8.4.10 Configure FICON ports. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
8.4.11 A step-by-step example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
Chapter 9. S/390 and zSeries host setup tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
9.1 Preparation and considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
9.2 OS/390 software configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
9.2.1 IOCP/HCD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
9.2.2 FICON host connectivity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
9.2.3 IOCP/HCD examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
9.2.4 WLM dynamic alias management. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
9.2.5 Custom volumes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
9.2.6 Large volumes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
9.2.7 MIH interval. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164

viii
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
9.2.8 DFSMS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
9.2.9 Volume initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
9.2.10 ESS Copy Services. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
9.3 VM/ESA software configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
9.3.1 Parallel access volumes (PAV). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
9.3.2 ESS Copy Services. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
9.4 VSE/ESA software configuration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
9.4.1 ESS Copy Services. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
9.5 TPF software configuration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
9.5.1 PPRC and FlashCopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
9.6 Linux software configuration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
Chapter 10. Data migration in the zSeries environment . . . . . . . . . . . . . . . . . . . . . . . 171
10.1 Migration planning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
10.2 Data migration for MVS environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
10.2.1 Migration methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
10.3 Data migration for VM environments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
10.4 Data migration for VSE environments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
10.5 Data migration across ESSs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
10.6 IBM migration services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
Chapter 11. Managing and monitoring the ESS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
11.1 General considerations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
11.2 Managing tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
11.2.1 IBM TotalStorage Enterprise Storage Server Specialist. . . . . . . . . . . . . . . . . . 178
11.2.2 Control Unit Initiated Reconfiguration (CUIR). . . . . . . . . . . . . . . . . . . . . . . . . . 178
11.3 Monitoring tools. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
11.3.1 RMF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
11.3.2 CRR and DFSMS Optimizer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
11.3.3 SMF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
11.3.4 SIM and EREP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
11.4 Logical subsystem resource usage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
11.4.1 IDCAMS LISTDATA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
11.4.2 IDCAMS SETCACHE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
11.5 MVS system commands. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
11.5.1 DEVSERV command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
11.5.2 DISPLAY MATRIX command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
11.6 Messages and codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
11.6.1 IEA434I. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
11.6.2 IEA435I. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
11.6.3 IEA480E SIM message. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
11.6.4 IEE169I. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
11.6.5 IOS090I. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
11.6.6 CUIR messages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
11.7 IBM StorWatch Enterprise Storage Server Expert . . . . . . . . . . . . . . . . . . . . . . . . . . 196
Chapter 12. ESS Copy Services for S/390. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
12.1 S/390 Copy Services: introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
12.1.1 Peer-to-Peer Remote Copy (PPRC). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
12.1.2 FlashCopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
12.1.3 Extended Remote Copy (XRC). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
12.1.4 Concurrent Copy (CC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
12.2 ESS Copy Services: considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
12.3 PPRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200

Contents
ix
12.3.1 Rules for configuring PPRC links . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
12.3.2 How to invoke PPRC for S/390. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
12.4 FlashCopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
Part 3. Implementation in the open systems environment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
Chapter 13. Open systems support. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
13.1 General considerations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
13.2 Operating system requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
13.3 I/O adapters: considerations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
13.4 Subsystem Device Driver (SDD). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
13.5 Command Line Interface (CLI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
13.6 Additional considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
Chapter 14. Open systems host setup tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
14.1 General. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
14.2 AIX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
14.2.1 I/O adapters in the host server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
14.2.2 WWPN of the Fibre Channel I/O adapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
14.2.3 Host recognition of newly assigned disks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
14.2.4 ESS LUNs assigned as host hdisks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
14.2.5 FlashCopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
14.3 Compaq (DEC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
14.3.1 I/O adapters in the host server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
14.3.2 WWPN of the Fibre Channel I/O adapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
14.3.3 Host recognition of newly assigned disks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
14.3.4 ESS LUNs assigned as host disks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
14.4 Hewlett Packard 9000. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
14.4.1 I/O adapters in the host server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
14.4.2 WWPN of the Fibre Channel I/O adapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
14.4.3 Host recognition of the newly assigned disks. . . . . . . . . . . . . . . . . . . . . . . . . . 220
14.4.4 ESS LUNs assigned as host disks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
14.5 AS/400 and iSeries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
14.5.1 I/O adapters in the host server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
14.5.2 WWPN of the Fibre Channel I/O adapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
14.5.3 Host recognition of the newly assigned disks. . . . . . . . . . . . . . . . . . . . . . . . . . 223
14.5.4 ESS LUNs assigned as host disks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
14.6 Intel based servers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
14.6.1 I/O adapters in the host server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
14.6.2 WWPN of the Fibre Channel I/O adapter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
14.6.3 Host recognition of newly assigned disks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
14.6.4 ESS LUNs assigned as host disks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
14.7 Linux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
14.7.1 Kernel configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
14.7.2 Loading the I/O adapter card as a module . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
14.7.3 I/O adapters in the host server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
14.7.4 WWPN of the Fibre Channel I/O adapter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
14.7.5 Host recognition of newly assigned disks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
14.7.6 ESS LUNs assigned as host disks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
14.7.7 Partitioning the disks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
14.7.8 Creating and using a filesystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
14.8 Novell NetWare. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
14.8.1 I/O adapters in the host server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
14.8.2 WWPN of the Fibre Channel I/O adapter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230

x
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
14.8.3 Host recognition of newly assigned disks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
14.8.4 ESS LUNs assigned as host disks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
14.9 Sun . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
14.9.1 I/O adapters in the host server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
14.9.2 WWPN of the Fibre Channel I/O adapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
14.9.3 Host recognition of newly assigned disks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
14.9.4 ESS LUNs assigned as host disks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
14.10 NUMA-Q . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
14.10.1 I/O adapters in the host server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
14.10.2 WWPN of Fiber Channel I/O adapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
14.10.3 Host recognition of newly assigned disks. . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
14.10.4 ESS LUNs assigned as host disks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
Chapter 15. ESS configuration for open systems fixed block storage . . . . . . . . . . . 235
15.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
15.1.1 Logical subsystem (LSS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
15.1.2 FB logical volumes and SCSI LUNs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
15.1.3 FB logical volumes and Fibre Channel LUNs. . . . . . . . . . . . . . . . . . . . . . . . . . 239
15.2 Fixed block storage configuration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
15.2.1 Task sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
15.2.2 Connecting to ESS Specialist. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
15.2.3 Configure Disk Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
15.2.4 Modify host systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
15.2.5 Configure Host Adapter Ports. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
15.2.6 Adding fixed block volumes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
15.2.7 Modify Volume Assignments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
15.3 Sample procedures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
15.3.1 Formatting RAID-5 disk groups. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
15.3.2 Creating a SCSI logical host. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
15.3.3 Creating a Fibre Channel logical host. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260
15.3.4 Assigning SCSI Host Adapters to a logical host. . . . . . . . . . . . . . . . . . . . . . . . 260
15.3.5 Assigning Switched Fibre Channel host adapters to a logical host. . . . . . . . . . 261
15.3.6 Assigning Direct Fibre Channel host adapters to a logical host . . . . . . . . . . . . 261
15.3.7 Carving LUNs from RAID5 Disk Groups. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
15.3.8 Assigning LUNs to Fibre Channel host adapters . . . . . . . . . . . . . . . . . . . . . . . 263
15.3.9 Assigning LUNs to SCSI Host Adapters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
15.3.10 Verifying a LUN and SCSI IDs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
Chapter 16. Data migration in the open systems environment. . . . . . . . . . . . . . . . . . 265
16.1 Migrating in the open system environment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266
16.1.1 Method selection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266
16.1.2 Replacing existing storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266
16.2 Data migration for iSeries systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
16.3 Data migration for UNIX systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
16.3.1 Migration methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
16.3.2 The AIX logical volume manager (LVM). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
16.4 Data migration for Windows NT and 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
16.4.1 Migration methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
16.5 Data migration across ESSs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
16.6 IBM migration services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
Chapter 17. Managing and monitoring the ESS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
17.1 IBM StorWatch Enterprise Storage Server Expert . . . . . . . . . . . . . . . . . . . . . . . . . . 276
17.2 General considerations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276

Contents
xi
17.3 Managing and monitoring tools. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276
17.3.1 AIX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276
17.3.2 HP-UX and Sun Solaris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
17.3.3 OS/400 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278
17.3.4 Windows Disk Administrator. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
17.3.5 Linux. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
17.3.6 NUMA-Q. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
Chapter 18. ESS Copy Services for open systems . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
18.1 ESS Copy Services: considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
18.2 Preliminary setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
18.3 PPRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
18.3.1 Rules for configuring PPRC links . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
18.3.2 Planning for PPRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
18.3.3 PPRC Read from Secondary option. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
18.3.4 How to invoke PPRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
18.4 FlashCopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
18.4.1 Invoking FlashCopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
18.4.2 Invoking FlashCopy scenario for AIX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
Appendix A. ESS configuration planning process . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
Considerations for planning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294
Where should you start? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294
Gathering information to submit the ESS Configuration worksheets . . . . . . . . . . . . . . 294
Completing the Communication Resources worksheets. . . . . . . . . . . . . . . . . . . . . . . . 295
Where you can find the reference material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
Submitting the ESS Configuration worksheets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
Configuration Worksheet Assistance details. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
Related publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
IBM Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
Referenced Web sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
How to get IBM Redbooks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
IBM Redbooks collections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
Special notices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301

xii
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment

© Copyright IBM Corp. 2002
xiii
Figures
1-1 ESS: Starting Web site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1-2 The More Information area at the ESS Web page. . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1-3 ESS documentation Web site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2-1 The expanded planning team . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2-2 S/390 logical paths using ESCON. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2-3 SCSI logical configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2-4 Fibre logical configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2-5 INRANGE 9801 storage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2-6 LSS logical to physical mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3-1 Single phase ESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3-2 Caution: Floor loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3-3 ESS Master Console: (FC 2717) — connections. . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3-4 ESS Master Console: Main screen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3-5 ESS Master Console: Selecting remote server services. . . . . . . . . . . . . . . . . . . . . . 41
3-6 ESS Master Console: Enabling remote services. . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3-7 Call home configuration of the old ESSNet (FC 2715) with modem expander . . . . . 42
3-8 Ethernet private LAN configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3-9 Network view 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3-10 Network view 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3-11 Network view 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3-12 ESSNet A. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3-13 ESSNet A and B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4-1 ESS cluster configuration for normal operation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4-2 ESS cluster failover process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
4-3 ESS cluster failback process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
4-4 Cooling fan groups in the ESS base frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
4-5 Service terminal functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4-6 ESS communications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4-7 Function of subsystem device driver (SDD). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
4-8 Multiple ESCON switch configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
4-9 Physical configuration of ESCON paths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
5-1 Entering the URL using the cluster name of your ESS . . . . . . . . . . . . . . . . . . . . . . . 83
5-2 Entering the URL using the TCP/IP address. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
5-3 ESS Specialist Welcome screen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
5-4 ESS Specialist function selection panel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
5-5 ESS Specialist help icon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
5-6 Modify users panel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
6-1 ESS Expert: Welcome panel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
6-2 ESS Expert: Managing the ESS Expert. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
6-3 ESS Expert: Managing ESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
8-1 Relationship between S/390 LCUs and SSA loops. . . . . . . . . . . . . . . . . . . . . . . . . 103
8-2 Disk group to LSS association. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
8-3 Disk groups in the Storage Allocation panel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
8-4 ESS Specialist: S/390 panels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
8-5 ESS Specialist: Estimated time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
8-6 ESS Specialist: Error 1109 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
8-7 ESS Specialist: Storage Allocation - Graphical View panel. . . . . . . . . . . . . . . . . . . 122
8-8 ESS Specialist: S/390 Storage panel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

xiv
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
8-9 ESS Specialist: S/390 Storage panel with devices . . . . . . . . . . . . . . . . . . . . . . . . . 123
8-10 ESS Specialist: Configure LCU panel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
8-11 ESS Specialist: Procedure to configure an LCU . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
8-12 ESS Specialist: Warning 1554. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
8-13 ESS Specialist: Warning 1550. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
8-14 ESS Specialist: Configure Disk Groups panel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
8-15 ESS Specialist: Procedure to configure disk groups . . . . . . . . . . . . . . . . . . . . . . . . 131
8-16 ESS Specialist: Warning 1982. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
8-17 ESS Specialist: Add Volumes panel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
8-18 ESS Specialist: Error 1901 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
8-19 ESS Specialist: Procedure to add volumes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
8-20 ESS Specialist: Configuring PAV aliases. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
8-21 ESS Specialist: Configure Host Adapter Ports panel. . . . . . . . . . . . . . . . . . . . . . . . 139
8-22 Sample configuration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
9-1 HCD panel control unit 2105 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
9-2 ESS Specialist LCU and device details. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
9-3 HCD panel CUADD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
9-4 HCD panel address range. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
9-5 HCD Add Device panel for 16 base devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
9-6 HCD Define Device Parameters panel with WLMPAV parameter. . . . . . . . . . . . . . 152
9-7 HCD panel define alias device. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
9-8 Activate WLM dynamic alias management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
9-9 DEVSERV QPAVS command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
11-1 MVS messages for quiesce request . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
11-2 MVS messages for resume request. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
11-3 Vary path online rejected. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
11-4 DEVSERV PATHS output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
11-5 DEVSERV QUERY DASD output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
11-6 DEVSERV QUERY DASD UCB output. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
11-7 DEVSERV QPAVS VOLUME output. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
11-8 DEVSERV QPAVS output. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
11-9 Unlisted devices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
11-10 Display Channel path status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
11-11 D M=DEV - Base device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
11-12 D M=DEV - Alias device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
11-13 Service Information Message (SIM). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
13-1 ESS - open systems support. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
14-1 Example of a modified /etc/fstab file . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
14-2 Example of how to label disks with volume sizes . . . . . . . . . . . . . . . . . . . . . . . . . . 219
14-3 Example listing of LUN assignment for Compaq servers. . . . . . . . . . . . . . . . . . . . . 219
14-4 SCSI adapter identification - HP-UX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
14-5 Fibre Channel adapter identification - HP-UX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
14-6 Host view of newly assigned disks - HP-UX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
14-7 LUN serial number to disk number - HP-UX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
14-8 Identifying the I/O adapters - iSeries and AS/400 . . . . . . . . . . . . . . . . . . . . . . . . . . 222
14-9 Resource detailed information for I/O adapters - iSeries and AS/400. . . . . . . . . . . 222
14-10 Identifying the WWPN - AS/400 and iSeries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
14-11 Listing of the assigned LUNs - iSeries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
14-12 Listing of the assigned LUNs - Intel based servers . . . . . . . . . . . . . . . . . . . . . . . . . 226
14-13 Adapter device list . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
14-14 list devices output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
14-15 LUNs listing, sample output -Sun. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
14-16 Output for rslist2105s - LUN serial numbers - Sun . . . . . . . . . . . . . . . . . . . . . . . . . 233

Figures
xv
14-17 Identifying the WWPN - NUMA-Q servers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
15-1 ESS Specialist: Storage allocation graphical view. . . . . . . . . . . . . . . . . . . . . . . . . . 236
15-2 Relationship between clusters, LSSs, disk groups and LUNs. . . . . . . . . . . . . . . . . 237
15-3 Relationship between LSSs DAs and SSA loops . . . . . . . . . . . . . . . . . . . . . . . . . . 238
15-4 Recommended configuration sequence for FB storage. . . . . . . . . . . . . . . . . . . . . . 240
15-5 Configuring fixed block disk groups. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
15-6 Configuring fixed block storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
15-7 ESS Specialist: Message 1402 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
15-8 ESS Specialist: Warning 1802. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
15-9 ESS Specialist: Modify Host System panel for SCSI. . . . . . . . . . . . . . . . . . . . . . . . 244
15-10 ESS Specialist: Modify host systems panel for Fibre Channel . . . . . . . . . . . . . . . . 244
15-11 ESS Specialist: Fibre-Channel Ports. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
15-12 ESS Specialist: Host Type drop down box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246
15-13 ESS Specialist: Configure SCSI host adapter ports . . . . . . . . . . . . . . . . . . . . . . . . 247
15-14 ESS Specialist: SCSI bus configuration window . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
15-15 ESS Specialist: Warning 1610. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
15-16 ESS Specialist: SCSI ID map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
15-17 ESS Specialist: Configure Fibre port panel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
15-18 ESS Specialist: Add Volumes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
15-19 ESS Specialist: Non-RAID disk group. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
15-20 ESS Specialist: Add Volumes (2 of 2) panel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254
15-21 ESS Specialist: modify volume assignments panel. . . . . . . . . . . . . . . . . . . . . . . . . 256
15-22 ESS Specialist: Dual row SCSI port assignments. . . . . . . . . . . . . . . . . . . . . . . . . . 257
17-1 AIX smitty panel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
17-2 Veritas Volume Manager GUI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
17-3 DST main menu. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278
18-1 ESS Copy Services: Read from Secondary allowed. . . . . . . . . . . . . . . . . . . . . . . . 287
18-2 ESS Copy Services: First wizard window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290
18-3 Establishing a pair with No Copy background. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290
18-4 Wizard window with example task name and description . . . . . . . . . . . . . . . . . . . . 291

xvi
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment

© Copyright IBM Corp. 2002
xvii
Tables
2-1 ESS RAID-5 disk storage capacity by model (rounded) . . . . . . . . . . . . . . . . . . . . . . 19
2-2 JBOD array capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2-3 RAID-5 array capacity - FB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3-1 Dimensions and weights for 2105 racks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3-2 Differences between ESS Master Console and ESSNet console . . . . . . . . . . . . . . . 42
5-1 Web browsers supported by ESS Web interfaces. . . . . . . . . . . . . . . . . . . . . . . . . . . 81
8-1 Array capacities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
8-2 CKD logical device capacities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
8-3 Number of volumes in a RAID array . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
8-4 PAV alias ratios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
8-5 Device address layout example 2 - 256 devices . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
8-6 Device address layout example 3 - non-contiguous address ranges . . . . . . . . . . . 113
8-7 Device address layout example 4 - 128 devices . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
8-8 Device address layout example 5 - more aliases . . . . . . . . . . . . . . . . . . . . . . . . . . 114
8-9 Number of volumes supported by the standard configuration process . . . . . . . . . . 115
8-10 S/390 storage requirements worksheet. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
9-1 FICON and ESCON comparison. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
9-2 Dynamic alias management algorithms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
9-3 Effects of WLMPAV settings on base and alias devices in HCD. . . . . . . . . . . . . . . 157
9-4 Linux implementations in zSeries servers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
13-1 Open systems support summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
14-1 Example of disk information for installed devices for a Compaq AlphaServer. . . . . 218
16-1 Example migration procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267

xviii
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment

© Copyright IBM Corp. 2002
xix
Preface
Since the initial availability of the IBM TotalStorage Enterprise Storage Server (ESS) there
have been many modifications and additional functionality. The original ESS Implementation
Guide was written in November 1999. The 2002 version is an updated guide that includes all
of the changes since then.
This IBM Redbook is a guide for the installation, implementation, and administration activities
of the ESS in both the S/390 and open systems environments. It will help you plan and
accomplish the installation, tailoring, and configuration of the ESS in your environment. It
explains how you can use the functions available for the ESS to efficiently manage your disk
storage data as well as the ESS, once it is operative.
We cover the latest announcements on the ESS: disk capacity intermix; 72.8 GB capacity
disk drive; flexible configurations; Control Unit Initiated Reconfiguration (CUIR) support; large
volume support (LVS); read from secondary; ESS Master Console; Subsystem Device Driver
(SDD) and the Command Line Interface (CLI) support for additional operating systems;
INRANGE Channel Extender support; TPF support for PPRC and FlashCopy.
We also provide information on the new models F10 and F20; FICON native host attachment;
new Fibre Channel/FICON host adapters (short wave and long wave); Linux support for
Intel-based servers and zSeries servers; iSeries and AS/400 support for copy services; and
new cache options.
The team that wrote this redbook
This redbook was produced by a team of specialists from around the world working at the
International Technical Support Organization, San Jose Center.
Gustavo Castets is a Project Leader with the IBM International Technical Support
Organization, San Jose Center. He writes extensively and teaches IBM classes worldwide on
areas of Disk Storage Systems. Before joining the ITSO, Gustavo worked in System Sales as
Field Technical Support Specialist, and has worked for more than 22 years in many IT areas,
in Buenos Aires, for IBM Argentina.
Donald (Chuck) Laing is a Disk Administrator in the south delivery center in Texas. He has
13 years of experience in the UNIX field, and has worked at IBM for 3 years. Before joining
IBM, Chuck was a Hardware CE on UNIX systems and taught basic UNIX at Midland College
in Midland Texas. He holds a degree in Computer Science from Midland College. His areas of
expertise include Open Systems UNIX System/Disk Administration.
Jukka Myyryläinen is an Advisory IT Specialist with IBM Global Services, Finland. He has
15 years of experience in storage product implementations. He has contributed to several
storage related redbooks in the past.
Diane Williams is a Consulting IT Specialist with the San Jose Advanced Technical Support
group. She has 23 years of experience in several areas of IT. Her areas of expertise include
disk performance.

xx
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Gustavo Castets, Jukka Myyryläinen, Donald (Chuck) Laing, Diane Williams
The authors of previous editions of this book are:
Mark Blunden, IBM Australia
Ivan Avsic, IBM Slovenia
Ian Black, IBM Australia
Peter Crowhurts, IBM Australia
Carlos Sadao Miyabara, Banco Itau, Brazil
Thanks to the following people for their diligence in reviewing this manual:
Don Maggert, IBM, Tucson, ESP Management Office
James Cammarata, IBM, Chicago, DFSMS Device Support
Jennifer Eaton, IBM, San Jose, User Technology - Information Development
Jim Garadis, IBM, San Jose, IRT - SSG Brand Support
Leslie Barton, IBM, San Jose, DFSMS Device Support
Nicholas Clayton, IBM, UK, EMEA ATS for Storage
Omar F Escola, IBM, Argentina, Hardware Support Specialist
Pat Blaney, IBM, San Jose, Storage Systems ATS
Paul Coles, IBM, UK, ITS Storage Specialist
Paulus Usong, IBM, San Jose, Storage Systems ATS
Phil Lee, IBM, Canada, ESS Configuration HelpLine
Rainer Wolafka, BM, San Jose, Open Systems Validation Lab
Scott Cardinell, IBM, Tucson, Microcode Tools
Steven Winch, IBM, Rochester, iSeries Hardware Test
Thomas Fiege, IBM, San Jose, SSD RAS Engineering

Preface
xxi
Thanks to the following people for their invaluable contributions to this project:
Amine Hajji, IBM, San Jose, Advanced Product Development
Amy L Therrien, IBM, San Jose, ESS Web Development
Brian Kraemer, IBM, San Jose, RAS Microcode Develpment
Cindy Sauer, IBM, San Jose, SSG RFA
Edsel Carson, IBM, San Jose, Staff Information Development
Ellen Vuong, IBM, San Jose, ESS Web Development
Gail Spear, IBM, Tucson, ESS Strategy and Architecture - Microcode
Henry Caudillo, IBM, San Jose, User Technology - Information Development
Hilda Wu, IBM, San Jose, Expert Development
Jack Flynn, IBM, San Jose, Device Support Architecture
Jane Madriaga-Miller, IBM, San Jose, ESS Web Development
Jed Dyreng, IBM, San Jose, Service Panel Development
Jeff Ferrier, IBM, San Jose, SSD RAS Engineering
Jeffrey Steffan, IBM, San Jose, DASD RAS
Joe Writz, IBM, Rochester, Storage IO Subsystem Develpment
John Bernauer, IBM, San Jose, RAS Programming
Kurt Lovrien, IBM, Tucson, Microcode Development
Lynsey V Bishop, IBM, San Jose, User Technology - Information Development
Maria Magana, IBM, San Jose, ESS Web Development
Marilyn Pullen, IBM, San Jose, User Technology - Information Development
Michael Washuk, IBM, Tucson, Shark Project Management
Mike Janini, IBM, San Jose, Open Systems Validation Lab
Richard Heffel, IBM, San Jose, Open Systems Lab
Robert Moon, IBM, San Jose, Open Systems Validation Lab
Sheryl Reed, IBM, San Jose, DFSMS Device Support
Stefan Jaquet, IBM, San Jose, DASD Web Development
Steven Winters, IBM, San Jose, DFSMS Optimizer Development
Walter Weaver, IBM, San Jose, Information Development
William H Turner, IBM, San Jose, User Technology - Information Development
Notice
This publication is intended to help IT professionals who are implementing the IBM
TotalStorage Enterprise Storage Server in S/390 and open environments. The information in
this publication is not intended as the specification of any programming interfaces that are
provided by IBM or third parties. See the PUBLICATIONS section of the IBM Programming
Announcement for the IBM TotalStorage Enterprise Storage Server, Models F10 and F20, for
more information about what publications are considered to be product documentation.

xxii
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
IBM trademarks
The following terms are trademarks of the International Business Machines Corporation in the
United States and/or other countries:
Comments welcome
Your comments are important to us!
We want our Redbooks to be as helpful as possible. Send us your comments about this or
other Redbooks in one of the following ways:
Use the online Contact us review redbook form found at:
ibm.com/redbooks
Send your comments in an Internet note to:
redbook@us.ibm.com
Mail your comments to the address on page ii.
e (logo)®
Redbooks Logo™
AIX®
AS/400®
CICS®
CUA®
DB2®
DFS™
DFSMS/MVS®
DFSMS/VM®
DFSMSdfp™
DFSMSdss™
DFSMShsm™
DFSORT™
DYNIX®
DYNIX/ptx®
Enterprise Storage Server™
Enterprise Systems Connection®
Architecture®
ESCON®
FICON™
FlashCopy™
IBM®
IMS™
IMS/ESA®
iSeries™
Magstar®
MVS™
Notes®NUMA-Q®
NUMACenter™
OS/390®
OS/400®
Perform™
PR/SM™
Predictive Failure Analysis®
PR/SM™
pSeries™
RACF®
RAMAC®
Redbooks™
RETAIN®
RMF™
RS/6000®
S/390®
Seascape®
Sequent®
SP™
StorWatch™
System/390®
Tivoli®
TotalStorage™
VM/ESA®
VSE/ESA™
xSeries™
z/OS™
z/VM™
zSeries™

© Copyright IBM Corp. 2002
1
Part 1
Characteristics,
planning, and
installation
In this part of the book we present the characteristics of the IBM TotalStorage Enterprise
Storage Server. We then discuss considerations for the planning and installation activities you
will be conducting for your ESS implementation. Finally, we present two ESS specific tools,
the ESS Specialist and the ESS Expert. This part of the book is intended for readers from
both open systems and S/390 environments.
Part 1

2
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment

© Copyright IBM Corp. 2002
3
Chapter 1.
Introduction and positioning
In this chapter we provide an overview of the IBM TotalStorage Enterprise Storage Server
and the functions and features it provides. We also offer some positioning information to give
you an understanding of when the ESS could be used in your environment. In addition, we
present an overview of the terminology and definitions that will be used throughout this book.
This chapter covers the following topics:
Overview of the IBM TotalStorage Enterprise Storage Server components
IBM TotalStorage Enterprise Storage Server positioning
IBM TotalStorage Enterprise Storage Server functions and features
An explanation of general terms used throughout this book
1

4
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
1.1 Overview
The IBM TotalStorage Enterprise Storage Server (ESS) is a member of the Seascape family.
It consists of a storage server and attached disk storage devices. The storage server provides
integrated caching and RAID support for the attached disk devices. The disk devices are
attached via a Serial Storage Architecture (SSA) interface. The ESS can be configured in a
variety of ways to provide scalability in capacity and performance.
Redundancy within the ESS provides continuous availability. It is packaged in one or more
enclosures, each with dual line cords and redundant power. The redundant power system
allows the ESS to continue normal operation when one of the line cords is deactivated, or a
power supply fails.
The ESS provides the image of a set of logical disk devices to attached servers. The logical
devices are configured to emulate disk device types that are compatible with the attached
servers. The logical devices access a logical volume that is implemented using portions of
multiple disk drives.
The following host I/O interface attachments are supported:
SCSI-3 Parallel Interface
Fibre Channel (FCP)
ESCON
FICON
On SCSI interfaces, the ESS emulates a variety of fixed block devices with either 512 or 520
byte blocks. Host systems with SCSI-2 or SCSI-3 interfaces can attach to the ESS. The ESS
provides multiple SCSI I/O interfaces (buses), each with multiple SCSI targets, and each with
multiple disk logical units. The storage provided by the ESS for SCSI interfaces can be
configured so that it is shared among multiple SCSI interfaces if desired.
On FCP the ESS emulates a variety of fixed block devices with either 512 or 520 byte blocks.
It allows 1 target per port, with 4096 LUNs per target in the ESS, and up to 128 concurrent
port logins per Fibre Channel port (maximum 512 concurrent port logins per ESS subsystem).
FCP ports can be configured as point-to-point or fibre channel arbitrated loop (FC-AL). FCP
servers can be directly attached to an ESS, or can be SAN attached using a fibre channel
director or switch.
On ESCON and FICON interfaces, the ESS runs as a 2105 control unit with one or more
S/390 logical control units (LCUs) attaching variable size IBM 3390 devices in either 3390 or
3380 track format. The ESS provides multiple ESCON and/or FICON interfaces and a set of
control unit images, each with multiple disk devices. The storage provided by the ESS for
ESCON and FICON interfaces is configured so that it is accessible from any of the ESCON or
FICON ports.
Fibre Channel/FICON host adapters for SCSI-FCP or FICON attachment are only available
on the ESS models F10 and F20 (FC 3021 for long wave and FC 3023 for short wave). The
original short-wave Fibre Channel host adapters (FC 3022) are available on both the previous
“E” models and the current “F” models of the ESS.
The ESS is composed of the following major components:
Host adapters: The ESS can have up to sixteen host adapters (HA). Each host adapter
provides one or two host interfaces. A host adapter can communicate with either cluster
complex.

Chapter 1. Introduction and positioning
5
Clusters: The storage server is composed of two clusters that provide the facilities with
advanced functions to control and manage data transfer. Should one cluster fail, the
remaining cluster can take over the functions of the failing cluster. A cluster is composed
of the following subcomponents:
– Cluster complex: The cluster complex provides the management functions for the
ESS. It consists of cluster processors, cluster memory, cache, nonvolatile storage
(NVS) and related logic.
• Cluster processor: The cluster complex contains four cluster processors (CPs)
configured as symmetrical multiprocessors (SMPs). The cluster processors execute
the licensed internal code (LIC) that controls operation of the cluster.
• Cluster memory/cache: The cluster memory or cluster cache is used to store
instructions and data for the cluster processors. The cache memory is used to store
the data read from, or to be written to, the disk drives. The cache memory is
accessible by the cluster complex, by the device adapters in the cluster, and by the
host adapters.
• Nonvolatile storage: The nonvolatile storage (NVS) is used to store a nonvolatile
copy of active written data. In normal operation, the NVS in one cluster stores active
write data belonging to logical disks managed by the other cluster. The ESS always
has two copies of write data — one in volatile cache on one cluster and one in NVS
on the other cluster — so that data integrity is ensured if a cluster fails before data
is destaged to disk.
– Device adapters: Each cluster has associated four device adapters (DAs). Each
device adapter provides two serial storage architecture loop device interfaces. Disk
drives are attached to a pair of device adapters, one in each cluster, so that the drives
are accessible from either cluster. At any given time, a disk drive is managed by only
one device adapter.
– Disk drives: These provide the primary nonvolatile storage medium for any host data
stored within the ESS. The disk drives are grouped into ranks and are managed by the
clusters.
As a member of the IBM Seascape family, the ESS provides the outboard intelligence
required by Storage Area Network (SAN) solutions, offloading key functions from host
servers, which frees up valuable processing power for applications. As a comprehensive
SAN-based storage solution, the ESS provides considerable management flexibility to meet
the fast-paced demands of current and future applications.
Following are some of the many factors that make the IBM ESS an ideal SAN solution:
Support for all major server platforms, including zSeries, iSeries, pSeries, xSeries, other
Windows NT and Windows 2000 servers, other Netware servers, and many varieties of
UNIX
Fibre Channel attachment capability
Extensive management capabilities through a Web interface
Excellent scalability:
– From 420 GB to over 22 TB
– Drive capacity intermix (the ability to mix drive sizes), thus giving you an increased
number of configuration options
Performance optimized to your heterogeneous environment needs
– High bandwidth and efficient algorithms that provide solutions for both online and batch
applications

6
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
– Innovations such as Parallel Access Volumes (PAV) to reduce resource contention and
dramatically improve performance in the z/OS environment
Availability required to support e-business applications:
– Non-disruptive access to data while making a copy using Concurrent Copy in the
zSeries environment
– Business continuity through remote copy services — Peer-to-Peer Remote Copy
(PPRC) for all platforms, and Extended Remote Copy (XRC) for the z/OS users
– Rapid data duplication through FlashCopy providing extensive capabilities to exploit,
manage, and protect your information in a 24 x 7 environment
– Storage server availability through redundancy and nondisruptive service with design
for no single point of failure or repair
1.2 Functions and features
The ESS provides many functions that will help you manage your enterprise storage and
provide efficient usage of your storage resources. These functions include the following:
ESS Copy Services: These optional functions include the following:
– Flashcopy: This offers the ability to create time zero (T0) copies of logical volumes.
– Peer-to-Peer Remote Copy (PPRC): This gives the ability to create synchronous
volume copies over ESCON links.
– Extended Remote Copy (XRC): The ability to create asynchronous volume copies
over long distances for the z/OS users
– Concurrent Copy (CC): For the zSeries environment, this is the ability to create
volume or data set copies locally.
IBM TotalStorage Enterprise Storage Server Specialist (ESS Specialist): This is the
Web interface tool used to manage the ESS logical configuration.
High performance features: These are hardware or software features which include the
following:
– Parallel Access Volumes (PAV): This optional feature allows you to dramatically
reduce or eliminate IOSQ time on single z/OS or OS/390 images.
– Multiple Allegiance: This feature allows you to dramatically reduce or eliminate PEND
time for a multiple image z/OS or OS/390 environment.
– Fast write with full data integrity: Data is written to cache and NVS before I/O
completion is sent to the server issuing the I/O. Applications with sensitivity to write I/O
service times, such as database logging, benefit significantly with this feature.
– SCSI command tag queueing: For open systems logical disks, the ESS can perform
multiple I/Os in parallel, similar to the PAV in the zSeries processors. This improves
throughput to the logical disks since cache hits and I/Os involving different RAID disk
groups can be processed in parallel.
– Performance Enhanced CCW commands: These are provided for the zSeries users.
– I/O Priority Queuing: This allows z/OS users to define priority of application
workloads.
– Adaptive caching: The ESS uses adaptive caching algorithms to determine how
much data to stage from disk to cache in response to a cache miss.

Chapter 1. Introduction and positioning
7
– Sequential detect: The ESS detects a sequential read pattern and begins prestaging
data from disk to cache. This allows parallel access to the disks in an array, increasing
the sustained throughput the ESS can deliver for a single sequential read stream.
– Full-stride RAID-5 writes: The ESS can accumulate full RAID-5 stripes in cache
before initiating a destage operation from cache to disk. For sequential write streams,
writing full RAID-5 stripes in a single operation avoids the traditional RAID-5 write
penalty and allows significantly higher sequential bandwidth. Server applications need
not be RAID-5 aware in order to benefit from full-stride RAID-5 writes in the ESS.
– Custom Volumes: For zSeries users, this is the ability to create your own
custom-sized logical volumes.
– Large Volume support (LVS): This is the ability to define a 32,760 cylinder 3390-9
volumes for the z/OS and z/VM users.
Scalability features:
– Disk capacity is scalable from 420 GB to over 22 TB.
– Disk drive capacity intermix offers flexible configuration options.
Availability feature for S/390 servers:
– Control Unit Initiated Reconfiguration (CUIR): This helps by automating channel path
quiesce and resume actions thereby reducing the manual actions required during
selected ESS service actions.
Serviceability feature:
– Machine reported product data (MRPD): This is an aspect of vital product data (VPD)
which facilitates system management of the ESS. This provides the ESS with the
capability to report on its current hardware and software configuration. This information
can be transmitted back to IBM electronically, through fax, hardcopy or diskettes.
1.3 Explanation of terms
Since this book will be used by planning and implementation experts with very different
background and skills, following are some of the major terms and expressions used
throughout the book:
One decimal gigabyte (GB): 1,000,000,000 bytes. Throughout this book, all disk capacities
are expressed in decimal form
One binary GB: 1,073,741,824 bytes. Cache, NVS and memory capacities are expressed
in binary terms
S/390: This refers to the System 390, zSeries, 9672 G5 and G6, or previous S/390
processors.
Open: This refers to SCSI or Fibre Channel attached systems, such as UNIX, Intel based,
xSeries, iSeries, or pSeries environments
Device end: Task complete or I/O complete (in the zSeries I/O protocol)
Task end: Task complete or I/O complete (in the SCSI command-layer protocol)
LUN (Logical Unit Number): ESS logical unit number that is an open systems unit which is
created on an ESS array
hdisk: Device label referred to by open systems operating systems such as pSeries
running AIX
vpath: A virtual path to a LUN which has multiple SCSI or fibre paths to it. It is the host
device label used in place of a LUN that has only one path or connection to it.

8
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
ESCON: Enterprise Systems Connection Architecture which is an ESA/390 computer
peripheral interface. The I/O interface utilizes ESA/390 logical protocols over a serial
interface that configures attached units to a communication fabric.
FBA (Fixed Block Architecture): Track format used for logical devices used by open
systems and accessed through SCSI (Small Computer System Interface) ports in the
ESS. FBA devices in the ESS are often referred to as FB or SCSI devices.
CKD (Count Key Data): Track format used for emulated 3390 logical devices used by
S/390 systems and accessed through ESCON or FICON ports. CKD devices in the ESS
are often referred to as S/390 devices, ESCON or FICON devices.
JBOD (just a bunch of disks): Group of individual disk drives not configured as a RAID
(redundant array of independent disks) array.
Rank: A RAID array or an individual JBOD disk drive. A rank is a unit of physical disk
storage in the ESS that is allocated to either FB or CKD logical volumes.
Host adapter: To those familiar with UNIX, a host adapter is an adapter on the host.
Throughout this book, a host adapter (HA) is a SCSI, ESCON, FICON or Fibre Channel
adapter which resides in the ESS and is used to connect the ESS to the host server bus.
LSS (Logical Subsystem): This equates to a S/390 logical control unit (LCU). For the open
environment, LSS is a new term used by the ESS. An LSS is used internally to manage a
set of logical volumes associated with an individual device adapter. The LSS is made up of
ranks managed by a single device adapter. Ranks on different loops both managed by a
single device adapter can be part of the same LSS
Disk group: This refers to a group of eight drives that can be defined as either a RAID
array or non-RAID ranks.
SSR: This refers to the IBM System Support Representative who is responsible for the
physical installation and maintenance of the ESS.
FICON (FIber CONnection): This refers to an ESA/390 and zSeries computer peripheral
interface. The I/O interface uses ESA/390 and zSeries FICON protocols (FC-FS and
FC-SB-2) over a Fibre Channel serial interface that configures attached units to a FICON
supported Fibre Channel communication fabric.
CUIR (Control Unit Initiated Reconfiguration): This refers to an S/390 feature which
automates channel path quiesce and resume actions in support of selected ESS service
actions.
1.4 Platform support
The ESS provides support for many different platform connections. These include most forms
of UNIX, iSeries, Intel based servers running NT, Windows 2000, Novell Netware, or Linux,
and the zSeries processors. For a detailed description of the latest server support for both
hardware and software levels please refer to the following Web address:
http://www.storage.ibm.com/hardsoft/products/ess/supserver.htm
1.5 Getting ESS-related information
Along the different steps needed to implement the ESS in your environment, the people
involved in the installation activities will be referencing other ESS-related manuals. In this
section we summarize the ESS documentation and manuals that are referenced throughout
this redbook, and explain where to get that information.

Chapter 1. Introduction and positioning
9
1.5.1 ESS Web site
The following URL takes you to an ESS Web site where you can start most of your searches
for additional information on the IBM TotalStorage Enterprise Storage Server:
http://www.storage.ibm.com/hardsoft/products/ess/ess.htm
The Web page you will see at this URL is shown in Figure 1-1.
Figure 1-1 ESS: Starting Web site
If you move down the scroll bar to see the bottom of this Web page, you will see a More
Information
area in the page (see Figure 1-2).
Figure 1-2 The More Information area at the ESS Web page
This More Information area of the ESS-related Web pages will be the starting point for much
of the information and material you will be needing during your implementation activities.

10
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
1.5.2 List of supported servers
Click the option Supported Servers (see Figure 1-2) and you will reach the following URL:
http://www.storage.ibm.com/hardsoft/products/ess/supserver.htm
From this ESS Web page, you can make the following selections to get the corresponding
material:

ESS Open Systems Support Summary
ESS with IBM zSeries and S/390 Systems Support Summary
ESS Supported Servers. This is a PDF download with the most complete and detailed
support information for server attachment

SDD (Subsystem Device Driver). If you click this option, it will take you to the SDD Web
site at:
http://ssddom01.storage.ibm.com/techsup/swtechsup.nsf/support/sddupdates
At this site you will find SDD installation support material (downloads, fixes). From here
you can also get, whether viewing online or download as PDF, the
IBM TotalStorage
Enterprise Storage Server Subsystem Device Driver Installation and User’s Guide,
GC26-7442.

ESS HBA FC Supported Versions. This is the Host Bus Adapter Supported Fibre Versions
Matrix. This is a matrix which correlates the different models of Fibre Channel I/O
adapters to the drivers and to the server operating system level requirements.

ESS Host Systems Attachment Guide. When clicking this option, you either view online or
you download the PDF of the
IBM TotalStorage Enterprise Storage Server Host Systems
Attachment Guide 2105 Models E10, E20, F10 and F20,
SC26-7296 manual.

ESS Copy Services Command Line Interface User’s Guide. When clicking this option, you
either view online or you download the PDF of the
IBM TotalStorage Enterprise Storage
Server Copy Services Command-Line Interface Reference,
SC26-7434 manual.
1.5.3 ESS documentation Web site
If when at any of the ESS-related Web pages you click the option Reference Information
(see Figure 1-2) you will reach the following URL:
http://ssddom02.storage.ibm.com/disk/ess/documentation.html
The page you will see is shown in Figure 1-3.

Chapter 1. Introduction and positioning
11
Figure 1-3 ESS documentation Web site
From this Web site, you can either view online (left click) or download the corresponding
PDFs (
Save Target As option after right clicking) for the following ESS documentation:
An ESS overview of the ESS characteristics
The ESS specifications sheet
The
IBM TotalStorage Enterprise Storage Server Host Systems Attachment Guide 2105
Models E10, E20, F10 and F20,
SC26-7296
The
IBM TotalStorage Enterprise Storage Server Introduction and Planning Guide,
GC26-7294
The
IBM Enterprise Storage Server SCSI Command Reference 2105 Models E10, E20,
F10 and F20,
SC26-7434

IBM Enterprise Storge Server System/390 Command Reference 2105 Models E10, E20,
F10 and F20,
SG26-7298
The
IBM TotalStorage Enterprise Storage Server User’s Guide 2105 Models E10, E20,
F10 and F20,
SC26-7296
The
IBM TotalStorage Enterprise Storge Server Web Interface User’s Guide, SC26-7346
The
IBM TotalStorage Enterprise Storage Server Copy Services Command-Line Interface
Reference,
SC26-7434
The
IBM TotalStorage Enterprise Storage Server Configuration Planner, SC26-7353
The
IBM TotalStorage Enterprise Storage Server Quick Configuration Guide, SC26-7354
The
ESS Configuration Worksheets

12
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
1.5.4 Redbooks
Besides this redbook, for your ESS implementation planning and hands-on activities you will
be referencing other ESS-related redbooks that go into more detail on specific tasks involved
in the implementation process.
To obtain these related redbooks, you must go to the IBM Redbooks site at:
http://www.redbooks.ibm.com
Alternatively, you can enter:
http://www.ibm.com/redbooks
Once here, you click the Redbooks Online option (on the left navigation bar), and you can get
to:
http://publib-b.boulder.ibm.com/Redbooks.nsf/portals/
From this site, you can make the appropriate selections to get the redbooks you may be
needing to reference.

© Copyright IBM Corp. 2002
13
Chapter 2.
Configuration planning
This chapter overviews a set of planning tasks that should be considered to ensure that the
ESS will be sized and configured according to your storage needs and performance metrics.
The configuration planning process will most likely be an iterative process until all parties
involved in the planning agree that a valid configuration has been reached.
2

14
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
2.1 Introduction
The Enterprise Storage Server (ESS) is designed for sharing its storage capacity between a
mix of both S/390 and Open server host systems.
As a result, the people involved in the capacity requirements and the logical configuration
planning will be from several different groups that are not traditionally combined, as shown in
Figure 2-1.
Figure 2-1 The expanded planning team
The disk capacity requirements from all people intending to store data in the ESS will need to
be combined to decide the total ESS physical capacity. Additional details relating to the
number of server hosts to connect, and what the host’s view of the storage allocation will be,
are used to build the ESS logical storage configuration.
This logical plan is then used to configure the ESS internally using the IBM TotalStorage
Enterprise Storage Server Specialist (ESS Specialist) as part of the installation process. For
further information on planning the ESS configuration refer to the IBM publication
IBM
TotalStorage Enterprise Storage Server Configuration Planner,
SC26-7353 (see Section 1.5,
“Getting ESS-related information” on page 8 for information on how to get this IBM manual).
2.2 Capacity planning
This section is intended to aid the sizing of the physical storage capacity required for your IBM
TotalStorage Enterprise Storage Server. The tables included in this section will enable you to
calculate the total number of 8-pack disk arrays required when ordering the ESS. The process
involves:
Calculating the number of arrays required for SCSI and /or FCP fixed block format data
Calculating the number of arrays required for S/390 CKD format data
Combining the two for a total ESS capacity requirement
z/OS Storage Manager

Chapter 2. Configuration planning
15
It is assumed that you have a base knowledge of the ESS architecture, functions and terms
used when describing the ESS. If the terms or concepts used are unfamiliar, it is
recommended that you read the IBM Redbook entitled
IBM TotalStorage Enterprise Storage
Server,
SG24-5465.
2.2.1 Capacity and configuration rules
This section is intended as a quick reference of rules and restrictions for the different
components and functions available on the ESS.
Physical features
Physical features of the ESS include:
Maximum 16 host adapter (HA) cards, ESCON, SCSI, Fibre Channel/FICON, or
combination of those HAs. ESCON and SCSI adapter cards have two ports, therefore, you
can have a maximum of 32 physical host connections when you use ESCON, SCSI or a
combination of only ESCON and SCSI HA. Fibre Channel adapter cards, for FICON and
SCSI over Fibre Channel (FCP), have a single port, thereby giving you a maximum of 16
host connections when using only Fibre Channel HA cards.(Fibre Channel host adapter
cards do not support the simultaneous transmission of both protocols, they must be
configured for one or the other.)
Four Device Adapter (DA) pairs with two loops per DA pair for a total of 8 loops on the
ESS.
Two to six 8-pack arrays per loop.
All the drives within an 8-pack are of the same capacity - 9.1 GB, 18.2 GB, 36.4 GB or
72.8 GB.
When the disk packs on the loops are of the same capacity, each RAID-5 loop will have
two spares. Therefore, the first two arrays on the loop will be 6+P+S, the next four arrays
will be 7+P.
Multiple drive sizes are supported on the same loop.The first two RAID-5 arrays
configured on each loop for each drive size are configured as 6+P+S, providing two
spares for each drive size.
Logical features
Logical features of the ESS include:
The arrays in a loop (maximum 6) may be assigned to any of the 8 LSSs on the DA pair.
You can have RAID and JBOD ranks on the same LSS or loop.
You can have CKD and FB formatted arrays/ranks on the same loop, but not on the same
LSS.
A rank can belong to only one of the LSSs on a DA pair.
The CKD ranks on both loops accessed by one device adapter can be assigned to 1 or 2
of the LSSs of the device adapter.
The FB ranks on both loops accessed by one device adapter can be assigned to 1 or 2 of
the LSSs of the device adapter.

16
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
0/8/16 CKD LSSs per ESS
S/390 ESCON host:
– Maximum 4096 devices per ESS
– Maximum 256 devices per LCU
– Maximum 1024 devices per ESCON channel (CHPID)
– Maximum 64 logical paths per ESCON port
– Maximum 128 logical paths per LCU with up to 64 path groups
– Maximum 2048 logical paths per ESS
S/390 FICON host:
– Maximum 4096 devices per ESS
– Maximum 256 devices per LCU
– Maximum 4096 devices per FICON channel (ESS maximum devices)
– Maximum 256 logical paths per FICON port
– Maximum 128 logical paths per LCU
– Maximum 2048 logical paths per ESS
0/8/16 Fixed Block LSSs per ESS
SCSI host:
– Maximum 960 LUNs (15 target x 64 LUN) per ESS host adapter port. This maximum is
supported by the ESS but may be further limited by the host operating system support.
– Maximum 256 devices per LSS, and 4096 FB devices per ESS.
– 1 to 15 SCSI targets per SCSI bus/port.
– Either up to 8 or up to 32 LUNs per target, depending on the operating system support.
– Up to four initiators (hosts) per SCSI port, for selected operating system platforms.
– LUNs may be shared between hosts, with hosts controlling contention.
– The Subsystem Device Driver (SDD) manages all the LUNs configured on the ports
with which it works.
FCP host:
– Maximum 256 LUNs per target (port login) for host that do not support the SCSI
command “Report LUNs”, and 4096 LUNs per target (port login) for hosts that support
the SCSI command “Report LUNs”.
– One target per Fibre Channel port login.
– Up to 128 port logins per ESS Fibre Channel host adapter port. Each of the up to 128
server host I/O adapters performing a port login to an ESS Fibre Channel host adapter
port will all see LUNs on a single SCSI target, but they’ll potentially see different LUNs.
– Maximum of 512 port logins per ESS.
– Maximum 256 devices per LSS, and 4096 FB devices per ESS.
– LUNs have affinity to the server Fibre Channel I/O adapter, independent of the ESS
Fibre Channel host adapter port the server is attached to. Three implementations can
be used to avoid integrity problems: install Subsystem Device Driver (SDD) in the
server; create zones in the switch fabric; use the ESS function “Access Control by Host
by Port” to restrict the host’s access to a single port.
S/390 host view
Figure 2-2 shows an example of the S/390 logical path limits and considerations for an
ESCON implementation.

Chapter 2. Configuration planning
17
Figure 2-2 S/390 logical paths using ESCON
SCSI host view
Figure 2-3 shows an example of the logical SCSI bus view of the ESS.
Figure 2-3 SCSI logical configuration
S/390
LPAR
LPAR
...
S/390
LPAR
LPAR
...
S/390
LPAR
LPAR
...
S/390
LPAR
LPAR
...
4 CECs with 8 LPARs Each - 32 Way Sysplex
2105 Enterprise Storage Server
CU
0
...
CU
F
32 Links x 64 LPs
= 2048 LPs
32 LPAR X 16 CU
X 4 Paths
= 2048 LPs
16 Links
(8 Adapters)
16 Links
(8 Adapters)
32 LPAR x 4 Paths
= 128 LPs/CU
= 32 PGs/CU
LP=Logical Path
PG=Path Group
16 CUs / 4096 Devices
8 LPARs x 16 CUs
x 4 Paths
= 512 LPs/CEC
4 Links / 4 CU
Due to 1024
Dev/Channel Limit
= 16 Links/CEC
ESCD ESCD

32 Logical Devices
32 LUNs
32 LUNs

18
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Figure 2-4 shows an example of the Fibre Channel logical configuration of the ESS.
Figure 2-4 Fibre logical configuration
2.2.2 Storage capacity options
For capacity planning purposes, the effective data capacity of the ESS is reduced from it’s
theoretical maximum, using the raw disk sizes 9.1 GB, 18.2 GB, 36.4 or 72.8 GB, due to the
following:

Disk sector overheads:
– Disk sector size is 524 bytes, data area 512 bytes
– Metadata area on each disk

The formatting of disks into:
– RAID-5 or non-RAID support
– SCSI and/or Fibre Channel FB devices or S/390 CKD devices
ESS model capacity
Table 2-1 details the minimum and maximum RAID-5 storage capacity available by the ESS
Model and is mapped against the available physical disk drive sizes. The table is designed to
assist you in the selection of an appropriately sized disk drive/ ESS model combination that
will meet your current and future capacity needs.
It is suggested that you select a configuration closest to your needs, since this will simplify the
ordering process and speed the installation process.
Fibre Channel host adapters
Up to 16 Fibre Channel host adapters
One port with an SC connector type
per adapter (Gigabit Link Module)
Long wave or short wave option
100 MB/sec full duplex
up to 10 km distance with long wave /
500m short wave
Fibre Channel Protocol (FCP)
Supports switched fabric
Fully participates in Storage Area
Networks (SANs)
LUN masking
Storage Area Network
UNIX
Intel
400
iSeries

Chapter 2. Configuration planning
19
Table 2-1 ESS RAID-5 disk storage capacity by model (rounded)
2.2.3 Capacity planning by storage format
The next step in your planning requires you to separate your total storage needs into the
different format types, SCSI and/or Fibre Channel FB or S/390 CKD. The tables in this section
can be used to aid you in determining the number of 8-pack arrays required for each format
and, therefore, the total for the ESS.
To aid in planning, the RAID array capacity data is shown separately for ESS arrays 1 to 8 for
the Model F10 and 1 to 16 (base frame), and 17 to 48 (expansion frame) for the Model F20.
This is due to the fact that the first two RAID-5 arrays on a loop contain a spare disk (the
exception being JBODs). Therefore all the RAID-5 arrays in the base ESS will contain a spare
disk, 6+P+S. The RAID-5 arrays in the expansion frame will all be 7+P arrays if the
configuration consists of similar capacity disk drives only.

But if you choose to intermix drive
capacities on the same loop, because the first two RAID-5 arrays for each capacity must be
6+P+S, then the expansion frame can also start to become populated with 6+P+S arrays.
For your existing ESSs, when planning for the capacity requirements, you may want to
consider the possibility of upgrading the installed eight-packs to the higher capacity disk
drives of 36.4 GB or 72.8 GB.
Why total the arrays?
The minimum unit of storage installable or upgradeable within the
ESS is two 8-packs.
Effective data capacity
Each group of eight disks can be formatted into RAID-5 or non-RAID JBOD support. In the
case of RAID support, a single partition, called a Rank in ESS terminology, is allocated
across the entire array of disks. For non-RAID JBOD support, the single partition is allocated
on a single disk drive, one for each physical disk. Logical volumes or logical disk devices are
placed within these partitions, or ranks.
Therefore, the first decision to effect the total data capacity available is how many RAID-5 or
non-RAID JBOD disk ranks are required. For example, assuming a disk group of 8 x 36.4GB
disks the maximum available capacity for:

JBOD is 280 GB.

RAID-5 is 245 GB (if a 7+P disk group).
The effective data capacity of a rank is also dependent on how you configure the logical
volumes; SCSI FB or S/390 CKD. As a result, you need to calculate the effective data
capacity required in terms of arrays for each of the logical volume types.
When defining standard logical volume sizes or formats onto a disk or array, you may not be
able to fully utilize the storage space available. This is true if the disk or array size is not a
multiple of the LV size. You can, therefore, define smaller sized LVs, or S/390 custom
volumes, to best utilize the remaining disk storage.
ESS 9.1 GB 18.2 GB 36.4 GB 72.8 GB
Minimum F10/F20 420 GB 420 GB 841 GB 1683 GB
Maximum F10 420 GB 841 GB 1683 GB 3367 GB
Maximum F20 Base 841 GB 1683 GB 3367 GB 6734 GB
Maximum F20 w/Expansion 2804 GB 5610 GB 11224 GB 22449 GB

20
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
JBOD data capacity
A JBOD has one non-interleaved partition on each physical disk device in the array, thus
there are 8 JBODs/ranks per group of drives. The Logical Volume (LV) size within a JBOD is
0.5 GB to 72 GB, dependent on the physical disk drive installed. A JBOD can be defined as
either SCSI FB or CKD.
A JBOD does not require a spare disk, and therefore will take over the entire 8 disks as
non-RAID data disks, regardless of the position. For JBOD array capacities based on disk
capacity, see Table 2-2.
Note
: You must remember that JBOD ranks are not fault tolerant, so they should be used in
conjunction with operating system mirroring techniques to provide fault tolerance.
Table 2-2 JBOD array capacity
SCSI or Fibre Channel fixed block data capacity
A SCSI FB RAID-5 rank has one partition that is striped across all the disks within the array
and, therefore, the maximum storage size available for placing Logical Volumes is the
capacity of the entire array, as shown in Table 2-3. UNIX, AIX or Windows NT, can allocate
LVs from 0.5 to 491 GB. Multiple LVs can be placed in an array/rank.
OS/400 supports 9337 devices with SCSI and 2105 devices for Fibre Channel Logical
Volume sizes of 4.19 GB, 8.59 GB, 17.54 GB, 35.16 GB, 36 GB and 70.56 GB. Multiples of
these can be placed into an array. OS/400 volumes can be intermixed on an ESS array with
other open systems LUNs.
Table 2-3 RAID-5 array capacity - FB
S/390 CKD data capacity
S/390 effective data capacity will differ from the FB capacity in a rank due to the 3390
standard device formatting required. Also CKD has additional overheads, for example the
count and key data (CKD), the home address (HA), and the record zero (R0) fields.
The CKD track format can be placed on a RAID-5 array or non-RAID JBOD rank. The RAID-5
array can be formatted with interleaved logical volumes or non-interleaved logical volumes.
Disk drive capacity JBOD array capacity (8 + no parity or spares)
9.1 GB 69 GB
18.2 GB 140 GB
36.4 GB 280 GB
72.8 GB 561 GB
Disk Drive Capacity Arrays 1 - 16 (F20)
Arrays 1 - 8 (F10)
(6 + P + S)
Arrays 17 - 48 (F20)
(7 + P) (see Note)
9.1 GB 52 GB 61GB
18.2 GB 105 GB 122 GB
36.4 GB 210 GB 245 GB
72.8 GB 420 GB 491 GB
Note: Arrays 17 to 48 will all be 7+P only if there is no disk drive capacity intermix. If the configuration
consists of intermixed capacity disk drives, then arrays 17 to 48, some of them, will be 6+P+S. See
explanation in 2.2.3, “Capacity planning by storage format” on page 19.

Chapter 2. Configuration planning
21
Interleaved:
Interleaved arrays have two partitions striped across the array. The first partition
is in interleaved mode and accounts for the major portion of the array capacity and is used to
place 3390 or 3380 standard logical volumes (not custom volumes, not large 32760 cylinder
volumes). LVs are placed within the partition in groups of four until there isn’t enough storage
space in the array to create another group of four still keeping 5,000 cylinders left over for the
non-interleaved partition.
The second non-interleaved mode partition in the array will be a minimum of 5,000 cylinders
of 3390 volumes and can be used to define additional Logical Volumes (either standard 3390
Models or Custom Volumes).
Interleaved applies to arrays with disk capacities of 9.1 GB, 18.2 GB and 36.4 GB.
Non-interleaved:
A non-interleaved mode array has only one partition that is striped across
the array. The entire partition can be used for logical volumes (either standard 3390 models or
custom volume).
The maximum CKD logical volume size is dependent on the 3390 model type that has been
formatted on the array (3390-2 1.89 GB, 3390-3 2.83 GB, 3390-9 8.49 GB) or the size of the
large capacity volume, 32760 cylinders. A custom volume can be from 1 to 10017 cylinders in
size, or from 1 to 32760 cylinders in size for a system with Large Volume Support.
S/390 capacity tables can be found in “CKD logical volumes” on page 106 which show the
various disk capacities and the number of logical volumes that can be defined.
2.2.4 ESS Copy Services
When using the ESS Copy Services functions offered by the ESS, the following
considerations for capacity, pathing and ESS resources should be factored into the planning:

FlashCopy provides a time zero (T0) image copy within the same LSS or LCU.
Consideration needs to be made for the number of logical volumes that you may wish to
use as target logical volumes since the disk capacity must be within the same LSS or LCU.

PPRC will require additional disk storage reserved in the second (remote) ESS which will
host the target volumes.

XRC for S/390 will also require additional disk storage in the second (remote) ESS for the
target volumes. You can use ESCON or FICON channels to connect the sites if the
distance allows it, else you have to use channel extenders and telecommunication lines.
One of the key elements in the sizing of an XRC configuration is the required bandwidth
between the SDM (System Data Mover) and the primary ESS.

PPRC requires that a primary and secondary ESS be connected by ESCON connections,
for both S/390 and the open systems environments. These paths are unidirectional and
the primary ESCON adapter ports are dedicated to PPRC and cannot be used for normal
host data traffic while in a PPRC session. Remember to consider these points during
planning, since they may effect the number and type of host adapters required.

When using ESS Copy Services, one ESS cluster is the ESS Copy Services primary
server which is the central place for collection of all information. A second ESS cluster can
be nominated as the backup server. You will need to decide which ESS cluster to
nominate for each role to ensure maximum availability. The IBM SSR will need this
information during the ESS installation.
For detailed information on ESS Copy Services see
Implementing ESS Copy Services on
UNIX and Windows NT/2000,
SG24-5757 and
Implementing ESS Copy Services on S/390,
SG24-5680.

22
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
INRANGE
Previous PPRC / INRANGE 9801 support is enhanced with greatly increased PPRC
ESTABLISH or RESYNC performance at long distances. INRANGE supports new uses of
ESS PPRC for data migration, data copy, or incremental resync at long distances.

New ESS PPRC usages: open systems and/or S/390 data migration, data copy at long
distances

Utilizes new level of microcode in existing INRANGE 9801 Storage Network Subsystem
hardware, combined with latest ESS microcode. See Figure 2-5.

INRANGE supports OC3/T1/T3/E1/E3/ATM/ IP channel extension of PPRC

ESTABLISH/RESYNC performance improvement: up to 40% reduction in elapsed time at
short distances, up to 90% reduction in elapsed time at long distances
Figure 2-5 INRANGE 9801 storage
INRANGE can be used for long distance disaster recovery in two ways:

Send log files long distance using PPRC as part of the log file archive process, with
excellent performance.

If the primary site can tolerate application quiesce as part of usage, then you can run
PPRC pairs in a RESYNC/SUSPEND loop operation.
Do a RESYNC at appropriate intervals to create a backup copy.
This will require application-sponsored checkpoints or quiesces to insure production primary
volume performance objectives are met.
Note:
This enhancement does not change distance or performance for continuous
synchronous PPRC.

Chapter 2. Configuration planning
23
2.2.5 Flexible configurations
The ESS has the ability to support mixed drive sizes (intermix). With drive intermix, a vast
number of configuration options are available, so price and performance can be optimized to
meet specific server and application requirements within a single ESS.
Intermix support is available on all ESS models. Any type and quantity of available disk
8-pack features can be intermixed as long as 8-packs of a given DDM capacity are purchased
and added to a loop in pairs and supported on a given ESS model.
Intermix support requires the new Flexible Capacity Option features (#9500 and
#9600).These new features replace the Standard Configuration features previously used
when specifying ESS capacity and drive type. With the Flexible Capacity Option, capacity is
now ordered by explicitly selecting the type and quantity of disk 8-pack features desired.
2.3 Logical configuration planning
The logical configuration of the ESS involves defining how the ESS is seen from the attached
hosts. The UNIX, AIX, NT system hosts will see the ESS as SCSI and FCP generic devices,
the AS/400 will see it as a 9337 or 2105 external disk, and the S/390 host will see it as a 2105
control unit with one or more logical control units (LCUs) attaching IBM 3390 devices.
The basic steps required are as follows:

Draw up a logical map or plan of the storage subsystems and devices you want to emulate
within the ESS.

Map these to the physical hardware arrays within the ESS.

Review the ESS capacity and adjust as necessary.

Combine all the decisions made and document them in the form of a specification or
logical configuration plan, which can then be used to setup the ESS using the ESS
Specialist during installation.

Give the logical specification to the host systems or storage software specialist to enable
the appropriate I/O definitions to be completed
Note: The steps shown in this section do not reflect the logical worksheet process which is
available to help you pre-define your logical layout, and then help the IBM SSR input these
requirements at installation time. For more information on the initial configuration process,
please refer to Appendix A, “ESS configuration planning process” on page 293.
2.3.1 Logical subsystem mapping
Determine the number of SCSI and/or FCP or CKD subsystems (LSS/LCU) to be defined.
Multiple targets/LUNs can share the same associated logical device.
Decisions that need to be made for SCSI and FCP devices include the following:

How will the hosts be connected to the ESS?

Determine RAID or non-RAID format.

Determine the LV or device size.

Define the number of targets and the number of LUNs per target.

Will LUNs be shared on two or more hosts?

Which will be the host attachment: Fibre Channel, SCSI?

Addressing requirements: Are specific SCSI IDs required for host and target? It is
recommended to set the initiator ID to 7.

Consider that iSeries and AS/400 6501 SCSI IOP can only access 1 to 8 logical disks.

24
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
A S/390 LSS relates directly to a S/390 logical control unit (LCU) with its associated devices
or volumes. A four digit subsystem identifier SSID will need to be assigned to each logical
control unit. Also when planning for S/390 CKD LSS mapping, PAV devices must be taken
into account towards the 256 LSS device limit and also the 1024 devices limit per ESCON
CHPID (16384 for FICON).
LSS/LCU to physical mapping
This is how the LSS physical mapping is implemented.
S/390:
The CUADDR address is determined by the DA and loop that the LCU is associated
with when defined using the ESS Specialist. For example, in Figure 2-6, the CUADDR for
cluster 1, adapter 1, loop B is 08.
SCSI and Fiber Channel:
The mapping of the SCSI and Fibre Channel FB LSS to
target/LUN is also dependent on the DA and loop association, and is automatically assigned
during ESS Specialist setup. The SCSI and Fibre Channel FB LSS numbers are shown in
Figure 2-6 below.
Figure 2-6 LSS logical to physical mapping
Disk group allocation
The disk groups on a loop are allocated evenly between the two device adapters. Therefore,
on a fully populated loop, only three arrays will be available to the LSSs on each adapter. For
example, LCU 00 or 08 will have access to disk groups 2, 4, and 8on the loop, with LCU 01 or
09 accessing disk groups 1, 3and 5. The ESS Specialist manages this for you.
Logical device mapping
This is how the logical device mapping is implemented.
DA to LSS Mapping
Cluster 2Cluster 1
LSS LSS
FB CKD CKD FB
03 13
Adapter 2
0B 1B
05 15
Adapter 3
0D 1D
07 17
Adapter 4
0F 1F
01 11
Adapter 1
09 19
10 00
Adapter 1
18 08
12 02
Adapter 2
1A 0A
14 04
Adapter 3
1C 0C
16 06
Adapter 4
1E 0E
A
A
A
A
A
A
A
A
B
B
B
B
B
B
B
B
Cluster 1, Adapter 1, Loop B, Group 2

Chapter 2. Configuration planning
25
Array allocation:
The first step has to do with the definition of which arrays/ranks are to be
assigned/mapped to which LSS/LCU.
For S/390 you must associate at least one disk group with the first CKD LSS for the device
adapter. If there is at least one disk group associated with the first CKD LSS for the device
adapter, other CKD disk groups can be associated with either the first or the second CKD
LSS for that device adapter.
For open systems when a disk group is formatted as an FB array, if there are already 193 or
more LUNs defined in the first LSS for a device adapter. the new disk group will be added to
the second LSS for the device adapter.
RAID or JBOD:
The disk arrays to be attached to the LSS/LCU then need to be defined as
RAID-5 or non-RAID JBOD disks. This will determine the rank storage capacity available. For
example, a single disk rank for JBOD and a full disk array for RAID-5.
Logical Volumes:
For each array/rank, specify the number and size of the logical volumes or
devices required.
SCSI and Fibre Channel:
You can map the same logical devices to different SCSI and/or
Fibre Channel ports to provide shared logical volumes. However, the host software must
handle the data contention.
A JBOD can have more than one logical volume assigned to it.
The ESS maximum number of SCSI LUNs (or LVs) per target is 64 and 16,000 for Fibre
Channel, however, it is host operating system or I/O adapter dependent. For example, many
servers operating system only support 8 LUNs per target for SCSI-attached storage and 32
LUNs per target for Fibre Channel attached storage. Similarly with Fibre Channel the
non-Report LUNs host support 256 LUNs per target, while the Report LUNs hosts support
4096 LUNs per target.
S/390:
Decide whether you will be using Interleaved or non-interleaved volumes. The CKD
logical volumes defined in the interleaved partition, will match full 3390-2 (1.89GB) or 3390-3
(2.89GB) or 3390-9 (8.51GB). For the remaining non-interleaved partition, which will be the
custom volumes you will be defining?.
A decision must be made on the logical device type to emulate, 3390 Model 2, Model 3,
Model 9, or large capacity volumes, or 3380 track format. If using custom volumes, decide
how many volumes are required and how many cylinders are to be allocated to the volumes.
Custom volumes can only be defined on non-interleaved partitions.
Parallel Access Volumes:
If you need to use PAVs, the LCU needs to have the PAV option
enabled via the ESS Specialist (
Note
: the PPRC feature has to be previously activated by the
SSR during the ESS installation or later if this feature is added on an existing ESS). The
volumes that require PAV aliases need to be identified, along with the number of aliases to be
available to each volume.
Host connections
The number of connections between a server and the ESS must be decided based on the
bandwidth requirements of the server. For zSeries servers, this will be the number of FICON
or ESCON channels in channel path groups attached to the ESS. For open systems servers,
it will be the number of SCSI busses or Fibre Channel I/O adapter-to-ESS port connections.
The ESS SCSI logical volumes are specifically assigned to an HA port via the ESS Specialist
at ESS setup. If using the IBM SDD, the physical SCSI and/ or Fibre Channel connections
should be spread across different HA bays.

26
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
For Fibre Channel open systems servers, LUNs are assigned to a sever host bus adapter by
its world-wide port name (WWPN). A server host bus adapter can be restricted using the ESS
Specialist to accessing its LUNs through selected ESS Fibre Channel host adapter ports.
When planning how much capacity to attach to a SCSI or Fibre Channel interface, the
following conservative guidelines are suggested:

For Ultra-SCSI (40 MB/sec): Up to 120 GB per SCSI bus

For SCSI-2 (20 MB/sec): Up to 80 GB per SCSI bus

We recommend that no more than 400 GB are assigned per Fibre Channel port.
For S/390 this is fairly straight forward since it is no different than standard device
considerations. It is recommended that eight ESCON channels for each z/OS or S/390 image
be used. All host ESCON adapters are able to address all LCUs defined within the ESS.
However, you should spread the physical ESCON connections from one host image across
the HA adapter bays. For ESCON configurations be sure to review the logical path limitations.
With FICON things become simpler because many ESCON limits no long exist. Less FICON
channels will be used when compared to ESCON: a 4:1 consolidation can typically be
achieved. Also with FICON, thanks to its greater bandwidth and increased logical pathing
characteristics, more LCUs in the ESS can be logically daisy chained thus addressing more
devices with less channels and consequently less host adapters needed.
PPRC and XRC also have an additional requirement for ESCON host adapter ports. PPRC
primary or source volume ESS ESCON ports are dedicated and, therefore, cannot be shared
with normal host traffic while the PPRC session is active. The ports on the secondary device
can, however, be used for other host activity. For XRC (S/390 users) dedicated host adapter
ports are recommended. Also in the sizing an XRC configuration the required bandwidth
between the SDM (System Data Mover) and the primary ESS must be determined, having in
mind which are the peak demands.
Planning for the future
It is a good idea to consider your future capacity requirements and the effect this may have on
your planned ESS configuration. This may prevent a lot of data relocation work later on. Items
to consider include:

If upgrading the ESS capacity, where will the next group of arrays/ranks be placed? Also
keep in mind that arrays are installed in pairs and must be installed in fixed order.

Will this require movement or redefinition of LSS/LCU and its logical volumes due to
performance, capacity or device number limitations?

Reformatting the array from RAID-5 to JBOD, or changing the logical volume size or the
CKD device type, will require unloading and then reloading of data.
2.4 Planning for performance
This section provides some considerations on the physical and logical configuration that may
contribute to better performance. Reviewing these points is recommended before finalizing
the physical or logical plan for your ESS. The topics in this section should be complemented
with the corresponding information presented in the following chapters.
Note:
The previous attachment guidelines are based on assumptions about the average
sequential bandwidth requirements per GB of data stored.

Chapter 2. Configuration planning
27
Disk drive capacity
There are configuration decisions to be made when ordering your ESS. One is the disk drive
capacity. You have the option to choose 9.1 GB, 18.2 GB, 36.4 GB or 72.8 GB disks drives.
The greater capacity 72.8 GB drives provide superior seek and throughput characteristics,
thereby allowing each drive to satisfy big demanding I/O loads. Moreover, access densities
(I/O operations per second per GB disk space) continue to decline, so workloads migrated to
72.8 GB drives will often not notice any material difference in performance compared to
smaller capacity disk drives.
Some customers have very high demand workloads that may be very cache-unfriendly or
have a very high random write content. These workloads may still require the lower capacity
drives (more drives are required for a given capacity). For these workloads, customers may
want to consider purchasing lower capacity drives or, for environments with mixed workloads
with different characteristics, an intermix of drive capacities.
Host adapters
When physically connecting a host system to the ESS host adapter ports (ESCON, FICON,
Fibre Channel or SCSI bus port), multiple connections from the same host system should be
spread across the host adapter bays. This not only provides better availability due to the
possible loss of a bay, but also spreads the I/O load across different PCI buses in the cluster
and maximizes the available data bandwidth.
S/390:
For high performance workloads, attaching 16 ESCON channels per MVS image and
using eight-path groups to LCUs will spread the I/O channel activity at peak times. This may
not be practical in some sysplex environments, but it is recommended that at least eight
ESCON channels be used. Be sure to check the logical path restrictions.
The 8 channels for a path group should be cabled to ports A and B on 4 host adapters, 1
adapter per HA Bay. See Figure 4-9 on page 76 for more explanation.
With 16 channels from a single image, you need to spread your data across volumes evenly
split into two groups--one group associated with ESS cluster 1, one set of four device
adapters (one host adapter per host adapter bay), the other associated with cluster 2, the
second set of four device adapters (one host adapter per host adapter bay).
One eight-path group should then provide access to the LCUs defined on ESS cluster 1, for
example, LCU CUADDR 0, 2, 4, 6, 8, A, C, E. The second eight-path group should provide
access to the LCUs defined on the other Cluster 2, CUADDR 1, 3, 5, 7, 9, B, D, F. See
Figure 2-6 on page 24 for details. This will maximize the bandwidth available through the host
adapters.
To maximize the operations per second from the eight-way path group, configure all eight
channels to the same IOP or SAP. However, this may not be achievable due to host channel
availability requirements.
For FICON the approach to a good performance configuration is similar but the requirements
differ because of FICON better performance characteristics and improved logical pathing
limits. For installations where ESCON host adapters were in the 25-50% utilization, then one
FICON host adapter is enough where four ESCON host adapters were used. Plan at least
four FICON host adapters per ESS. A more typical configuration would have eight FICON
HAs per ESS (eight FICON HAs may be necessary to exploit ESS full bandwidth for some
workloads).

28
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Spread FICON HAs across all four HA bays. This should result in minimally one FICON HA
per bay, or in a typically configured ESS, two FICON HAs per bay.
For each control unit definition in HCD/IOCP, define a minimum of four FICON channels per
ESS/LSS path group (or typically eight).
Subsystem Device Driver:
The Subsystem Device Driver (SDD) for open systems can
provide multiple SCSI or Fibre Channel bus paths to the same target/LUN group, thus
spreading the I/O workload. This I/O load balancing improves the performance of the server
I/O operations.
Cache
If you are only defining one type of LSS/LCU (CKD or FB) within the ESS, it is recommended
that you set the other LSS quantity to zero. This releases additional cache storage for data,
rather than holding control information about the LSS. For example, if only using SCSI LSS,
set the S/390 CU LSS value to zero. The LSS/LCU maximum values are set by the IBM SSR
during installation. You must have in mind that if you later need to change the LSS definitions
(if a requirement of mixing workloads appears in your installation) then you will need to re-IML
both ESS clusters together.
Cache options for the ESS are 8 GB, 16 GB, 24 GB and 32 GB. Consideration for the
application use of the ESS cache should be factored in when ordering the cache size. In
general larger cache sizes can be very beneficial for S/390 workloads, but are much less
likely to help open systems where the I/O caching is resolved in the host application buffers.
LSS/LCU
For a given host it could be beneficial to use as many LSS/LCUs as possible, from the
configured LSS/LCUs in the ESS (Note: for S/390 ESCON path requirements it may be that 8
rather than 16 LSSs is the recommended configuration). This will enable to spread the logical
volumes of the server across a greater number of LSS/LCUs, and therefore DAs and loops,
thus getting more of the ESS back-end bandwidth.
Spreading the devices of a given server across more LSS/LCUs (using more DAs and loops)
will also make the I/O to be directed across the two ESS clusters. This allows to take
advantage of the bandwidth and processing capacity of both clusters.
If the host is dedicated to a specific application, this dispersion of the I/O could be beneficial
and more easily be implemented and monitored. But when more than one application is
running simultaneously then you may have overlapping I/So from the different
applications/servers and then the benefits of spreading the logical volumes of a given server
may not be so evident. Also you must consider that using more LSS/LCUs for the logical
volumes of a given server will have to be considered when planning your application
implementation of FlashCopy procedures (FlashCopy works with source and target within the
same LSS/LCU).
Data isolation
For performance reasons, contrary to the spreading, you may wish to isolate volumes
containing data with a particular activity pattern or attribute. You can use the ESS design to
your advantage by isolating or grouping together volumes at either the Cluster, target
LSS/LCU (DA), or array/rank level by selecting the disk groups where you allocate the logical
volumes.

Chapter 2. Configuration planning
29
RAID-5
RAID-5 arrays will perform better than JBOD for reads and sequential operations because the
data is striped across multiple disks and more I/O can be done in parallel.
S/390:
The ESS provides additional performance options for S/390 RAID arrays. The
arrays/ranks can be partitioned into either interleaved logical volume mode or non-interleaved
logical volume mode.
Using interleaved mode will give more even performance across all logical volumes, but at the
expense of some additional definition work that needs to be done if you want to utilize the
entire array/rank storage capacity. For example, you will need to define custom volumes in the
non-interleaved partition appended after the interleaved partition.
Custom volumes: S/390
Custom volumes are variable-sized S/390 formatted volumes. They could be used to place a
single critical performance dataset where we would like to have it on its own volume to
minimize contention. Custom volumes can also assist you in circumventing hardware
serialization problems at the volume level, for example reserve/release.
If your operating system does not support PAVs, then custom volumes could be used to
reduce volume contention.
Parallel Access Volumes (PAV): S/390
Parallel Access Volumes provide the ability to almost eliminate IOSQ time, a major contributor
to response time. PAVs will improve the performance of volumes are experiencing large IOSQ
values. In general 1 PAV for each 3390-3 or 3 PAVs for each 3390-9 is recommended.
Larger volumes should use more aliases to keep IOS queueing down, but also allow more
aliases to be defined for the LCU since there will be less base defined devices. Please refer to
Section 8.1.6, “Parallel Access Volumes” on page 110 for detailed considerations.
Workload manager: S/390
Workload Manager (WLM) in GOAL mode provides resources to workloads within a sysplex
to ensure that predefined performance metrics are met. The WLM therefore plays an
important part in the ESS performance by controlling Dynamic PAV and enabling I/O Priority
Queueing.

30
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment

© Copyright IBM Corp. 2002
31
Chapter 3.
Physical installation planning
This chapter discusses various physical considerations involved in planning the physical
installation of a new ESS in your environment. Some of the topics covered are host
attachments, electrical and cooling, floor loading, network, and communication requirements.
You should consult the latest version of this guide for the current information as you plan for
the physical installation of your ESS.
3

32
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
3.1 General considerations
Successful installation of an ESS requires careful planning. The main considerations when
planning for the physical installation of a new ESS are the following:

Interface cables for UNIX, NT, AS/400 and S/390 hosts

Electrical power

Cooling and airflow

Physical placement

Floor loading

Call home

Ethernet connection

Cabling Schemes and distances

Installation planning references
Always look at the most recent information for physical planning in
IBM TotalStorage
Enterprise Storage Server Introduction and Planning Guide,
GC26-7294.
3.2 Host attachments
If you are configuring open systems SCSI/Fibre Channel and S/390 ESCON/FICON logical
disks in your ESS, you need to locate the ESS appropriately, within the various cabling
schemes found in the information you have gathered from the references in the installation
planning section. Please refer to Chapter 1 of the
IBM TotalStorage Enterprise Storage
Server Installation and Planning,
GC26-7294.
3.2.1 Daisy chaining SCSI adapters
We do not recommend daisy chaining SCSI host adapters together. Although it is technically
possible, it has proven to be problematic.
3.2.2 Unique target IDs
Care must be taken to ensure that the target ID of each adapter on the bus is unique.
A single ESS SCSI port may contain up to 15 target IDs. You need to ensure that each target
ID assigned to logical devices attached to that port is unique on that SCSI bus. Use the IBM
TotalStorage Enterprise Storage Server Specialist (ESS Specialist), storage allocation, and
configure SCSI ports panel to display internally assigned target IDs for an SCSI adapter port
in the ESS.
It is preferable to define first the hosts that will access the ESS on the SCSI bus, then the IDs
of all other initiators and non-ESS devices on the bus. On the ESS Specialist, configure SCSI
ports panel, define your host IDs, then enter the other IDs as unrelated hosts or devices,
before adding logical volumes to the port. This will ensure that the ESS assigns
non-conflicting IDs for the logical volumes you add. Refer to “Configure Host Adapter Ports”
on page 247 to see the process involved to do this.
3.2.3 Fibre Channel host attachment
The ESS supports attachment to a Fibre Channel arbitrated loop and Direct native Fibre
Channel connection or through a switch fabric.

Chapter 3. Physical installation planning
33
Direct Fibre Channel connection or through 2109 Fabric switches
To complement the information in this section, please refer to
Implementing Fibre Channel
Attachment on the ESS
, SG24-6113.
3.3 S/390 ESCON/FICON cables
This section discusses various considerations regarding S/390 ESCON/FICON cables.
ESCON cable considerations
ESCON cables may be used to attach the ESS directly to an S/390 host, or to an ESCON
director or channel extender. They may also be used to connect to another ESS, either
directly or via ESCON Director, for Peer-to-Peer Remote Copy (PPRC). The ESS supports all
models of the IBM 9032 ESCON director. It also supports the IBM 9036 channel extender to
the length allowed by the 9036, and the 9729 to a maximum distance of 103 km. Customers
may wish to use methods of extending ESCON channels with which they already have
experience.
ESCON cables come in a standard length of 30 meters, but can be obtained in various
lengths. The maximum length of an ESCON link from the ESS to the host channel port,
ESCON switch, or extender is 3 km. This is using 62.5 micron fiber, or 2 km using 50 micron
fiber.
FICON cable considerations
The ESS Models F10 and F20 support native FICON connection over short-wave or
long-wave fibre links to provide FICON attachment to zSeries or S/390 servers. For the
FICON cable options refer to the chapter 12 in the IBM Redbook,
FICON Native
Implementation and Reference Guide,
GC24-6266.
PPRC considerations
If PPRC is to be used, ESCON connections must be provided between ESS subsystems
containing primary and secondary PPRC volumes. This is the case even if there are no S/390
host connections and no S/390 volumes on either ESS. Any ESCON connections needed to
support PPRC must be considered when planning the ESS configuration. ESCON host
adapters must be installed to support the PPRC links.
For detailed information on PPRC for the ESS, refer to Chapter 12, “ESS Copy Services for
S/390” on page 197, and the corresponding chapter in
IBM TotalStorage Enterprise Storage
Server,
SG24-5465.
TCP/IP Connection for PPRC:
If the ESS Copy Services Web browser interface will be used
to manage PPRC, then TCP/IP connections are needed between each participating ESS
subsystem, and the PC running the Web browser. At each ESS, the LAN connection is
implemented using 100BaseT ethernet.

34
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
3.4 Electrical power
You need to take note of the following considerations concerning electrical power for the ESS.
3.4.1 Dual line cords
All ESS models, and the expansion frames, all have dual line cords. The two line cords to
each ESS frame should be supplied by separate AC power distribution systems. This makes
a simultaneous power interruption to both line cords less likely. The ESS continues to operate
normally if power to one of the line cords is interrupted.
Because each line cord must be able to support the entire ESS frame in the event that power
to the other line cord is interrupted, it is important that the AC supply circuit for each line cord
has sufficient current capacity to supply the entire ESS.
See Chapter 3, "Installation Planning" in the
IBM TotalStorage Enterprise Storage Server
Introduction and Planning Guide
, GC26-7294 for details on power consumption, voltage, and
current requirements for the ESS.
3.4.2 Single phase power
A single phase line cord cannot be plugged into a normal wall outlet. It needs a circuit capable
of supplying 50 or 60 amps (country and location dependent) at a nominal 208 volts or higher,
as shown in Figure 3-1. A wall outlet typically can supply only 20 or 30 amps at 120 volts.
Figure 3-1 Single phase ESS
Note that the maximum data capacities available in the single-phase models of the ESS are
smaller than those of the three phase models. Capacity is limited by the number of disk drives
that can be supplied by the power available from a 50/60 amp single phase supply.
3.4.3 Three phase power
There are no special considerations for three phase power to the ESS.
Phase rotation at the line cord connection is not critical, but where a three phase line cord
connector is provided, IBM recommends that the receptacle be connected to provide
counter-clockwise phase rotation as you view the plug face.
The wall breaker in the circuit supplying the ESS must be capable of withstanding an inrush
current of 100 amps at power on time of the ESS. Inrush current at power on, and operating
current, are higher than normal if the ESS is operated with power on only one of the two line
cords.
Single phase ESS
does not use wall
outlet.
50/60A circuit required.

Chapter 3. Physical installation planning
35
3.4.4 Other considerations
This section discusses considerations regarding power connections and remote control.
Power connectors
The cable connectors supplied with various line cords, and the required receptacles are given
in Chapter 3, "Installation Planning" in the
IBM TotalStorage Enterprise Storage Server
Introduction and Planning Guide
, GC26-7294. The length of the line cords is 14 feet, except in
Chicago, where the length is 6 feet.
Convenience outlet power
The ESS base frame contains a 14-foot multiple-outlet extension cord. The outlets are
mounted within the rack and are provided for use by the service representative to power any
required service tools. You must supply a standard power outlet for the extension cord.
Additionally, three outlets are required for the first installed ESS to supply the ESS Net PC,
display, and ethernet hub.
Remote power control
If you have ordered the Remote Power Control Facility (FC1001), the 2105 can support up to
eight S/360 Remote Power Control (RPC) interfaces to allow remote power control from
S/390 hosts. If you plan to use remote power control, you need to obtain the RPC cables in
the required lengths.
3.5 Physical location and floor loading
In this section, we discuss restrictions on where the ESS can be physically located due to
cable length, floor loading, and service clearance requirements.
3.5.1 Transit
You need to consider the size and weight of the ESS rack in its shipping container, as well as
how it will be moved from the loading dock to the final installation location.
Arbo crate
The weight of a fully configured 2105 frame ranges from 2160 pounds (980 kg) to 2910
pounds (1320 kg). Additionally, each frame is shipped in a wooden Arbo crate that weighs
395 pounds (179 kg) empty.
To check for specification changes and updates on the Web, go to:
http://www.storage.ibm.com/hardsoft/products/ess/essspec.htm
Refer to Table 3-1 for dimensions of the Arbo crate.

36
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Table 3-1 Dimensions and weights for 2105 racks
The 2105 should be protected from any mechanical shocks during transit. It is preferable to
leave the 2105 frame inside the Arbo crate until it is at the final install location, if possible. If
the ESS rack must be moved from the loading dock without the Arbo crate, on its own casters,
take precautions to minimize shocks caused by passing over door gaps in elevators, ramps
and so forth.
IBM recommends that any ramp used to move a 2105 have a maximum gradient of 12.5
degrees, and be no more than 12 feet (3.6 meters) in length.
3.5.2 Physical placement
If you plan to attach the ESS to SCSI hosts, the ESS needs to be located within close
proximity of the SCSI hosts to allow connection within the 25 meter SCSI cable length limit.
The 2105 requires service clearances of 32 inches (81 cm) at the front and 45 inches (114
cm) at the rear. Racks can be placed side by side if floor loading restrictions allow this. If later
installation of an ESS expansion rack is planned, you should allow room for the additional
rack adjacent to the base 2105 to which it will attach. The expansion frame can be installed to
either side of the base frame.
When an expansion rack is attached to a 2105 Model F10 or F20 base frame, a spacer
section a few inches wide is inserted between the two frames. This section serves two
purposes:

To provide a place to loop the excess length in the SSA cables connecting the expansion
rack to the base frame

To keep the floor load within the required limit when both frames are fully configured
Model Dimensions,
inches (cm)
Maximum
Weight (fully
configured),
pounds (kg)
Dimensions of
Arbo Crate,
inches (cm)
Maximum
Weight in
Arbo Crate,
pounds (kg)
2105 Models F10
and F20
Height: 75.25
(179.6)
Width: 54.5
(138.3)
Depth: 35.8 (91)
2160 (980kg) Height: 81.0
(206)
Width: 62.0 (158)
Depth: 41.7 (106)
2985 (1354)
2105 Expansion
frame
Height: 70.7
(179.6)
Width: 54.4
(138.3)
Depth: 35.8 (91)
2590 (1175kg) Height: 81.0
(206)
Width: 62.0 (158)
Depth: 41.7 (106)
3305 (1500)
Note: Height of installed 2105 includes casters and covers, but excludes the top-hat. With top-hat,
height is 75.3 inches (191 cm)
Caution:
A fully configured frame in the crate can weigh up to 3305 pounds (1500
kilograms). This is the approximate weight of a medium-sized car. Ensure that the path for
the 2105 from the loading dock to the final location, including any ramps and elevators, can
accommodate frames of this size and weight. This is a safety issue.

Chapter 3. Physical installation planning
37
The ESS needs a single cable access hole located under the front of the machine at center
for the ESCON and SCSI cables and two line cords. See Chapter 3, "Installation Planning" in
the
IBM TotalStorage Enterprise Storage Server Introduction and Planning Guide
,
GC26-7294, for information on the size and location of cable access holes needed in the floor.
We suggest that the ESS be located over three floor tiles, and that the center tile have
approximately 8" cut off. This creates a hole 24" by 8", which is wide enough to span the cable
restraints at the bottom front of the ESS without weakening the load bearing floor tiles.
3.5.3 Floor loading
The 2105 does not require a raised floor. However, if multiple cable connections will cause a
problem in a non-raised floor environment, it is suggested that a raised floor be used. Ensure
that the ESS is included in the raised floor equipment plan for the site.
A fully configured ESS with an expansion frame is very heavy (5500 pounds, 2495 kg).
Ensure that not only is the building floor loading capacity sufficient for the ESS configuration
you plan to install, but that the raised floor is also rated to support the load. Because of the
high point load on each floor tile supporting part of the weight of the ESS (727 lb, 430 kg for
each of the four casters on a fully configured 2105 expansion frame), the floor tiles and
support structure must be capable of supporting such loads. This is a safety issue. See
Figure 3-2.
Figure 3-2 Caution: Floor loading
It may be necessary to install additional stanchions under the floor tiles that support the ESS.
If installing more than one ESS frame, you may need to separate the frames by a specified
distance to keep within the load capacity of the floor.
See Chapter 3, "Installation Planning" in the
IBM TotalStorage Introduction and Planning
Guide, GC26-7294,
for additional weight and floor loading information.
>980
kg
Floor
Static

38
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
3.6 Cooling and airflow
Adequate cooling is critical to the long term reliability of electronic equipment in general, and
to hard disk drives in particular.
3.6.1 Operating temperature
The ESS should be maintained within an operating temperature range of 20 to 25 degrees
Celsius (68 to 77 degrees Fahrenheit). The optimum temperature is 22 degrees Celsius (72
degrees Fahrenheit). We strongly recommend that you avoid running the ESS, or any disk
storage equipment, at temperatures outside this temperature range.
Humidity
The humidity should be maintained between 40% and 50%. The optimum operating point is
45%.
3.6.2 Airflow
Adequate airflow needs to be maintained to ensure effective cooling. We recommend that two
full-vented floor tiles be located at the front of each 2105 frame, and two vented tiles at the
rear. The ESS takes in cooling air through vents in the front and rear covers, and exits it
through the top of the frame. The hole in the floor provided for cable access to the ESS will
not pass sufficient cooling air. It may be filled by cables, or be blocked with a fire-retardant
pillow.
Adjustments may need to be made to air conditioning equipment or ducting to ensure a good
flow of cool air up through the floor vents. The 2105, like any disk subsystem, works best and
most reliably when the temperature is maintained near the optimum.
Avoid placing racks of 2105 in confined corners of a room where there is insufficient above
floor airflow to remove the heat. See Chapter 3,
Installation Planning
in the IBM TotalStorage
Enterprise Storage Server
Introduction and Planning Guide, G
C26-7294, for information on
required temperature and humidity, and heat output of the 2105 models.
3.7 Call Home and Remote Support
Why is the “call home” feature needed?

To allow the ESS to automatically notify IBM when service is required

To allow remote access and service by IBM support specialists
A remote link capability enables call home support where the IBM TotalStorage Enterprise
Storage Server can initiate a service call if it detects or anticipates a problem occurring. The
remote link also enables service calls into the ESS for remote analysis and potential
correction of problems.
Service alerts describe a problem and the action and/or parts needed to correct the problem.
When a service alert is generated, an e-mail and/or SNMP can be issued to designated
users, and a notification can be sent via the ESSs call home function to an IBM Service Call
Screening Center. Upon notification, service personnel may be dispatched to provide repair
service or remote service actions may be initiated. For the call home function to work, there
must be an analog phone line available for the remote service support.
The Call Home and Remote Service Support functions have been enhanced with the
introduction of the IBM TotalStorage Enterprise Storage Server Master Console.

Chapter 3. Physical installation planning
39
3.7.1 IBM TotalStorage Enterprise Storage Server Master Console
The Remote Support Facility feature code (FC 2715), which included the ESSNet console,
has been superseded by ESS Master Console (FC 2717) for new ESS orders.
The IBM TotalStorage Enterprise Storage Server Master Console (ESS Master Console) is
provided with the first ESS installed in a site. This is a desktop machine preconfigured with
Linux and a Netscape browser. It provides a Web client function and allows connection into
the customer’s LAN. The ESS Master Console is the means for:

The IBM SSR to administer service functions. What previously was done using his Most
(Mobile Solution Terminal) laptop, now can be done from the Master Console.

The IBM SSR to set up the Call Home, Remote Support Facility and other ESS service
related options.

Launch the ESS Specialist. The ESS Master Console includes links to the ESS client user
interface. When you click on one of these links, it initiates the Web interface to ESS
Specialist and ESS Copy Services.
When ordering the first ESS, FC 2717 (ESS Master Console) must be included (Figure 3-3).
This feature provides the hardware to support the ESS configuration, call home, and remote
support capabilities by providing a dedicated console (processor, modem, monitor, keyboard)
and networking components (hub and Ethernet cables and multiport serial adapter). Up to
seven (7) ESS machines are supported per ESS Master Console. Six additional ESS
machines can be connected into the ESS Master Console if FC 2716 (Remote Support
Cables) is ordered with each additional ESS to be connected.
Figure 3-3 ESS Master Console: (FC 2717) — connections
Figure 3-3 shows the ESS Master Console connections and remote support functions. The
same Master Console can be used to launch the ESS Specialist or the ESS Copy Services
interfaces by clicking the
ESS Launcher
icon (see Figure 3-4). Optionally, you can access
ESS Specialist from your workstation if you connect the ESS Master Console into your
intranet by using one of the ports on the ESSNet hub as shown in Figure 3-3.
ESS
Cluster
Bay 2
ESS
Cluster
Bay 1
16 Port Serial Adapter
16 Port Ethernet Hub
Ethernet
Remote
Service
Customer Firewall
Optional
SSR Most
RS 232-S1
Modem
RS 232-S2
RS 232-S3
RS 232-S1
RS 232-S2
RS 232-S3
Master
Console
Call
Home
To Customer Network (optional)
Ethernet

40
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
The desktop machine will display the panel as shown in Figure 3-4 when the ESS Master
Console is installed or when the ESS Master Console icon is pressed from the desktop.
Your IBM SSR will be trained in the use of all the console panels and is the intended user of
the ESS Master Console.
Figure 3-4 ESS Master Console: Main screen
In order to allow remote server support,
Console Launcher
button must be clicked. Then the
IBM SSR will proceed to click the
Remote Services Settings
button as shown in Figure 3-5.
Note:
The panels from the ESS Master console used in this section are not intended to be
all inclusive. We provide these screens for you so that you recognize them as a component
of the ESS Master Console.

Chapter 3. Physical installation planning
41
Figure 3-5 ESS Master Console: Selecting remote server services
You will then be given the panel shown in Figure 3-6 to enable remote services. Here you
must select the use of the default password or supply a custom password.
Figure 3-6 ESS Master Console: Enabling remote services

42
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
3.7.2 ESS Master Console differences
The TotalStorage Enterprise Storage Server Master Console (ESS Master Console) improves
on the previous ESSNet concept. While providing the same overall function, it also adds new
capabilities. ESS Master Console does not use the ESS Ethernet connections and replaces
them with serial port connection to each ESS cluster creating a serial network (see Figure 3-4
on page 40).
It permits the elimination of the modem expander that was used with ESSNet (see
Figure 3-7). In its place is a Multiport Serial Adapter (MSA) which provides performance
improvements giving each ESS the ability to transfer control information independent of other
ESSs that might be connected to the ESS Master Console (see Figure 3-4 on page 40).
Figure 3-7 Call home configuration of the old ESSNet (FC 2715) with modem expander
Table 3-2 shows the differences between the ESS Master Console and the old ESSNet
console.
Table 3-2 Differences between ESS Master Console and ESSNet console
The ESS Master Console implementation offers, among others, the following enhancements
to the remote support function over the previous ESSNet implementation:

Activation of microcode engineering changes (ECs) from the ESS Master Console

Reduction of long-distance telephone costs for call home service

Improved data transmission rates and improved reliability for state saves and traces
Component ESSNet console ESS Master Console
Operating system Microsoft Windows 4.0 Linux
Software • Microsoft Internet Explorer
(MSIE) browser
• Netscape Browser (see note)
• Virus protection software
• Netscape Browser (see
note)
• Terminal emulation
• Virus protection software not
applicable
Communication • Modem
• Modem expander
• Modem
• 16-port serial adapter
Note: The default browser is Netscape 4.7.6. ESS Specialist does not support Netscape 6.0
Modem
Modem
Expander
DTE
Line
Power
Power
Port 2
Port 1Port 16
Enterprise Storage Server
Customer AC service
Telephone line
S3
S3
Cluster 0
Cluster 1
Maximum cable length
10 meters
Maximum cable length
45 meters

Chapter 3. Physical installation planning
43
3.8 Web client and Ethernet connection
The two clusters in the ESS communicate with each other through a 100baseT Ethernet
connection. With the addition of a hub, this Ethernet is used also as a link to the Web client,
and optionally can be connected to the user’s LAN to provide connectivity from anywhere
using TCP/IP.
3.8.1 ESS Specialist
The IBM TotalStorage Enterprise Storage Server Specialist (ESS Specialist) is a software
application that runs in the ESS clusters. It is the interface provided for the user to define and
maintain the configuration of the ESS subsystem. The ESS Specialist can be accessed using
a Web browser running in a Web client PC attached directly to the 2105 via ethernet, or in a
remote machine if the 2105 is connected into the user’s network. For detailed information on
the ESS Specialist, see Chapter 5, “IBM TotalStorage Enterprise Storage Server Specialist”
on page 79.
ESS Specialist Web client requirements
The speed of the Web client used to access the ESS Specialist functions in the ESS will have
a direct effect on the time taken to load the ESS Specialist, and how long it takes to refresh
pages in the browser. This is due to Java code being downloaded from the ESS to the Web
Client. The minimum configuration for the Web client is a Pentium 166 MHz with 32 MB RAM.
For best usability you should use a fast Pentium II machine. We recommend a Pentium II 233
MHz or higher, with 96 MB RAM.
Preferably, the Web client will have a 100BaseT Ethernet adapter installed. With this, only a
simple Ethernet hub and three cables are needed to connect the Web client to both clusters
of the 2105 in a minimum configuration. If the ESS is connected to the user’s LAN, the Web
client can have either token ring or Ethernet, to match the existing LAN.
Web browser requirements
The ESS Specialist Web user interface can be viewed using any browser that fully supports
the Java 1.1 standard. We recommend Netscape Navigator 4.7.6 running on Red Hat Linux
7.0 or NT. This browser can be downloaded free from Netscape’s Web site:
http://home.netscape.com/computing/download/index.html
You can also use one of the following:

Netscape Navigator or Communicator Version 4.06 or later, running on Windows 95,
Windows 98, Windows NT Version 4.0., or Red Hat Linux 7.1.

Microsoft Internet Explorer (MSIE) Version 4.0 or later, running on Windows 95, Windows
98, or Windows NT Version 4.0. MSIE at this level has the necessary support for Java 1.1,
but you may experience the following problem: When running some versions of MSIE, you
may receive error message 1196: "Unencrypted update request" when you attempt to
change the ESS configuration. This is related to a problem with handling secure
transactions when the browser is in a different domain than the ESS (
Note
: the security
restrictions with the MSIE browser have been resolved with the later levels of MSIE and
the G3+4 an later levels of ESS microcode).
For ESS Specialist Web browser interface prerequisites, refer also to 5.2, “ESS Specialist
prerequisites” on page 81.
If you experience a browser hangup or crash in Netscape or MSIE, restart the browser and try
the operation again. These problems are related to the Web browser, not the ESS Specialist
application or Windows operating system.

44
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
3.8.2 Minimal connection to the Web client
The minimum configuration needed to enable use of the Web client uses an Ethernet hub with
a minimum of three ports, and three 100BaseT cables, connected as shown in Figure 3-8.
The internal 10BaseT cable connecting the two clusters together is removed by the IBM
service representative during ESS installation, and a 100BaseT cable is connected between
each cluster and the Ethernet hub. A third cable connects the hub to the Web client. It is
important that both clusters be connected to the Ethernet hub so that connectivity between
them is maintained.
The IP addresses for the two clusters are assigned by the IBM SSR during installation, using
the interface on the service terminal.
The ESS Specialist can be run by specifying the IP address of either cluster directly as the
universal resource locator (URL) in the browser. For example, the IP address of the clusters
might be assigned as 192.168.0.1 for cluster 1, and 192.168.0.2 for cluster 2. You would use
one of these as the URL.
When planning for the Web client connection remember that feature 2717 (IBM TotalStorage
Enterprise Storage Server Master Console) consists of a dedicated console (processor,
modem, monitor, keyboard) and networking components (hub, Ethernet cables and serial
cables). Please refer to Section 3.7.1, “IBM TotalStorage Enterprise Storage Server Master
Console” on page 39 for further description of these hardware components.
Multiple ESSs on the Ethernet private LAN
Additional ESS clusters can be connected to the Ethernet hub, provided that they have
unique TCP/IP addresses. A single Web Client can be used to configure all connected ESS
subsystems.
Ethernet hubs can be daisy chained to connect extra ESS clusters beyond the capacity of a
single hub.
If two or more ESSs in a site participate in PPRC, and this is to be managed using the ESS
Copy Services function, they need to be connected to the same private LAN.
Figure 3-8 Ethernet private LAN configuration
Web Browser
10 BaseT Ethernet Cables
Ethernet Hub
ESS
Optional connection to LAN

Chapter 3. Physical installation planning
45
3.8.3 Connection to Local TCP/IP Network
One or more ESS can be connected to the local IP network by connecting the Ethernet hub to
the network. See Figure 3-8 for details. The user must supply any cables, bridges or hubs
necessary to make the connection.
Why connect to the TCP/IP network?
These are some reasons why you may want to connect to the TCP/IP network:

You may want to use the IBM TotalStorage Enterprise Storage Server Specialist (ESS
Specialist), or IBM StorWatch Enterprise Storage Server Expert (ESS Expert), from a
remote location.

You may want to use e-mail and/or SNMP notification.

Where primary or secondary volumes of PPRC pairs reside in ESS subsystems in remote
locations, you may wish to manage the PPRC using the ESS Copy Services function of
the ESS. This requires that the Web browser used for the ESS Specialist has TCP/IP
connectivity to each local and remote ESS participating in PPRC.
Additional requirements
There are additional requirements if an ESS is to participate in the user’s network:

Two static IP addresses for the clusters must be assigned by the user to be consistent with
the existing network.

A DNS supported host name must be provided for each cluster.

Have the ESS Master Console and both ESS cluster TCP/IP addresses (hostnames)
defined to the DNS server.
Use the Communication Resources Work Sheet at the back of the
IBM TotalStorage
Enterprise Storage Server Introduction and Planning Guide,
GC26-7294, to specify the
TCP/IP address, network mask, default gateway address, hostname, and nameserver TCP/IP
address for the ESS clusters.
The ESS Specialist can now be used from any TCP/IP-connected PC running a suitable
browser. For example, if cluster 0 is assigned the host name essclust0, and the local IP
domain is mydomain.mycompany.com, then the URL to use in the browser is:
http://essclust0.mydomain.mycompany.com
Security
IBM recommends that any ESS connected to a user’s LAN be protected from unauthorized
access by use of a firewall. The ESS Specialist provides password protection and four levels
of user authority to manage access by users. If required, ESS Specialist can be configured to
accept connection only from defined TCP/IP addresses.
3.8.4 ESS Copy Services behind the firewall
This section discusses ESS Copy Services behind the firewall.
Network Configurations and ESS Copy Services
To use ESS Copy Services in a network environment there are some questions you and your
network administrator will have to answer. Each answer will depend on the configuration you
are trying to implement.

46
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
For the purpose of narrowing the possible configurations, we have isolated the possible
configurations to be within the following boundaries. As in any configuration, if you do not
have the expertise available, IBM Global Services or your service provider could assist your
administrators in setting up the right configuration for your environment.
Each ESS Server is located within a private Network along with the CLI Host or Client
Browsers as shown in Figure 3-9.
The Primary and Alternate ESS Servers are separated by a Non-Secured or Public Network
with the CLI Host and Browser located at either end.
The configuration of the sub-area network with Network Address Translation (NAT)/Firewalls
in front of the ESS Server and placements of the CLI Host and Client Browsers.
Figure 3-9 Network view 1
In Figure 3-10, the network connection attaching to the ESS Servers and CLI Hosts is
connected within a customer intranet. This implies that each network is aware of each other
and are routable to each other. The firewalls protecting the intranet are not between network
A and network B. This is important when attaching the ESS server to your network. The
required IP addresses have been obtained and are properly configured on the ESS server
and CLI Host. There are no other required configuration changes to be performed to have the
ESS Specialist console, CLI host and client browsers talk to the ESS servers. Note that the
primary data transmissions and copies are done through the original connections.
Figure 3-10 Network view 2

Chapter 3. Physical installation planning
47
In Figure 3-11, a connection is defined to simulate the attachment of two offices located in
locations remote to each other. The two networks are attached to the Internet and tunneling
protocols have been implemented between the two sites to secure the management,
configuration and processing data being passed between the CLI Host, Client Browser, and
ESS Servers.
Again, because the two networks are unique to each other the same communication
channels are used to pass the management, configuration and processing data as are used
for communicating between the two remote sites. With the Virtual Private Network (VPN)
Tunneling being established by the routers that are connecting the two sites, the data is being
transmitted in a secured environment.
Figure 3-11 Network view 3
When connecting the CLI Host to an ESS Server using various security type procedures,
such as Network Address Translation (NAT) and filtering routers, there are some procedures
that should be followed to allow the passing of traffic from the CLI host to the ESS server. This
may require verifying the configurations and port opens by your network security
administrators.

48
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Here is the list of filters which have been tested with filtering router firewalls that will have to
be opened by the network owner of the router/firewall. Other techniques such as application
firewalls and stateful firewalls could also be used. Follow the filters with: TCP any established,
verifications, and/or stateful filter:

TCP Port 80 - WWW (this means port 80 has to be opened to view your Web browser,
over the World Wide Web)

TCP Port 443 - HTTPS

TCP Port 23 - Telnet

UDP Port 53 - Domain (DNS)

TCP Port 1703 - Nameserver - SOCKS

TCP Port 1705 - appletserver - SOCKS

TCP Port 1707 - applet status server - SOCKS

TCP Port 1709 - client heartbeat - SOCKS

TCP Port 1711 - cobra code - SOCKS/IIOP

TCP Port > 32000 - Heartbeat opens between ESS servers to permit TCP any established
When using NAT you must be careful where you put your CLI host. If it is located as CLI host
A, as shown in Figure 3-12, it will not be able to translate the Web updates correctly. This is
because the translation is done outside the data frames being transmitted to the ESS
Servers. The solution to having the CLI host located outside your firewall with NAT is to use
VPN tunneling techniques.
This can be done in two ways. The first is to simply install the client VPN software supporting
IPSEC Standard for your firewall/NAT device as in example A in Figure 3-13. The second is to
install a device(s) which supports VPN Tunneling to connect CLI Host A to the ESS server as
in example B in Figure 3-13. This would place the CLI Host in the same network. If you moved
CLI Host A to ESS Net A, you would simply apply the filters required for the two ESS servers
to allow them to talk to each other. CLI Host A now would be on the NAT side of the
router/firewall, which would then translate correctly and find the other ESS servers.
Figure 3-12 ESSNet A

Chapter 3. Physical installation planning
49
Figure 3-13 ESSNet A and B
Desk Top On Call (DTOC) or PCanywhere
Another solution, which is less complicated is to install DTOC or PCanywhere on an existing
NT server behind the firewall. The server must have the CPU computing specifications and
Memory allocation to handle ESS Copy Services remotely. By simply accessing this server
remotely and using Netscape browser to connect to the Web interface of the ESS, you can
perform the functions of the ESS Copy Services.
Customer and IBM personnel will use the information you enter in the work sheet during the
installation process. For example, the IBM SSR will enter TCP/IP values and enable the ESS
Specialist during the installation using a terminal connected to one of the serial ports in the
clusters. After this, the ESS Specialist can be accessed over the TCP/IP LAN.
TCP/IP configuration
If a private LAN configuration is to be used, that is, the private LAN comprising the ESS
clusters, Ethernet hub, and Web client only (ESSnet standard configuration), then no
connection will be made to the user’s LAN. See Figure 3-8 on page 44. The TCP/IP
addresses assigned to the clusters and Web Client can have any desired values. No
hostnames, nameserver, or gateway need to be specified. Standard values for the TCP/IP
addresses used in ESSnet are 172.31.1.1 for cluster 1, 172.31.1.2 for cluster 2, 172.31.1.250
for the Web client.

50
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
In many cases, the ESS network will be connected to the customer’s TCP/IP network via a
hub, router or bridge. All the information in the section "TCP/IP Configuration" of the
Communication Resources work sheet needs to be specified (see corresponding appendix in
IBM TotalStorage Enterprise Storage Server Introduction and Planning Guide,
GC26-7294).
Note
: If your LAN environment is something other than 10baseT, you need to ensure that the
hub or switch into which you connect the ESS is capable of operation at 10Mb/s.
ESS Copy Services primary and backup server TCP/IP configuration
If you plan to use the ESS Copy Services function of the ESS to manage PPRC or
FlashCopy, you need to specify the host name and TCP/IP address of the ESS Copy
Services primary server and the backup server. The primary server and backup server are
ESS storage clusters, usually in two different ESS subsystems.
This information is used by the SSR using the service terminal to configure each participating
ESS for copy services. If the Private LAN configuration is used, there will be no DNS
available. The host names must also be added to the HOSTS file by the SSR, using the
service terminal, to allow host name resolution. This needs to be done for each participating
ESS subsystem.
The ESS Copy Services primary and backup server TCP/IP configuration is not included in
the Communication Resources work sheet, it is included in the Communication Resources
work sheets for ESS Copy Services (see corresponding appendixes in
IBM TotalStorage
Enterprise Storage Server Introduction and Planning Guide,
GC26-7294).
ESS Specialist configuration
Users who will be authorized to use the ESS Specialist are specified here, along with the
authority level:

View

Operation

Configuration

Administration
The administrator can define other users, their authority, and passwords using the ESS
Specialist.
Note
: The default initial administrator user ID is storwatch. The password is specialist. At the
first logon to the ESS Specialist, one or more new administrator user IDs and passwords
should be defined. When you define one or more administrator user IDs, the storwatch user
ID is automatically deleted. To prevent unauthorized access, the ESS Specialist will not allow
you to define the storwatch user ID again. This is important because the ESS Specialist can
be used to reconfigure disk arrays and consequently erase customer data. However, if you
delete the last administrator userid, the storwatch userid will be restored.

Chapter 3. Physical installation planning
51
Call home and remote services
In this section of the Communication Resources work sheet you enable or disable these
functions:
Modem incoming calls
IBM recommends that incoming calls be enabled. This
allows diagnosis of problems by a remote specialist.
Modem outgoing calls
Outgoing calls must be enabled to allow the ESS to place a
call to IBM when service is required.
E-mail over LAN
This function would normally be enabled. It allows
specified people to receive e-mail notification of events
such as a modem incoming call being received.
Pager messages
The ESS can send messages to your pager if you enable
pager messages here.
The required information such as pager numbers and remote telephone numbers is specified
in the Communication Resources work sheet.
E-mail configuration
Here you specify the following:

Maximum error notification count per problem
The ESS sends error notifications (0-9) to e-mail addresses and to the call home
destination. Enter the total number of notifications (default is 1) that you want sent to each
recipient for each problem. A number greater than 1 increases opportunities for delivery.
Note: If you enter a value of 0, the ESS does not send an error notification.

E-mail destinations
The ESS sends error notifications and information to the destinations that you enter here if
you have attached your LAN to the ESSNet external hub.
– Enter the full e-mail address for each destination, for example:
maria@host.com
– For each recipient, check one box:
errors
,
information
,
all
, or
none
. This specifies
which notification the recipient receives.
IBM strongly recommends that you list at least one e-mail address. This allows the ESS to
notify you when it receives an incoming modem call from IBM service personnel.
Note:
The e-mail address must be inside the customer’s intranet (behind the firewall).

52
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment

© Copyright IBM Corp. 2002
53
Chapter 4.
Availability
In this chapter, we describe ESS high availability features. For example, in most cases of
internal component failure, recovery by the ESS is automatic and requires no user
intervention. Also discussed are maintenance, error notification, and configuration
considerations to ensure maximum availability.
4

54
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
4.1 ESS internal recovery
This section describes some possible ESS internal recovery situations, their effect on
availability, and the impact of repair. In most cases of internal component failure, recovery by
the ESS is automatic and requires no user intervention.
The main categories of possible ESS component failure and recovery are:

Host connection failures

Cluster failures

Device adapter failures

Disk failures

Power and cooling failures
The ESS is designed to provide uncompromised data integrity in all failure situations. If the
ESS is properly configured, data continues to be available in the event of any single
component failures.
4.1.1 Host adapters
A failure in an ESCON, FICON, FCP or SCSI host adapter may affect one or both of the ports
in the adapter.
Multiple host paths recommended
We recommend that wherever possible, you configure two or more paths from each SCSI,
FCP, ESCON, or FICON host to different host adapter bays in the ESS. This provides
redundancy in the event of a cable failure, or an adapter failure in the host or the ESS. It
enables a failing ESS host adapter or cable to be replaced with no disruption. Additionally,
performance can be significantly improved by configuring multiple physical paths to groups of
heavily used logical disks.
ESCON
For ESCON, there will normally be four or eight paths in a path group between each S/390
host and a S/390 LCU in the ESS. The loss of a single path usually can be tolerated until a
repair can be made, with little impact on performance if four or more paths are configured.
FICON
Plan for a minimum of four FICON channels per ESS. Typical configurations will have eight
FICON channels when the disk capacity reaches 3.4 TB.
As with ESCON adapters, spread the FICON host adapters across all the adapter bays.
Define a minimum of four FICON cards per path group, as reflected in the IOCP or IODEF.
SCSI and FCP
For SCSI and FCP, if the Subsystem Device Driver (SDD), or similar software that supports
multiple paths is running on the host, and two or more paths are configured from the host to
the ESS as recommended, a single host adapter failure will not interrupt access to the LUNs
supported by the failing path(s).

Chapter 4. Availability
55
Reconfigure spare SCSI adapter
If a failing SCSI adapter is the only path configured between a host and any logical volumes,
then data access is interrupted. If a spare SCSI port is present in the ESS, the logical
volumes can be reassigned to this adapter using the IBM TotalStorage Enterprise Storage
Server Specialist (ESS Specialist). The following steps establish a path between the host(s)
and the logical volumes using a new adapter port:
Associate the host(s) with the new port on the Configure Host Adapter Ports panel. Refer to
15.2.5, “Configure Host Adapter Ports” on page 247.
Add the desired logical volumes to the new port. See Section 15.2.7, “Modify Volume
Assignments” on page 256.
The host SCSI cable is moved to the spare adapter port to restore access until a repair is
made. If only one port on the adapter is failing, this needs to be done only for the hosts and
volumes normally accessed through that port. However, repair of the adapter involves
interrupting access through all four adapters in the I/O bay.
CPI cables
A problem in one of the cables interconnecting the storage clusters with the host adapter
bays can prevent communication between a cluster and a host adapter bay. A logic problem
in a HA bay or cluster can have a similar effect. A HA bay must be able to communicate with
both clusters because each write operation involves a transfer of data to NVS in one cluster
and to cache in the other cluster.
A logic failure on a host adapter bay planar board could prevent all four host adapter cards in
that bay from operating. In these situations, the entire HA bay is quiesced and is unavailable.
If the ESS is configured as recommended with multiple physical paths, access to data
continues.
4.1.2 Clusters
Cluster failures can be caused by logic or the internal disk drive used by the cluster to store
microcode and configuration data. If a cluster failure makes the cluster inoperative, or
prevents access from the cluster to disk arrays, or to NVS, the ESS automatically takes
recovery action by transferring the functions performed by the failing cluster to the opposite
cluster. The other cluster takes over the subsystem management and supports access from
the host adapters to all the logical volumes in the ESS.
Cluster failover can be manually induced by the SSR to perform functions such as installation
of microcode updates or to replace components when a failure does not cause automatic
failure, for example the ethernet adapter, diskette drive, or serial ports.

56
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Normal cluster configuration
During normal operation, each cluster manages its own logical volumes, communicating with
the HAs through the CPI hardware, and with the disk arrays through its own DAs. Data is
stored in cache installed in the cluster and write data is stored in the NVS of the opposite
cluster. The CPC (processors and associated logic) in the cluster, together with the
associated NVS in the other cluster comprise a subsystem. Refer to the diagram in
Figure 4-1.
Figure 4-1 ESS cluster configuration for normal operation
Normal Operation
NVS
2
NVS
1
CPC 1
CPC 2
Subsystem BSubsystem A
Subsystem A runs in
CPC 1 and NVS 1.
Subsystem B runs in
CPC 2 and NVS 2.
NVS 1 is physically
in Cluster 2.
NVS 2 is physically
in Cluster 1.
Cluster 1 Cluster 2

Chapter 4. Availability
57
Cluster failover
If a cluster fails in such a way that both cache and NVS in the cluster are not available, a copy
of modified data for both subsystem A and subsystem B still exist in the NVS and cache of the
other cluster. No data is lost.
The failover process is shown in Figure 4-2. In this example, cluster 1 has failed. CPC 2
normally communicates with NVS 2 which is physically located in cluster 1. NVS 2 is no
longer available. Because the only copy of modified data for subsystem B is in the volatile
cache of CPC 2, this data must be destaged quickly to disk.
The sole copy of modified data for subsystem A is in NVS 1. During failover, CPC 2
establishes communication with NVS 1 and copies the modified data into cache. CPC 2 then
takes over the logical volumes normally managed by CPC 1.
Subsystem A and subsystem B are now managed by CPC 2, using the cache in CPC2, and
NVS 1, both physically located in cluster 2. Access to the disk arrays is through the DAs in
CPC 2.
Figure 4-2 ESS cluster failover process
Because one cluster is now doing the work of two clusters, there may be a noticeable effect
on subsystem performance. But read and write access is maintained to all logical devices in
the ESS.
While cluster 1 is offline, it can be serviced while subsystem operation continues.
Failover..
Failure of Cluster 1:
Cache data in CPC 1 and modified
data in NVS 2 is unavailable.
CPC 2 switches to use NVS 1.
CPC 2 takes over functions
of Subsystem A.
High priority destage of Subsystem B
modified data from CPC 2 cache.
Copy Subsystem A modified data
from NVS 1 to CPC 2 cache.
Access to all arrays is through DAs
in Cluster 2.
Subsystem A
and
Subsystem B
NVS
2
NVS
1
CPC 1
CPC 2
Cluster 1
Cluster 2

58
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Cluster failback
When servicing of the failed cluster is complete, the failback process is used to return to the
normal configuration. Figure 4-3 shows the failback process.
Figure 4-3 ESS cluster failback process
The cluster resumes control of its logical subsystems and the associated drives. The host
adapters are instructed to direct IO operations for those LSS to the reactivated cluster.
Cache
Each of the two storage clusters has a high performance cache consisting of 3GB, 4GB, 8GB,
12GB or 16GB of Error Correcting Code (ECC) Synchronous Dynamic RAM (SDRAM). The
ECC provides single-bit, double-bit and multiple-bit error detection.
When data is written to cache, the memory controller calculates check bits and stores them
along with the data. When data is read from cache, the memory controller generates the
check bits again from the data and compares these with the stored check bits. If there is a
difference between the two sets of check bits, an error has occurred.The error may be among
the following:
A single data bit
The error is corrected automatically by hardware using the check
bit information
Two or more data bits
A check condition is generated. The data is automatically
recovered from either NVS (write data), or a disk drive (read
data).
One or more check bits
A check condition is generated. Data recovery is not needed
Errors occurring on cache cards are logged by the cluster. If analysis of cache errors
indicates that a cache card needs to be replaced, the ESS will Call Home.
Failback
NVS
2
NVS
1
CPC 1 CPC 2
Subsystem
A
Re-activate quiesced cluster.
CPC 2 starts using NVS 2.
Transfer Subsystem A to CPC 1.
Copy Subsystem A modified data
from NVS 1 to CPC 1 cache.
Return to normal processing.
Cluster 1
Cluster 2

Chapter 4. Availability
59
Device adapters
At any given time, a disk group (RAID-5 array or JBOD) set is under the control of a single
device adapter; the one corresponding to the LSS or LCU to which the array or JBOD
belongs. The rank is normally accessible only by the cluster containing that DA. If the DA has
a failure, management of its logical subsystems can be transferred to the opposite cluster by
cluster failover.
This results in the host adapters being instructed to direct I/O for the logical devices that are
normally managed by the cluster with the failed DA to the other cluster, and the other cluster
takes control of all disk groups in the SSA loops.
This is physically possible because each SSA loop is attached to both device adapters in a
DA pair, and either device adapter can manage access to all disks on both loops if the other
DA fails.
Non-disruptive repair of the DA can be done using cluster failover and failback in the normal
manner for repair of cluster components.
SSA loop problems
An SSA loop can be affected by a problem in a cable or in the Serial Interface Chip (SIC) in
one of the disk drives, which may prevent transmission of data, commands, and status past a
point in the loop in one or both directions.The device adapters each have one remaining path
around the loop to all disks (apart from the failing disk) on the loop. The failure reduces the
maximum bandwidth of the loop and may have a measurable effect on the performance of the
logical devices located in disks on the loop if they are heavily utilized.
4.1.3 Disk drives
The disk drives of a disk group can be configured as a RAID-5 array or as a set of non-RAID
disks (JBOD). For a disk drive failure, continued access to data depends on whether the
failed drive was configured as part of a RAID-5 array.
Disk drive failure in RAID-5
When a disk drive module (DDM) fails in a RAID-5 array, the DA starts an operation to
reconstruct the data on the failed drive onto one of the hot spare drives on the loop. It does
this by reading the corresponding data and parity in each stripe from the remaining drives in
the array, performing an exclusive-OR operation to recreate the data, then writing this data to
the spare drive.
Effect on performance
While this data reconstruction is going on, the DA can still service read and write requests to
the array from the hosts. There may be some degradation in performance while the sparing
operation is in progress, because the DA and loop resources are being used to do the
reconstruction. Additionally, any read requests for data on the failed drive (1/7 of the total
requests to a 6+P+S array or 1/8 for a 7+P array) require data to be read from the other
drives in the array to reconstruct the data. The remaining requests are satisfied by reading
the drive containing the data in the normal way.
Similarly, write operations are affected if the target for the write is the failed drive. RAID 5
recovery will be necessary for 1/7 of the total requests to a 6+P+S array and 1/8 for a 7+P
array.

60
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
For sequential read and write operations, data from the other drives needed to reconstruct
data for a failing drive may exist in the DA cache, or may be read to satisfy the read or write
request.Therefore few, if any extra read operations are needed for data reconstruction and
the impact of a failing drive for sequential operations is smaller.
Performance of the RAID 5 array returns to normal when the data reconstruction onto the
spare device completes. The time taken for sparing can vary, depending on the workload on
the array, SSA loop, and DA pair.
Drives are hot pluggable
Replacement of the failed drive does not affect operation of the ESS because the drives are
fully hot pluggable. In the event of a disk drive failure, replacement of the failed drive should
be done in a timely manner to ensure that a spare drive continues to be available on the loop.
Predictive Failure Analysis (PFA)
The drives used in the ESS incorporate Predictive Failure Analysis (PFA) and can anticipate
certain forms of failures by keeping internal statistics of read and write errors. If the error rates
exceed predetermined threshold values, the drive will be nominated for replacement.
Because the drive has not yet failed, data can be copied directly to a spare drive. This avoids
using RAID 5 recovery to reconstruct all the data onto the spare drive. The ESS will Call
Home to notify IBM of the need for a repair action.
Disk drive failure in non-RAID
For JBOD disks, the situation is as for any non-RAID protected drive. This means the data on
the drive is unavailable, unless software mirroring is in use to provide an alternate copy on
another device for every logical volume residing on the failed drive.
The broken drive can be replaced without affecting the operation of other disk drives on the
SSA loop. After replacement of the drive, data must be restored to the logical volumes on the
drive from backups using established procedures.
Note:
A disk drive removed from an ESS may still be functional. It may contain confidential
customer data. A drive from a RAID array contains only fragments from any one logical
device. A drive that is configured as a JBOD can contain one or more entire LUNS or several
3390 volumes. You may consider having a procedure in place to manage the secure disposal
of removed disks drives.
RAID-5 recommended for data availability
We recommend that RAID 5 be used because this provides for continued data availability in
the event of a drive failure without needing to provide software mirroring. Software mirroring
requires twice the storage capacity for each logical volume and uses host and ESS resources
for dual write operations. Where high performance is needed for random reads, JBOD may
be needed. For reads and sequential operations, any performance penalty on random write
operations should be mitigated by the large cache provided in each storage cluster, the cache
in the device adapters and the data striping across multiple drives in the RAID 5 array.
4.1.4 Power and cooling
Power for the ESS is designed so that there is no single point of failure that will prevent
access to data.

Chapter 4. Availability
61
AC power supplies
Interruption of AC power to one of the line cords is equivalent to failure of one of the two
primary AC power supplies.
Because each DC power supply for the electronic bays and the disk drive cages is supplied
from both AC power supplies, the failure of an AC power supply or interruption of power to
one line cord leaves all DC power supplies operating.
The electronics bay fans, the disk drive cage fans, and the drive cage power planar fans all
operate on DC from the DC power supplies. The fans continue to operate normally.
Either primary AC power supply can be replaced concurrently. Both clusters continue to
operate normally during AC power supply replacement.
To minimize the effect of any power interruption or disturbance, the two line cords for each
ESS frame should be connected to separate AC distribution systems. The loss of AC power
on one of the line cords will leave the ESS operational using the remaining line cord.
Power interruptions
If power to both ESS line cords is interrupted, the two batteries integrated into the ESS rack
maintain power to the clusters and disk drives for a period sufficient to allow a destage of
write data from cache to the drives, and for an orderly shutdown of the ESS. Because power
interruptions can occur two or more times within a short period, the batteries have sufficient
capacity to support more than one controlled shutdown. The length of time that the batteries
can support the rack depends primarily on the number of disk drives installed. Because the
batteries conform to the 2N philosophy, either battery alone can support an orderly shutdown.
The batteries can be replaced concurrently by the IBM SSR.
Emergency Power Off (EPO) switch
In the case of an emergency, the ESS can be powered off using the Emergency Power Off
(EPO) switch on the front panel. This switch causes immediate cutoff of power to all parts of
the 2105 for safety. No controlled shutdown takes place. Host systems using the ESS will be
affected.
Because no destage of write data can be done, and the internal microcode is not sequenced
down normally, recovery actions may be needed when power is restored to the ESS. The
ESS will perform internal recovery and destage data from NVS to disk when it is brought up
after power is restored. The EPO switch should be used only in an emergency.
DC power supplies
There are three groups of DC power supplies in the ESS base frame:

Three electronic bay DC supplies for cluster 1, and HA bays 1 and 2

Three electronic bay DC supplies for cluster 2, and HA bays 3 and 4

Four storage cage DC supplies for one disk drive cage, or six DC supplies for two disk
drive cages
The DC power supplies in the ESS are designed using an N+1 philosophy. Any single power
supply failure in a group leaves sufficient power capacity remaining to continue uninterrupted
operation. For example, the cages containing the disk drive bays each require two DC power
supplies. For a single disk drive cage, three DC supplies are provided. For two cages, five
supplies are installed. In the event of a DC power supply failure, the disk drives continue to
function normally, powered by the remaining DC supplies. The broken DC power supply can
be replaced with no impact to operation of the ESS.

62
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Cooling fans
Cooling fans in the ESS base frame are categorized into several groups:

Electronics bay fans

Storage cage fans

Storage cage power planar fans

AC supply fans
Each group of electronic bay fans cools an SMP (cluster) processor and associated logic, and
two host adapter bays. The storage cage fans located at the top of the ESS frame cool the
disk drive bays in the front and rear of the disk drive cages. The storage cage power planar
fans cool the DC power supplies for the disk drive cages. The AC supply fans cool the AC
supplies in the lower part of the ESS base frame. Refer to Figure 4-4
In accordance with the N+1 philosophy of the ESS, the failure of any one fan in a group
leaves sufficient cooling capacity for normal operation to continue.
If a fan fails, the remaining fans in the group increase speed to maintain normal airflow.
The cooling fans in each group can be individually replaced without the need to quiesce or
failover any ESS component. Operation of the ESS is unaffected during the repair.
Figure 4-4 Cooling fan groups in the ESS base frame
In addition to these groups of fans, there is a fan in each of the three Electronics cage DC
power supplies in each cluster.
The fans in the electronics cage DC supplies are not associated into a group. If any of these
fans fails, the corresponding power supply is quiesced. The remaining two DC supplies in the
cluster continue to function while the failing power supply is replaced.
Storage cage fans:
front (2), rear (2)
Storage cage power
planar fans (2) on
front side of planar
Electronics bay
DC supply fans (3)
Electronics bay
fans (4)
AC power supply
fans (2) on front
side of AC power
supplies
Cluster 2 Cluster 1
Rear View of ESS

Chapter 4. Availability
63
4.1.5 Ethernet
The Ethernet adapters in the clusters are used by the clusters to communicate with each
other, with a Web client to support ESS Specialist, and to send alerts using e-mail or SNMP. If
the ESS participates in PPRC, and the ESS Copy Services function of the ESS is used to
manage PPRC, ethernet is used for communication between the ESS Copy Services primary
or secondary server and the participating ESS subsystems. An Ethernet network failure can
isolate both clusters from the network or from each other. In either case, operation of both
clusters continues. The connection between the Web client and either or both clusters may no
longer be available. In this case, use of the ESS Specialist may not be possible.
If the network connection between the clusters and the outside network is not operational, the
clusters may log errors if they attempt to use the network and find it nonoperational. Use of
the ESS Specialist from a remote machine will not be possible, but may be OK on a locally
attached Web client. Management of PPRC may be disrupted.
In the event of a network failure that isolates the clusters from each other, some recovery
actions must be performed by the IBM SSR to return the subsystem to the normal state. If
one cluster is isolated from the network (either by Ethernet adapter or cable failure), it will not
be able to send an alert about the problem. At regular intervals, each cluster runs a system
check procedure. When the opposite cluster runs this procedure, it detects that the first
cluster is unreachable and logs the network failure. It sends notification using SNMP, e-mail or
pager with the failure information.
If the problem is in the internal Ethernet adapter, a cluster failover is required to effect a repair
action. After the failback process is complete, normal operation resumes.
4.2 Maintenance
As described in “ESS internal recovery” on page 54, internal failures in the ESS are
recovered automatically by the ESS without user intervention.
4.2.1 Maintenance strategy
The maintenance philosophy of the ESS is based on the following:
First time data error collection:
Data needed for analysis of the problem is collected at
the time of failure and logged.
Non-recreate methodology:
Sufficient information is collected at the time of the failure to
isolate the problem in most cases. The problem does not need to be recreated.
Log and trace data for a problem can be collected remotely by Product Engineering (PE)
support for detailed analysis (See 3.7, “Call Home and Remote Support” on page 38). The
IBM SSR is dispatched by IBM support to repair the problem.
The SSR uses the ESS Master Console (See Section 3.7.1, “IBM TotalStorage Enterprise
Storage Server Master Console” on page 39 for a description) to access the maintenance
procedures. This was previously done using the MoST (Mobile Service Terminal) service
terminal connected to a cluster to access the maintenance procedures (See Figure 4-5). The
problem is opened for service and the required resource is quiesced (for example, a cluster
could be made available for maintenance by failover). The SSR is guided through the repair
process by the maintenance procedures, and the repair validated by running automated
diagnostic tests. If validation is successful, the resource is returned to service, and the
problem is closed.

64
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Figure 4-5 Service terminal functions
The MoST is an IBM Thinkpad running a terminal emulator that will be used by the IBM SSR
if your ESS has not yet been upgraded to use the ESS Master Console feature (FC 2717). It
is used by the IBM SSR to configure communication functions during ESS installation, to run
procedures to upgrade storage or add host adapters, and to run maintenance procedures
when a repair is needed. It is connected to the S2 serial port in either of the clusters.
4.2.2 Upgrades
The 2105 Models E10, E20, F10 and F20 are designed to allow non-disruptive upgrade of
cache capacity and NVS capacity, and addition of device adapters, ESCON, SCSI and Fibre
Channel (SCSI-FCP for Open and FICON) host adapters, and disk drive bays.
The 2105 Models E10 and E20 comes configured with the full complement of 6 GB of cache,
384 MB of NVS, and four DA pairs factory installed. Consequently, the only field upgrades are
the addition of ESCON and SCSI host adapters, and the addition of disk storage capacity.
These field upgrades can be performed non-disruptively. The 2105 Models F10 and F20 have
options of 8 GB, 16 GB, 24 GB and 32GB of cache capacity.
An expansion frame can be non-disruptively attached to a base frame by an SSR under
guidance of the service terminal. The correct sequence must be followed when connecting
SSA cables during addition of disk drive bays, or addition of an expansion frame, to ensure
there is no disruption to operation of existing disk arrays. If only eight LSSs were defined at
installation time, a disruptive IML may be required after addition of an expansion frame to
activate new LSS/LCUs that may need to be configured to support the new storage.
Important:
The functions of the MoST Service Terminal can now be performed using the
ESS Master Console. See Section 3.7.1, “IBM TotalStorage Enterprise Storage Server
Master Console” on page 39.
MoST Service Terminal
Checkout machine
Run FRU replacement
Upgrade procedures for disks, HAs
Configure Call Home
Enable Remote Support
Configure TCP/IP for ESS Specialist

Chapter 4. Availability
65
4.2.3 Non-disruptive service actions
When maintenance is required, virtually all components in the ESS can be serviced
concurrently with continued operation, provided that the ESS is correctly configured.
Microcode updates
Microcode in the storage clusters can be updated concurrently by using the failover
procedure to make each cluster available in turn, then failing back after the new microcode
has been activated. Microcode is loaded from a CD ROM drive located in each cluster.
Configuration information can be copied between clusters on diskette.
Host adapters
ESCON, SCSI, and FCP (FICON and Open Fibre Channel) host adapters can be replaced
without using cluster failover. Servicing of a host adapter card requires that the adapter bay
containing up to four HA cards (four to eight ports) is quiesced. All the ports, ESCON, FICON,
SCSI or FCP, in the bay will be unavailable during the repair.
Although a storage cluster and two host adapter bays are supplied by a common group of
three DC power supplies in an N+1 configuration, DC can be independently switched off to
the cluster, or either HA bay under control of the service terminal. For maintenance on a HA
bay, DC power to the bay is switched off, and the bay is withdrawn to allow replacement of the
host adapter or planar board.
Nondisruptive service on a host adapter bay depends on correct configuration of the ESCON,
FICON, SCSI, FCP paths from each host to the ESS. There must be at least one alternative
path available through the other HA bays for each ESCON or SCSI port used in the bay being
repaired.
CPI cables
The common parts interconnect (CPI) cables connect the host adapter bays to the I/O
adapter cards in the storage clusters. Each HA bay has two CPI cables- one to each cluster.
The CPI cables can be replaced non disruptively, provided that the HA bay is quiesced.
Storage cluster components
Any component in a storage cluster can be replaced by using the failover/failback procedures
to make the cluster available for service. Cluster components include the SMP processors,
memory (cache), device adapters, NVS, I/O adapter cards, the service processor card, and
the planars on which these components are mounted. The associated CD-ROM, hard drive,
and diskette drive are also components in the cluster.
While the cluster is offline, and powered off for service, all host adapters remain operational.
Cluster hard drive replacement
Each cluster is responsible for loading microcode into its SMP processors during initialization,
monitoring and logging errors, and notification using the methods configured (for example,
e-mail, SNMP, pager, Call Home). The clusters also support recovery from internal failures,
and provide the maintenance facility used to make repairs to the ESS. These functions are
supported by a dedicated internal hard disk drive. The data residing on the drive falls into
three categories:

ESS microcode, or licensed internal code (LIC)

ESS customization data (for example, communications parameters)

Configuration information (hosts, ports, and logical volumes)

66
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
A full copy of all of this data is stored in each cluster. In the case of internal disk failure in one
cluster, the IBM SSR replaces the failed disk, then reloads the LIC from the original CD-ROM.
Configuration and customization data are saved on diskettes from the other cluster and
restored to the replacement hard disk. The cluster is then returned to service using the
failback process. The repair is concurrent with continued operation of the ESS.
Device adapters
Replacement of a failed DA requires that cluster failover has occurred to make the cluster
available for repair. This should happen automatically if a DA fails. Replacement of the DA
does not disrupt operation of the two associated SSA loops. After the DA card has been
replaced, the cluster will be powered on, tested and restored to service using failback.
SSA cables
The SSA cables connect the disk drive bays to the DAs. An SSA cable can be replaced
nondisruptivesly because the DAs reconfigure the loop to work without it. During the repair,
each DA has access to each drive on the loop through only one path around the loop instead
of the normal two paths.
Hard disk drives
Any one of the SSA disk drives can be hot plugged by exchanging the failed drive for a new
one. The correct maintenance procedure must be followed by the IBM SSR using the mobile
service terminal to change the drive, test the new drive and initiate a format operation to
change the format in the new drive from the conventional 512 byte sector SCSI format to the
524 byte sector format used in the ESS. The new drive becomes a spare.
Power components
Power components fall into three main categories:

Primary components supplying the whole rack, such as AC power supplies, line cords and
rack batteries

Secondary components supplying DC power to sections of the machine

Power control and sensing components
In general, primary components are 2N, secondary components are N+1, power control is
2N.
Primary power supplies and rack batteries
The primary power supplies, the line cords, and the two rack batteries can be replaced while
the ESS is operating. Both clusters continue to operate normally during replacement of these
components.
Electronics bay DC power supplies
These components are N+1, as described in Section, “DC power supplies” on page 61.
Three DC supplies are provided for each storage cluster. Each of the three power supplies
has an integrated cooling fan. If one of the DC supplies fails, the cluster and host adapter
bays continue to operate normally, powered by the remaining two DC supplies. The failing
power supply can be replaced nondisruptively without the need for cluster failover.
Disk drive cage DC power supplies
These components are N+1, and, like the electronics bay DC supplies are hot pluggable.

Chapter 4. Availability
67
Power control and sensing components
There are two rack power control (RPC) cards in the ESS. The RPC cards, the electronics
cage sense cards, and the storage cage fan and power sense cards can be replaced with the
rack power on.
Cooling fans
These components are N+1, as described in Section, “Cooling fans” on page 62. They can
be replaced individually, except for the fans in the electronics bay DC power supplies which
require replacement of the power supply.
4.2.4 Disruptive service actions
Replacement of the planar board for the disk drive cage DC power supplies requires that the
supplies be removed, and, therefore, cannot be completed nondisruptively.
Replacement of the UEPO operator panel card requires the ESS subsystem to be powered
off.
The probability of failure of these components is extremely small.
4.3 Error notification
The ESS storage clusters continuously monitor the operation of the ESS hardware and
microcode. If a failure is detected, or a situation occurs that requires notification to the
customer or IBM, it can be reported to system administrators or to IBM in a number of ways:

Call home

Simple network management protocol (SNMP)

E-mail

Pager

Service information message (SIM) - S/390 only

Environmental recording and editing program (EREP) - S/390 only
Figure 4-6 shows the methods available for communication with ESS internal functions. Error
notification uses a serial port and the Call Home modem to notify IBM. The Call Home
modem is also used to call a pager if this function is configured.
The IBM SSR uses a second serial port on one of the storage clusters to connect the service
terminal. This is used to configure the ESS during installation, and to perform maintenance
procedures.
The Ethernet connection is used for SNMP and e-mail notification. It is also the means by
which a local or remote Web client can be used to configure and monitor the ESS, using IBM
TotalStorage Enterprise Storage Server Specialist (ESS Specialist) and IBM StorWatch
Enterprise Storage Server Expert (ESS Expert).

68
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Figure 4-6 ESS communications
Any or all of these options can be used. Apart from SIM and EREP, they must be enabled
through the service terminal used by the IBM SSR during installation. For SNMP, e-mail, and
pager, the necessary information, such as e-mail addresses, is configured through the ESS
Specialist interface.
4.3.1 Call Home feature
This feature of the ESS enables it to dial the IBM Support Center directly in the event that
maintenance is required. The process is automatic; no customer intervention is needed. Call
Home must be enabled using the ESS Specialist. The Call Home function dials into a catcher
PC in the IBM support center. The PC connects to the IBM RETAIN system to create a
problem record. This record contains information on the failure and allows the support center
to analyze the problem.The problem record is used by IBM to manage and resolves
problems.
If required, the Support Center can dial into the ESS using remote support to collect
additional information.
Call home setup is done by the IBM SSR during installation of the ESS.
Remote support
If the remote support function is enabled, a remote IBM specialist can initiate a call into the
ESS and download log information, microcode traces, and other data for analysis to aid in the
diagnosis of problems in the ESS.
Remote support makes use of the same modem and serial ports in the storage clusters that
are used for Call Home (note that remote access does not compromise the security of your
data stored in the ESS.)
Enterprise
Storage
Server
Remote
Services
Local
Terminal
Web
Server
Serial
Ports
Ethernet
TCP/IP
SNMP
E-mail
IBM CE with Mobile Service
Terminal (MoST)
Product
Engineer
Catcher
System
Support
Center
Call
Home
Whoever
needs
to know
Pager
Storage Manager

Chapter 4. Availability
69
To enable remote support, you must use the ESS Specialist before the ESS will accept
incoming calls. Two levels of support can access the ESS through the cluster serial ports. The
first level, System Support Representative (SSR) support, gives limited access to SMIT. The
second level, Product Engineering (PE) support, needs root access to the operating system
in the storage clusters to enable diagnosis of problems. This level of access must be
specifically enabled by you, onsite, and it automatically disables after one week.
Neither support level gives access to your data. If the ESS accepts an incoming remote
support call, you are notified by e-mail.
4.3.2 Simple network management protocol (SNMP)
The ESS generates SNMP traps and supports a read-only management information base
(MIB). The ESS generates both generic and product-specific SNMP traps. Product-specific
traps provide information on problems that are detected by the ESS and require corrective
action. The ESS sends information to the trap addresses set by the SSR during the
installation procedure.
The ESS supports the usual generic traps, such as cold start and warm start. Product-specific
traps provide information on problems detected by the ESS that require action by the user or
by IBM. Product-specific traps contain the identifier assigned by the ESS for the problem that
caused the trap. The character string "Problemid=N" (where N is the problem identifier) is in
the trap’s description field. You can use this to find additional details on the problem using an
MIB browser.
Using the MIB browser to view problem details
To use the MIB browser to view problem details, proceed as follows:
1.Point the MIB browser to the problem that created the trap.
2.Look within the ibm2100 mib, ssProblem, ssProblemid. A display of this MIB variable is in
the form "Index:Problemid".
3.Using the problem identifier (ID) from the trap, determine the index value with which it is
associated.
4.Use the index value to determine which specific data is associated with a problem, and
look at other MIB variables within ssProblem.
4.3.3 E-mail
The ESS uses standard TCP/IP e-mail to send error reports and notifications to the e-mail
addresses you define. You can specify four levels of notification for each e-mail recipient, and
the number of times that notification is sent for each problem. This parameter also defines the
number of times the ESS will Call Home to IBM for each problem.
The ESS generates e-mail messages in two categories:

Information

Errors
Informational messages
These are some examples of informational messages:

A new level of Licensed Internal Code (LIC) has been installed.

New hardware has been installed.

A remote support specialist has dialled into the ESS.

70
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment

The SSR has run the customer-notification diagnostic test. This test verifies that e-mail
messages are being received at the addresses set up by the SSR or by the customer,
using ESS Specialist.
Error messages
The ESS sends error messages when it detects a situation that requires action by the
customer or IBM. The error messages typically contain the following fields:

Product manufacturer ID and date.

Rack location — The rack location is entered by the installer during the initial installation of
the product.

Product machine type and model number (assigned by IBM).

Product serial number (assigned by IBM).

Customer voice phone number.

LIC level of local storage server.

LIC level of remote storage server.

Report time/date stamp.

Problem ID. — The problem ID assigned to this problem by the storage facility. This
problem ID can be used to access detailed problem information.

Exception Symptom Code (ESC) — A detailed error code used to define a problem. It is
used by the IBM SSR to enter the maintenance procedures.

SRN (Reference Number) — A detailed code used by the IBM SSR.

Problem Status. See “Status” on page 71 for a list of the problem states.

Description — A description of the problem.

Additional message — Any additional information that is available.

Failing cluster — The cluster on which the failure occurred (1 or 2).

Reporting cluster — The cluster which reported the failure (1 or 2).

Failing resource — This coded resource name is used during the repair process by the
IBM SSR.

Failure occurred — Date and time when the failure first occurred.

Last occurrence — Date and time of the last recorded occurrence.

Failure count — The number of times this failure occurred.

Presentation interval — The time between successive reports of this problem.

Remaining presentations — The number of additional times this report will be sent.

Isolation procedure — A pointer to a special procedure in the service guide.

Failure actions — Actions the service provider should take.

Probable cause — Information for the service provider.

Failure cause — Information for the service provider.
The information above is also sent to IBM using the Call Home facility, if this has been
configured.

Chapter 4. Availability
71
Following are the fields that are most useful in identifying DDM failures:

Description

Failing cluster

Reporting cluster

Failing resource

Last occurrence
Status
A problem can exist in various states. Each state represents a state in the reporting or repair
of a problem. The problem states are defined as follows:
Pending
The Initial problem state. The problem will be reported via one of these
methods (SNMP trap or e-mail).
Received
This state is used to indicate that the notification has been received by a
host, either via e-mail or SNMP trap.
Open
A repair process has begun. The required resources have been removed
from use. Suspending a repair leaves a problem in the OPEN state.
Closed
A repair process has been completed. All resources have been successfully
returned to use.
Canceled
A service representative has chosen to cancel this problem.
Expired
Once a problem has been in the pending state for 30 days, it is changed to
this state.
Archived
Problems that have been closed, expired or canceled for more than 30 days
are archived.
4.3.4 Pager
You can optionally configure the ESS to send problem information to a pager using the Call
Home modem.
The Call Home, SNMP, e-mail and pager functions are configured using information you
provide on the Communication Resources Worksheet. For information on completing the
work sheet, refer to Appendix A on the Part, “Completing the Communication Resources
worksheets” on page 295.
4.3.5 S/390 notification
The two methods of error notification for S/390 systems are console messages, and by
logging errors for display by EREP.
Service information message (SIM)
In the S/390 environment, the ESS uses one of the channel paths to notify a S/390 host of
certain internal conditions. Sense data is logged by the host. Conditions in the ESS that need
to be brought to the attention of system operators usually result in the presentation of a
service information message (SIM) on a system console. This message identifies the ESS
and provides information about the error and the impact of repair.
For OS/390, the SIM is presented in message IEA480E. This message is sent to the system
that presents the next active I/O. This system could be any system in the complex and not
necessarily the production system.

72
Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
The ESS sends SIMs to System/390 host consoles for three types of SIMs:
DASD SIM
This SIM tracks disk drive module (DDM) failures and problems
Media SIM
This SIM tracks data check problems on the media.
Storage facility SIM
This SIM tracks control unit, power control, and other hardware
problems.
EREP
In S/390 systems, information about hardware errors and software errors and statistical
information is written to a data set on a system disk. For OS/390, the data set is called
LOGREC. The environmental recording and editing program (EREP) is used to select desired
types of entry in the data set, edit and summarize them. In particular, the event report lists
events in chronological order, and can be useful when analyzing problems such as missing
interrupts.
DEVSERV (DS) command
The DEVSERV command can be used to obtain information about a disk subsystem attached
to an S/390 host using the OS/390 operating system. See Section 11.5.1, “DEVSERV
command” on page 188.
4.4 Configuration for availability
Virtually all forms of internal failure in the ESS are recovered automatically by the ESS
without intervention by the user. This is achieved using mechanisms such as RAID 5, cluster
failover, and multiple redundant paths between the storage clusters, disk arrays and host
adapters.
Configuring for maximum availability of the ESS and access to data requires attention to
these considerations:

Maintaining connectivity between each host and the ESS in the event of a failure in a host
attachment path and during repair of the failing component

Using RAID 5 or host software mirroring to ensure continued access to data in the event of
a disk drive failure in the ESS

Ensuring the reliability of the power supply and cooling to the ESS racks
4.4.1 Maintaining connectivity
ESCON and FICON path failures include problems with host channels, problems in ESCON
or FICON cables, patch panels, extenders or switches, and faults in ESS ESCON or FICON
host adapters. SCSI and FCP path failures in open systems can be caused by problems in
host SCSI initiators, other devices on a SCSI interface, or in physical cabling or fibre.
Problems may be caused by incorrect configuration such as excessive cable length or
incorrect termination.
To maintain connectivity in the event of a path failure, at least one alternate physical path
must be provided from the host to the logical device in the ESS. An FB logical device in the
ESS can be logically assigned to two or more SCSI ports to enable sharing between hosts, or
to provide multiple physical paths to a single host.

Chapter 4. Availability
73
Subsystem device driver (SDD)
For UNIX, Windows NT, or AS/400 environments, multiple SCSI and/or FCP paths to a device
must be supported by software. IBM Subsystem Device Driver (SDD) is a software product
that provides availability and I/O load balancing for the pSeries, Intel PC based servers, HP,
Compaq and Sun environment. Other platforms may be supported by other non-IBM
products. Please refer to Section 13.4, “Subsystem Device Driver (SDD)” on page 208 for
support information.
SDD provides availability through automatic I/O path failover. If a failure occurs in the data
path between the host and the ESS, SDD automatically switches the I/O to another path.
SDD will also move the failed path back online after a repair is made.
SDD improves performance by sharing I/O operations to a common disk over multiple active
paths to distribute and balance the I/O workload. Refer to Figure 4-7.
Figure 4-7 Function of subsystem device driver (SDD)
Shared LUNs are supported by SSD
SDD and HACMP can be run in an environment where more than one host is attached to the
same LUN (a multi-host environment). This includes clustered hosts such as RS/6000
servers running HACMP and NT High Availability Clusters.
Supported systems
For a list of supported systems, please refer to Table 13-1 on page 207.
Host system
SCSI
Adapter
SCSI
Adapter
Logical Volume
Manager
Application
SCSI Adapter
Device Driver
ESS logical volume
Data Path Optimiser
SCSI Adapter
Device Driver

74

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Windows NT
SSD for Windows NT is installed using the install shield. It is self configuring. The Subsystem
Device Driver operates as a filter mechanism. Other paths to the drive appear offline, but they
are used in rotation.
AIX
SMIT is used to install and configure Subsystem Device Driver.
Conversion scripts replace hdisk devices by vpath devices for volume groups. Affected file
systems must be unmounted. The hdisk devices are still online, but to use the SDD functions,
the vpath devices must be used.
Datapath commands
For AIX, a command line path recovery command is provided. It allows you to query devices
and adapters. You can use the datapath command to vary paths online or offline. By simply
typing in the command
datapath
you may view the options available to the command.
Following are some examples:
datapath query adapter/device [n]
datapath set adapter <n> online|offline
datapath set device <n> path <m> online|offline
Trace function
In an environment with more than one path to a drive, errors could occur on a single path,
while others are working error free. SSD supports driver traces to assist in resolution of single
path and intermittent problems that otherwise can be difficult to isolate.
Refer also to the
IBM Subsystem Device Driver Installation and Users Guide
. You can obtain
a copy on the Web at:
http://ssddom02.storage.ibm.com/disk/ess/related.html
Once there, click the related PDF file.
We recommend that SDD, or a program that provides equivalent function in the host
operating system (for example, Veritas Volume Manager for the SUN Solaris OS) be used to
enable the configuration of two or more physical paths from each SCSI host to logical devices
in the ESS. The iSeries operating system has the needed multiple path support included.
Multiple SCSI paths from each host should be distributed to different adapter bays when
possible. This minimizes the impact of failure in host SCSI adapters, cables and ESS SCSI
adapters, and allows for nondisruptive repair of ESS SCSI adapters.
ESCON
In the S/390 environment, normal practice is to provide multiple paths from each host to a
disk subsystem. Typically, four paths are installed. The channels in each host that can access
each Logical Control Unit (LCU) in the ESS are defined in the HCD (or IOCDS) for that host.
Dynamic Path Selection (DPS) allows the channel subsystem to select any available
(non-busy) path to initiate an operation to the disk subsystem. Dynamic Path Reconnect
(DPR) allows the ESS to select any available path to a host to reconnect and resume a
disconnected operation, for example to transfer data after disconnection due to a cache miss.
These functions are part of the S/390 architecture and are managed by the channel
subsystem in the host and the ESS.

Chapter 4. Availability
75
A physical ESCON path is established when the ESS port sees light on the ESCON fiber (for
example, a cable is plugged in to an ESS host adapter, a processor or the ESS is powered
on, or a path is configured online by OS/390). At this time, logical paths are established
through the ESCON port between the host and some or all of the LCUs in the ESS, controlled
by the HCD definition for that host. This happens for each physical path between a host CEC
and the ESS. There may be multiple system images in a CEC. Logical paths are established
for each system image. The ESS then knows which ESCON paths can be used to
communicate between each LCU and each host.
At the host operating system level (for example, OS/390), the paths to be available for use for
each logical device are defined to the ESS at the time the device is brought online.
ESCON directors
Because a large number of hosts may be connected to the ESS, each using multiple paths,
the maximum 32 ESCON adapter ports that can be installed in the ESS may not be sufficient
to accommodate all the connections. Where SCSI adapters are installed, the number of
ESCON adapter ports will be fewer than 32. The solution to this problem is the use of IBM
9032 ESCON Directors to switch logical connections from multiple host channels to a single
physical port connected to the disk subsystem.
A logic or power failure in an ESCON Director can interrupt communication between hosts
and the ESS. We recommend that more than one ESCON Director be provided to ensure
continued availability. For example, four of the eight ESCON channels in a path group could
be configured to go through each of two ESCON Directors as shown in Figure 4-8. The
complete failure of either ESCON Director leaves half of the ESCON paths still operating. In a
large installation, there may be two eight-path ESCON groups connected to an ESS. Four
directors might be installed, each carrying two paths in an eight-path group.
Figure 4-8 Multiple ESCON switch configuration
Host A Host B Host C Host D
Escon
Director 1
Escon
Director 2
Enterprise Storage Server
If Escon Director 1
fails, 4 paths remain
from all hosts to
the ESS.

76

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
For availability, and to provide sufficient data bandwidth to provide maximum performance, we
recommend that each MVS host has eight ESCON paths to all LCUs it will access.
Assuming that the ESS has ESCON adapters in each HA bay, you should configure two paths
of each eight path group to the two ports in one ESCON adapter per bay. Refer to Figure 4-9.
The failure of one ESCON adapter card removes two of the eight paths. This is equivalent to
losing one path of four, and the impact should not be significant unless channel utilization is
very high.
Figure 4-9 Physical configuration of ESCON paths
This configuration is recommended for best performance. Refer to “Host adapters” on
page 27 for guidelines on configuring ESCON channels for performance.
4.4.2 Access to data
If a disk drive in the ESS fails, data continues to be available if the drive is a member of a
RAID-5 array. A spare drive is reconstructed into the array and performance returns to normal
after the reconstruction is complete.
If a drive configured as a JBOD fails, data on that drive is not available. There may be more
than one logical disk on the drive. None of these logical disks will be available. Where JBODs
are in use, continuous availability can be provided by host software mirroring. For example,
the AIX Logical Volume Manager (LVM) can be configured to mirror selected volumes. The
combination of JBOD and software mirroring provides excellent random write performance
with continuous data availability, at the expense of a doubling in required storage capacity per
volume and increased usage of resources such as SCSI paths.
IBM recommends that RAID-5 be used in preference to JBOD, except where performance on
random reads is critical.
Remote copy
Techniques such as RAID-5 and software mirroring can provide data availability in the event
of a single disk drive failure in the ESS. However, if the ESS loses power, or receives physical
damage (for example, fire or water damage), that prevents its continued operation, data
availability depends on backups of the data having been made and safely stored, perhaps
offsite. An alternate site may be needed to continue processing.
Path group
Path group
HA Bay
ESCON
SCSI
SCSI
ESCON
HA Bay
ESCON
SCSI
SCSI
ESCON
HA Bay
ESCON
SCSI
SCSI
ESCON
HA Bay
ESCON
SCSI
SCSI
ESCON

Chapter 4. Availability
77
Peer-to-Peer Remote Copy (PPRC)
Peer-to-Peer Remote Copy (PPRC) provides a means to maintain a synchronous copy of
selected volumes on another ESS located in the same site, or elsewhere up to 103 kilometers
away (using channel extenders). PPRC can be used for both S/390 volumes and FB volumes.
PPRC for FB and CKD volumes can be managed using the ESS Copy Services function of
the ESS. Additionally, for S/390 volumes only, PPRC can be controlled from a S/390 host
using commands similar to those used for PPRC on IBM 3990 and IBM RAMAC storage
subsystems. A feature code must be ordered to enable PPRC on the ESS.
Extended Remote Copy (XRC)
Extended Remote Copy (XRC) can be used in the z/OS environment to create and maintain
asynchronous copies of ESS CKD volumes. The ESS can have primary or secondary
volumes of XRC pairs. The primary and secondary volumes in an XRC pair can reside in
different types of disk subsystem. For example, a primary device might be in an ESS while the
secondary is in a 3990-6/RAMAC subsystem up to several thousand kilometers away.
If the ESS is to act as a primary control unit, the appropriate feature code must be installed.
No feature code is needed for an ESS used only for secondary volumes of XRC pairs.
The
IBM TotalStorage Enterprise Storage Server,
SG24-5465 contains detailed information
on PPRC and XRC for the ESS.
4.4.3 Power and cooling
The ESS can continue to function only if power is available on at least one of the two line
cords on each rack. The internal batteries provide power for a sufficient time to override
power disturbances, and to destage write data and sequence down the ESS in the event of
total power loss. The batteries cannot support continued operation without AC power.
The two line cords should be connected to independent AC distribution systems to minimize
the possibility of losing power to both line cords simultaneously.
Although the ESS may continue to operate at temperatures outside the extreme operating
range of 16 to 32 degrees Celsius (60 to 90 degrees Fahrenheit), continued operation at
these temperatures may affect the future reliability of internal components, especially disk
drives. If a failure in air conditioning equipment results in an ambient temperature outside this
range, the ESS should be powered off until recommended operating conditions are restored.

78

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment

© Copyright IBM Corp. 2002
79
Chapter 5.
IBM TotalStorage Enterprise
Storage Server Specialist
In this chapter we describe the IBM TotalStorage Enterprise Storage Server Specialist (ESS
Specialist), and the Web interface for the ESS. This is the configuration and administration
interface of the ESS. It is the most important interface between the hardware, the software,
and the storage administrators.
5

80

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
5.1 ESS Specialist
The ESS includes the ESS Specialist, a network enabled management tool that allows the
storage administrator to monitor and manage storage from the IBM TotalStorage Enterprise
Storage Server Master Console (ESS Master Console), or from a remote workstation using a
Web browser.
Using a secure Internet connection (LAN with a Web browser), such as Netscape Navigator,
or Microsoft Internet Explorer, your storage administrator can coordinate the consolidation
effort and easily integrate storage capacity into the ESS.
The ESS Specialist provides you with the ability to do the following:

Monitor error logs — If a problem occurs, a description of the problem including the failed
component, the problem severity, and who is to be automatically notified is described.

View the ESS status — Logical schematic of the ESS environment including the host
attached ports, controller and cache storage, device adapters, devices and host icons may
be checked.

View and update the configuration — A color schemed view of the storage, including the
amount of space allocated and assigned to one or more hosts, space allocated and not
yet assigned, and space not allocated to logical volumes may be viewed.

Add host systems or delete host systems.

Configure host ports.

Add volumes, remove volumes, and reassign volumes between different servers.
Volumes can be reassigned between hosts as follows:
– Removing volumes (or unassigning volumes from hosts). Volumes can be removed by
removing all logically attached host connections to the logical volume.
– Adding volumes. Volumes can be added from subsystem capacity that has never been
defined or after an array has been reinitialized.
– Reclaiming previously defined logical volumes.

View communication resource settings, such as TCP/IP configuration and users.

View cluster LIC levels.
You can view the active level, next level yet to be activated, and the previous level.

Select one of the following authorization levels for each user:
– Viewer. A viewer can view the current configuration and status information.
– Operator. An operator can perform view and operation functions, such as changing the
remote service and PE password
– Configurator. A configurator can view the current configuration and status information
and can make changes to the configuration.
– Administrator. An administrator can define new user IDs, delete old IDs, and assign,
change or revoke passwords and levels of authorization.

Web support for ESS Copy Services (PPRC and FlashCopy).
The remainder of the chapter will give you general topics of the ESS Specialist. For ESS
Specialist as it relates to S/390, refer to Chapter 8, “CKD storage configuration” on page 101.
For ESS Specialist as it relates to fixed block storage, refer to Chapter 15, “ESS configuration
for open systems fixed block storage” on page 235.
For detailed information of how to use the ESS Specialist, refer to
IBM TotalStorage
Enterprise Storge Server Web Interface User’s Guide,
SC26-7346.

Chapter 5. IBM TotalStorage Enterprise Storage Server Specialist
81
5.2 ESS Specialist prerequisites
The ESS Specialist interface (and the ESS Copy Services interface) is a set of Java applets,
which are programs that are dynamically loaded by the browser, and which execute within
your browser. When you request a change to the configuration, the Java applets
communicate with the microcode running on the ESS clusters to retrieve the current
configuration data, submit the requested configuration change, and display the outcome of
the request.
The ability to run programs (applets) inside a Web browser is provided by Java technology.
You must use a browser that contains the proper Java Virtual Machine (JVM) implementation
to support these applets. The browser software provided by different companies, and even
different versions of the same browser, vary widely with respect to their JVM support.
Consequently, not all browsers are capable of supporting the ESS Specialist or ESS Copy
Services.
The ESS Web interfaces support both the Netscape Navigator and the Microsoft Internet
Explorer (MSIE) versions listed in Table 5-1.
Table 5-1 Web browsers supported by ESS Web interfaces
The minimum recommended browser hardware is a 166 MHz processor and 32 MB memory,
however, greatly improved performance can be achieved with larger configurations (233MHz
and 128 MB RAM).
The ESS Specialist is not enabled by default. At installation time, the IBM SSR will assign the
customer specified IP address and hostname alias to each cluster controller within the ESS.
Once complete, other setup functions can be performed by using the ESS Specialist.
Netscape level (See Note 1) MSIE level (See Notes 2, 3, 4)
Netscape 4.04 with JDK 1.1 fixpack MSIE 4.x with Microsoft Java Virtual Machine
(JVM) 4.0 or 5.0
Netscape 4.05 with JDK 1.1 fixpack MSIE 5.x with Microsoft JVM 4.0 or 5.0
Netscape 4.06 (no fixpack required)
Netscape 4.5x (no fixpack required)
Netscape 4.7x (no fixpack required)
Notes:
1. The ESS Web interfaces do not support Netscape above version 4.7.x
2. If your ESS is running with ESS LIC earlier than level 1.3.0 or SC01206, the performance of the
ESS Web interfaces on MSIE 5.0 with JVM 5.0 is slower than with Netscape. It is recommended that
you use Netscape as the browser or move to LIC level 1.3.0 or higher
3. MSIE 5.0 with JVM 4.0 is supported with all levels of ESS code. However, it is not recommend that
you change JVM 5.x to JVM 4.0 on the ESSNet machine in order to improve performance. It is not
trivial to change the JVM to a lower level.
4. The ESS Master Console running Linux does not support the MSIE browser.

82

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
5.3 ESS Specialist components
The ESS Specialist is a software package residing inside the ESS and includes the following
components:

Storage function operation interface (SFOI)

Information server

Web server

Client code
The SFOI, Information server and Web server belong to the ESS itself and run inside the
cluster controllers. The client software is a set of Java applets running inside a Web browser
on the client workstation (see additional description in 5.2, “ESS Specialist prerequisites” on
page 81).
The SFOI is an internal application programming interface (API) to the ESS server functions
required by the ESS Specialist. The ESS Specialist uses this API to perform queries of the
current configuration and service setup for the facility. All functions, including the ESS
Specialist, SFOI and ESS Information server, runs on both cluster controllers of the ESS. To
enable the ESS Specialist to present a consolidated view of the storage facility, the SFOI
maintains a cross-cluster socket connection. This ensures that the configuration information
from both clusters is available to you even though network connection to only one of the two
cluster controllers is established.
The Information server is essentially a communication link that communicates on one side
with the client applets through an Ethernet TCP/IP socket connection and on the other side
with the SFOI. It provides the most up-to-date configuration information to the client applets,
so that it can be displayed to the end user or to allow the end user to change the
configuration.
The ESS Specialist Web server is set up for secure connection through the Secure Socket
Layer (SSL). The data sent between the Web server and the Web browser on the client side
is encrypted through the use of public and private key pairs.
The client code of the ESS Specialist consists of a set of hypertext markup language (HTML)
pages and Java applets, which are executed from inside a Java-enabled Web browser
running on the end-user’s network connected workstation. The user has the option of
connecting to either of the two cluster controllers by specifying a universal resource locator
(URL) consistent with the controller’s IP address or hostname.
5.4 Connecting to your ESS Specialist
To connect to the ESS Specialist through the browser, enter the URL of one of the two
clusters of your ESS machine. You can connect through either cluster but we recommend that
once you start updating and modifying from one cluster that you continue to make your
changes through that cluster, for the duration of the change. This is to give both clusters time
to synchronize.
For ease of identification, you could add a suffix such as c0 or c1 to the selected hostname for
your ESS to represent the different clusters. For example, ESSa90c0 for cluster 0 as shown in
Figure 5-1. Then bookmark it for ease of use. You can also enter in the TCP/IP address
assigned to the cluster (See Figure 5-2). When assigning the TCP/IP address, we
recommend that you make the last field of the second cluster only one digit higher than the
first cluster. For example if the left cluster or first cluster is 9.67.51.170, then make the right
cluster of second cluster 9.67.51.171.

Chapter 5. IBM TotalStorage Enterprise Storage Server Specialist
83
Figure 5-1 Entering the URL using the cluster name of your ESS
Figure 5-2 Entering the URL using the TCP/IP address
5.4.1 ESS Master Console
To aid you in getting the ESS set up and configured, the ESS will optionally come with a
private network already set up and a PC system to access the network and the ESS
Specialist (See Section 3.7.1, “IBM TotalStorage Enterprise Storage Server Master Console”
on page 39). The ESS Specialist will always be available to the user or IBM support
personnel for configuration or status. In addition, the PC can be set up to be a proxy server to
the ESS Specialist from anywhere in the customer environment.
5.4.2 ESS Specialist panels
The ESS Specialist is the major interface between your storage administrator and the ESS. It
is used to configure the ESS logical storage environment and provide administrative and
problem status information.
This section gives a very brief introduction to some of the ESS Specialist panels and features.
For detailed information about the use and panels of the ESS Specialist, refer to
IBM
TotalStorage Enterprise Storge Server Web Interface User’s Guide,
SC26-7346.
For ESS Specialist as it relates to S/390, refer to Chapter 8, “CKD storage configuration” on
page 101. For ESS Specialist as it relates to fixed block storage, refer to Chapter 15, “ESS
configuration for open systems fixed block storage” on page 235.
The ESS Specialist is frame enabled. When running inside a Web browser, it presents two
main frames on the Admin panel — a navigation frame on the left with hyperlink buttons for
accessing ESS Specialist functions and the working frame on the right. At the bottom of the
panel is the message area where messages are displayed while applets are processing or
the client browser is waiting to receive data from the server.

84

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
5.4.3 Welcome panel
From the screen similar to the one shown in Figure 5-3, you may select to enter the ESS
Specialist, the ESS Copy Services, or Tools from the navigation frame.
Figure 5-3 ESS Specialist Welcome screen
After selecting the ESS Specialist, options for managing the ESS can be selected from a
panel similar to the one in Figure 5-4. This is the introduction and welcome screen. You can
see the serial number, model, machine type and world wide node name of your ESS.
Note:
The screen shown in Figure 5-3 should not be closed, just minimize it. This is
because closing it will affect the associated applets.

Chapter 5. IBM TotalStorage Enterprise Storage Server Specialist
85
Figure 5-4 ESS Specialist function selection panel
5.4.4 Help
The help system includes content help, task help, specific scenario assistance and a
glossary. You can press the help icon at any time to bring up a new browser window that will
display the help text. The help icon as shown in Figure 5-5 is located on the top right-hand
side of all panels.
Figure 5-5 ESS Specialist help icon
5.4.5 Security
Accessing the ESS through the ESS Specialist requires a valid username and password. As
an option, the ESS Specialist can be configured to require that a user originate from a specific
IP address or a range of IP addresses. By selecting the Users button from the ESS Specialist
panel as shown in Figure 5-4, a user can be restricted to one of four different authorization
levels — viewer, operator, configurator or administrator. See Figure 5-6.

86

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Figure 5-6 Modify users panel
A viewer can view the current configuration and status information, while a configurator can
view the current configuration, status information and can make changes to the configuration.
An administrator can define new user IDs, delete old IDs and assign, change and revoke
passwords and levels of authorization, as well as configure the ESS.
All data sent between the ESS and the ESS Specialist interface in your Web browser is
encrypted to avoid unauthorized modification of configuration commands during transit. As a
result, the Web browser will warn the user that an encrypted site is being accessed by
displaying a sequence of certificate windows.
These windows tell your browser that the Web site represented by the ESS URL should be
treated as a trusted site, and that encrypted communications should be allowed between your
browser and the site.

© Copyright IBM Corp. 2002
87
Chapter 6.
IBM StorWatch Enterprise
Storage Server Expert
In this chapter, we discuss the IBM StorWatch Enterprise Storage Server Expert (ESS
Expert), an IBM offering which is a storage management tool for managing the ESS assets,
capacity and performance. We explain what it is and how it relates to the ESS.
6

88

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
6.1 Using ESS Expert
The IBM StorWatch Enterprise Storage Server Expert (ESS Expert) provides storage
resource management for the IBM TotalStorage Enterprise Storage Server and the IBM
TotalStorage Enterprise Tape Library (ETL). This chapter will only cover the applications and
functions of the Expert as it relates to the ESS (the ESS Expert).
The ESS Expert is a program product which can be purchased and used as a storage
monitoring and management tool for the ESS. this tool is accessed using a Web browser
interface.
The ESS Expert prepares reports and graphical charts from the data it collects. Data may be
collected from any of the ESSs attached to the same network as the ESS Expert.
ESS Expert helps you with the following tasks:

Asset management:
Data is collected and presented in either summary or detail reports
on the ESS. The type of information reported includes system names, LIC levels, device
nodes, clusters and features installed.

Capacity management:
ESS Expert can be used for central tracking and reporting of the
ESS capacity. Assigned storage, unassigned storage and freespace can be reported on
each ESS.

Performance management:
Performance data is collected and prepared for
presentation. The storage administrator may view the number of I/O requests based on
cluster, device adapter, rank or volume. Performance reports on caching statistics and disk
utilization can be observed to analyze the I/O subsystem performance.
The information collected by the ESS Expert along with host based tools will help the storage
administrator to manage the capacity and performance of the ESS. Further details on using
the ESS Expert can be found in
IBM StorWatch Enterprise Storage Server Expert Hands-on
Usage Guide,
SG24-6102

which details the planning, implementation and usage of the ESS
Expert.You can download a copy of this guide from the Web at the following location.
6.2 Accessing ESS Expert
The ESS Expert can be accessed using a Web browser running in a server attached directly
to the ESS via Ethernet, or in a remote machine if the ESS is connected into the user’s
network.
To familiarize you with the ESS Expert, we are providing some screen shots. Enter the IP
address of the ESS Expert from your Web browser and you’ll be given the sign-on screen as
shown in Figure 6-1.This is your doorway into the ESS Expert.
The ESS Expert interface consists of these major components:

The information frame:
This displays status and messages. You can press the help icon
located in the information frame any time for additional help.

The navigation frame:
This presents folders to access the ESS Expert applications and
tasks.

The work frame:
This is the main area in the center of the browser window which
contains the actual information being displayed by the interface, and it allows you to
perform tasks. Every screen displayed in the work frame contains a page-help icon.

Chapter 6. IBM StorWatch Enterprise Storage Server Expert
89
When you click the help icon, it links to the help system and displays the help information
specific to the current work frame screen. The help system is always contained in a separate
browser window. Therefore, you may have to switch from the main browser window to the
help browser window after clicking the help icon to actually see the displayed help
information.
Figure 6-1 ESS Expert: Welcome panel
6.3 Navigating through ESS Expert
Once you have signed on to the ESS Expert, you’ll be presented with the
Introduction
panel.
We have expanded the Manage StorWatch folder in the navigation frame for you in
Figure 6-2. From this folder you can perform tasks which are related to managing the ESS
Expert system. These tasks include discovering and managing the nodes (clusters of the
ESS), performing administrative tasks for users of the ESS Expert, administering the ESS
Expert database, and monitoring the ESS Expert tasks. Further information on these tasks
can be found by invoking the help function or by reviewing the IBM Redbook entitled,
IBM
StorWatch Enterprise Storage Server Expert Hands-on Usage Guide,
SG24-6102.

90

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Figure 6-2 ESS Expert: Managing the ESS Expert
Now we have expanded the Manage ESS folder in the navigation frame and presented it in
Figure 6-3. It is from this application section that you access the tasks for managing the
assets, managing the capacity and managing the performance of your ESS. For managing
assets and capacity, you may want to set up a task to run periodically to gather the
information on your ESS. You may then display pertinent information about the ESS, such as
the ESS serial numbers and IP addresses, active, previous and next LIC levels. From a
capacity perspective, you can view reports which show how the storage has been assigned,
how much cache and NVS is available for each cluster and other capacity related information.
When using the managing performance for the ESS, you have the capability to set up tasks
which can be run regularly or one time only to collect the performance statistics of the ESS
being monitored. In order to view the reports of this collected data, you will need to run a data
preparation task. Once the data is prepared in report format, you can view the performance of
your ESS based on the cluster, device adapters or raid rank. Refer to
IBM StorWatch
Enterprise Storage Server Expert Hands-on Usage Guide,
SG24-6102 for sample
performance reports and how they can be used.

Chapter 6. IBM StorWatch Enterprise Storage Server Expert
91
Figure 6-3 ESS Expert: Managing ESS
The reports presented by the ESS Expert supplement the host based tools which monitor
system performance. The complementation depends on the host. For instance, the z/OS
systems may use RMF from which they can obtain the caching statistics based on logical
volume along with the rank statistics. With a UNIX or NT based system,
iostats
or the
performance monitor will give you a further picture of the ESS by providing the caching
statistics for you to monitor and use to tune your ESS.
6.3.1 Additional documentation on ESS Expert
We suggest that you go to
IBM StorWatch Enterprise Storage Server Expert Hands-on Usage
Guide,
SG24-6102 for more detailed information on the ESS Expert. This publication shows
how to plan for the ESS Expert and how to install, implement, and use the tool.

92

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment

© Copyright IBM Corp. 2002
93
Part 2
Implementation in the
zSeries environment
In this part we discuss the processes for implementing the IBM TotalStorage Enterprise
Storage Server in the zSeries environment. We first cover software requirements for the
zSeries operating systems, then discuss in length how to design and implement an ESS
logical configuration that meets your needs. We also cover monitoring and reporting
enhancements in support of the ESS, discuss migration scenarios, and the use of copy
services with the ESS.
Part 2

94

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment

© Copyright IBM Corp. 2002
95
Chapter 7.
zSeries systems support
In this chapter we first cover the S/390 and zSeries operating system basic software support
as of the initial ESS announcement. We then cover software support for the more recent ESS
enhancements. We list the required minimum software release levels for the zSeries
operating systems that support the ESS. We do not provide PTF level information in this
redbook. For information regarding APARs and PTFs required in each case, please refer to
the appropriate Preventive Service Planning (PSP) information, or contact your IBM
representative.
7

96

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
7.1 Basic support
The ESS is supported by the z/OS, OS/390, z/VM, VM/ESA, VSE/ESA, and TPF operating
systems.
7.1.1 OS/390 and z/OS support
For the currently supported releases of OS/390 and z/OS the ESS runs in what is called
Exploitation mode. Exploitation mode means that the host recognizes the ESS as a 2105
control unit and can exploit the PAV capability. You use CNTLUNIT UNIT=2105 to define the
ESS. The currently supported releases are DFSMS/MVS 1.4 and higher.
Exploitation support is provided with PTFs for OS/390 1.3-2.6 with DFSMS/MVS 1.4. Support
is integrated in OS/390 2.7 with DFSMS/MVS 1.5 and above, and in z/OS 1.1 and above.
ICKDSF (refresh) release 16, EREP version 3 release 5, and DFSORT for MVS/ ESA release
13 are also prerequisites.
The minimum software level for dynamic PAV support is OS/390 2.7 with DFSMS/MVS 1.5.
Earlier releases only support static PAV. For dynamic PAV, the Workload Manager (WLM)
must run in goal mode.
Previous releases of OS/390 and DFSMS/MVS now not supported, will have to run the ESS
in either Transparency or Toleration mode. These modes of operating the ESS must be
considered as transitional because the recommendation is to have all your systems at the
currently supported levels of software and so capable to run the ESS in Exploitation mode.
Transparency support provides the base functions of an IBM 3990 Model 6 Storage Control.
OS/390 sees the ESS as multiple 3990 control units. PAV support is not available in
transparency mode. The host cannot share an I/O Definition File (IODF) with an exploiting
system. You define the logical control units to the host system using IOCP specification
CNTLUNIT UNIT=3990.
Toleration support permits a level of OS/390 which does not support PAV to share the ESS
with releases of OS/390 and z/OS which do. The OS/390 host recognizes the ESS as a 2105
control unit and recognizes the base and alias addresses needed to support PAV. With
toleration support, only non-PAV UCBs are built, so the system cannot utilize PAVs. However,
the host can share an IODF with exploitation-capable systems. You define ESS logical control
units using the new unit type, CNTLUNIT UNIT=2105.
7.1.2 VM/ESA support
In this section we discuss VM/ESA support.
CP native support
The following support is provided:

VM/ESA supports the ESS as an emulation of multiple 3990 Model 6 Storage Controls
with up to 256 unit addresses each.

No native use of new functions (no use of new CCWs, no PAV support for CP/CMS use).
Note:
For latest information on software levels and software maintenance requirements
you should check the corresponding preventive service planning (PSP) bucket.

Chapter 7. zSeries systems support
97
Guest support
The following support is provided:

VM/ESA 2.3.0 with an enabling APAR:
Guest use of read/write track CCWs only

VM/ESA 2.4.0 with an enabling APAR:
Guest use of exploitation functions
VM/ESA 2.4.0 support for z/OS and OS/390 guests
The following support is provided:

Parallel Access Volumes:
Supported for guest use only

FlashCopy:
Supported for guest use only
Utilization of ESS functions
PPRC is supported in VM/ESA and managed by ICKDSF or the IBM TotalStorage Enterprise
Storage Server Specialist (ESS Specialist). FlashCopy can be managed by the ESS Copy
Services Web interface.
Multiple allegiance and I/O Queueing are ESS hardware functions, independent of software
support. In a shared environment VM/ESA can take advantage of it. The priority byte,
however, is not set by VM/ESA and, therefore, I/O Queueing is not applicable.
7.1.3 VSE/ESA support
VSE/ESA supports the ESS as an emulation of multiple 3990 Model 6 Storage Controls with
up to 256 unit addresses each. PAV support is not available. New read/write track CCW is not
supported.
VSE/ESA support levels
ESS is supported from VSE/ESA 2.1.0 and up. No PTFs are required, however, we
recommend you contact your IBM support center.
Utilization of ESS functions
PPRC is supported in VSE/ESA and managed by ICKDSF or the ESS Copy Services Web
interface. FlashCopy can be managed by ESS Copy Services Web interface.
Multiple Allegiance and I/O Queueing are the ESS hardware functions. In a shared
environment VSE/ESA could take advantage of it, but the priority byte, however, is not set by
VSE/ESA and therefore I/O Queueing is not applicable.
7.1.4 TPF support
TPF supports the ESS as an emulation of multiple 3990 Model 3 TPF Storage Controls with
up to 256 unit addresses each. ESS supports the TPF Multi Path Locking Facility.
TPF support levels
ESS is supported with
TPF 4.1
.
Utilization of ESS functions
With applied corresponding PTFs TPF 4.1 is capable of using the new performance
enhanced Read Track/Write Track CCWs.
Multiple Allegiance function is available in TPF environments as an RPQ. TPF takes
advantage from Multiple Allegiance and I/O Queueing functions.

98

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
FICON attachment support
TPF supports native FICON attachment. This support requires TPF Version 4, Release 1.
The ESS has to be an F10 or F20 at LIC level 1.5.0, or later.
PPRC and FlashCopy support
PPRC and FlashCopy are supported on TPF. This requires TPF Version 4, Release 1.
Additional PTFs may be required. PPRC and FlashCopy are supported on the ESS models
F10 and F20 with LIC level 1.5.0 or later.
7.1.5 Linux
On zSeries 900 and on S/390 servers you can run Linux natively as a stand-alone or as a
logical partition (LPAR). In addition, the S/390 Virtual Image Facility for Linux, and z/VM V4
enable you to run more Linux images than can be deployed using LPARs and provide
capabilities to help create and manage these images.
Current Linux for S/390 distributions supported with the ESS are SuSE Linux Enterprise
Server for S/390 and Turbo Linux Server 6 for zSeries and S/390. For the most current
information on the Linux for S/390 releases and distributions that are supported with the ESS,
refer to:
http://www.storage.ibm.com/hardsoft/products/ess/supserver.htm
7.2 FlashCopy support
The minimum levels for FlashCopy support are:

DFSMS/MVS 1.3

VM/ESA 2.4.0 provides FlashCopy support for guests

VSE/ESA 2.5.

TPF 4.1
See PSP bucket information for required PTFs.
7.3 FICON support
Native FICON attachment is supported on ESS models F10 and F20 with LIC level 1.5.0, or
later, and using the Fibre Channel/FICON host adapters (FC 3021 or 3023). FICON support
requires the following minimum levels of S/390 and zSeries operating systems:

z/OS 1.1, or later, is required.

OS/390 2.8 is the minimum level.

System Automation for OS/390 1.3 is the minimum level.

VM/ESA 2.2 is the minimum level.

VSE/ESA 2.1.0 is the minimum level.

TPF 4.1 is the minimum level.
PTFs are required for FICON support. See the PSP bucket for detailed information, or consult
your IBM representative.

Chapter 7. zSeries systems support
99
7.4 Large Volume Support
The ESS initially supported custom volumes of up to 10017 cylinders, the size of the largest
standard volume, the 3390 model 9. This was the limit set by the operating system software.
The ESS Large Volume Support (LVS) enhancement, announced in November 2001 has now
increased the upper limit to 32760 cylinders, approximately 27.8 GB. The enhancement is
provided as a combination of ESS licensed internal code (LIC) changes and system software
changes, available for z/OS, OS/390, and z/VM.
Requirements
Large Volume Support is available on ESS models F10 and F20 with ESS LIC level 1.5.0, or
later, and requires one of the following operating system levels:

z/OS V1.1, or later + PTF

OS/390 V2.10 + PTF

ICKDSF R16 + PTF

DFSORT R14 + PTF

z/VM V3.1, or later + PTF
For DFSMS/MVS components this support is provided as a Small Programming
Enhancement (SPE) on OS/390 Release 10 (HDZ11F0) and integrated into z/OS Release
1.3 (HDZ11G0).
z/VM supports these large volumes as native device as well as for guest support.
VSE does not support large volumes.
Install large volume support on your systems before defining large volumes on the ESS.
Large volume support needs to be installed on all systems in a sysplex prior to sharing data
sets on large volumes. Shared system/application data sets can not be placed on large
volumes until all system images in a Sysplex have the large volume support installed.
Installation of PTFs on some components will require a system IPL to activate.
Please check PSP bucket information for required PTFs. Check with your OEM software
product vendors for changes to their products which may be required in support of large
volumes.
Coexistence support
For DFSMS/MVS 1.4 and 1.5, a coexistence PTF is provided that will allow these system
levels to coexist in the same Sysplex with LVS systems. You must install this PTF in order to
prevent unpredictable results that may arise from systems without large volume support
accessing volumes that have more than 10017 cylinders. The coexistence PTF will:

Prevent a device with more than10017 cylinders from being varied online to the system.

Prevent a device from coming online during an IPL if it is configured with more than 10017
cylinder.
Coexistence PTFs will also be required by DFSMShsm on all releases prior to OS/390 R10
because of the updates being made to the record format in the DFSMShsm control data sets.
Coexistence support will not be available for DFSMS 2.10 and higher. Install the large volume
support prior to using data on large volumes at these release levels. No coexistence support
will be available for DFSMS/MVS 1.3 (unsupported release), or earlier.

100

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
7.5 CUIR support
In large configurations, quiescing channel paths in preparation for upgrades or service
actions is a complex, time consuming, and potentially error prone process. Control Unit
Initiated Reconfiguration (CUIR) automates the process of quiescing and re-enabling channel
paths, reducing the time required for service actions and reducing the operations staff
required efforts, and reduces the possibility for human error.
CUIR requires ESS licensed internal code (LIC) level 1.5.0, or later, and is supported on
zSeries and S/390 platforms by the following operating systems:

z/OS V1.1, or later + PTF

OS/390 V2.10 + PTF

z/VM V3.1, or later + PTF
Many of the support PTFs require a system level IPL to activate. CUIR support will also
require host channel microcode updates (MCLs). Please check PSP bucket information for
required PTFs.

© Copyright IBM Corp. 2002
101
Chapter 8.
CKD storage configuration
In this chapter we provide guidance and recommendations on how to configure an IBM
TotalStorage Enterprise Storage Server (ESS) to meet your needs. We begin by introducing
the logical configuration components. We then describe the procedures using the IBM
TotalStorage Enterprise Storage Server Specialist (ESS Specialist) to configure the ESS for
use with S/390 and zSeries hosts.
8

102

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
8.1 Basic concepts
For the S/390 and zSeries environments, the ESS appears to the host as a 2105 control unit
with one or more logical control units (LCUs) attaching 3390 devices in either 3390 or 3380
track format. The number and type of storage controls and devices an ESS emulates, that is
the
logical configuration
of the ESS, can be selected by the user. One of the key tasks of ESS
implementation planning is to decide, what the logical configuration of an ESS should be.
During installation, the logical configuration is defined on the ESS so that it will present to the
host the desired image. Later, it can be modified if requirements change.
The target configuration you will implement depends on several factors, such as the installed
capacity of the ESS, your host system configuration, host software level, and the type of
logical volumes you need for your applications. You need to plan for the number and type of
logical control units and logical volumes you will define. If your system supports Parallel
Access Volumes (PAV), you need to plan how to allocate them. You will also need to decide
on which configuration method to use. These items should be planned well in advance of the
actual installation in order to guarantee a successful implementation of the ESS.
In this section we introduce the basic concepts you need to understand in order to configure
an ESS for use on S/390 and zSeries systems.
8.1.1 ESS Specialist
The IBM TotalStorage Enterprise Storage Server Specialist (ESS Specialist) is a software
component that runs in the ESS clusters and provides the interface for the user to define the
logical configuration of the ESS. User communication with ESS Specialist is achieved using a
Web browser such as Netscape or Internet Explorer, running on a Web client. The Web client
is connected to one or more ESS subsystems using the TCP/IP protocol through the 10baseT
Ethernet LAN adapter in each ESS cluster.
Refer to 3.8.1, “ESS Specialist” on page 43 for ESS Specialist requirements, and also to
Chapter 5, “IBM TotalStorage Enterprise Storage Server Specialist” on page 79 for an
introduction to the ESS Specialist.
8.1.2 Logical Control Units
The ESS can emulate up to 16 ESCON or FICON attached Logical Control Units (LCUs) with
emulated 3390 devices. The LCUs you want the ESS to present to the host must be
configured on the ESS. The LCU and its logical volumes must also be defined in the host
IODF to enable host communication with those volumes. Each LCU corresponds to one CKD
format ESS Logical Subsystem (LSS). The terms LCU and LSS are therefore often used
interchangeably in documentation for CKD storage. The term LCU only refers to a CKD LSS.
An LCU can be defined to emulate one of the following control units:

3990 Model 6

3990 Model 3

3990 Model 3 TPF

3990 in Siemens BS2000 compatibility mode.
Every installed ESCON and FICON host adapter port is capable of accessing all configured
LCUs. ESCON and FICON attached hosts are identified to the ESS when you make the
physical connection between the hosts and the storage server. These hosts are seen as a
single ESCON or FICON net in the ESS Specialist for improved graphical presentation.

Chapter 8. CKD storage configuration
103
The LCUs are numbered from 0 to F. (The ESS Specialist typically uses a three digit notation,
0x000 to 0x00F.) These map directly to LSS numbers 00 to 0F. See Figure 8-1.
An LCU can have up to 256 logical volumes defined to it, one per unit address 00-FF. The
ESS can therefore have up to 16 * 256 = 4096 CKD logical volumes allocated. If Parallel
Access Volumes (PAVs) are enabled for the LCU, each PAV alias takes up one of the 256
addresses.
Each LCU is uniquely associated with one cluster, and one device adapter (DA). On each
device adapter you can configure up to two CKD LCUs. The even numbered LCUs are
managed by ESS Cluster 1, the odd numbered by Cluster 2. Under normal conditions all I/O
to the LCU is done through that particular cluster and DA. Each LCU is also associated with
half of the disk groups on the two loops attached to the device adapter. (See explanation of
disk groups later in section 8.1.3, “Disk groups and ranks” on page 104.) A LCU can be
assigned physical disk capacity from either or both of the loops on the DA. The associations
between LCUs, DAs, and clusters are predefined and shown in Figure 8-1.
Figure 8-1 Relationship between S/390 LCUs and SSA loops
For example, LCUs 4 and C are managed by Cluster 1 and DA 03 (in adapter pair 3). The
logical volumes belonging to LCUs 4 and C are physically located on one or more (maximum
of six) disk groups on SSA loops A and B of DA pair 3. SSA loops always contain an even
number of disk groups, evenly divided between the two clusters. Thus loops A and B of DA
pair 3 may also have one or more disk groups that belong to LCUs 5 and D, and are managed
by Cluster 2 and by DA 13 (also in DA pair 3).
In the event of a cluster failover, the other cluster can take over all LCUs and disk groups.
Data would then be accessed through the DAs in the surviving cluster.
At ESS installation time the IBM System Support Representative (IBM SSR) can configure
the ESS to support either 0, 8, or 16 CKD LSSs and 0, 8, or 16 FB LSSs. The default for both
is 16. If you do not plan to use your ESS for open systems, have the maximum LSS value for
FB set to 0. This will save some 2MB of cache memory for each LSS. However, we
recommend that you keep the number of CKD LSSs in the default 16, as this will maximize
DA 11
DA 12
DA 13
DA 14
DA 01
DA 02
DA 03
DA 04
Cluster 1 Cluster 2
LSS LSS
00
LCU LCU
000
08008
02002
0A00A
04004
0C00C
06
006
0E00E
01 001
09
009
03 003
0B 00B
05 005
0D 00D
07
007
0F 00F
Loop A
Loop B
DA pair 1
Loop A
Loop B
DA pair 2
Loop A
Loop B
DA pair 3
Loop A
Loop B
DA pair 4

104

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
the number of devices you can define on the ESS. This does not mean that you have to
configure all 16 LCUs initially, it just means that 16 are available for you to configure should
you later need them. Changing the Maximum LSS values after the initial configuration will
require an IML of the ESS.
8.1.3 Disk groups and ranks
A
disk group
in ESS is a group of eight disk drive modules (DDMs) of the same size on an
SSA loop. As physical capacity on the ESS is always installed in pairs of disk 8-packs of the
same size on a loop, there can be 0, 2, 4, or 6 disk groups on any loop.
The grouping of DDMs into disk groups is done by the ESS when 8-packs are installed in the
ESS. You cannot select which DDMs belong to each disk group. With the ESS Specialist you
cannot display which DDMs belong to a certain disk group.
A disk group must be configured before its capacity can be allocated for host system use. You
can configure the disks in a disk group to form a RAID-5 array, or leave them as eight
individual non-RAID disks (JBOD). The RAID-5 array, or the non-RAID disks must further be
configured in either 3390 track format or 3380 track format for use on zSeries hosts.
Alternatively, they may be configured as FB storage for open systems use.
The formatted disks are referred to as
ranks
. Thus, a RAID array is called a RAID rank, and
the individual disks are called non-RAID ranks (JBOD ranks). A rank is the unit on which
logical volumes are later allocated. A CKD rank is assigned to a LCU when the rank is
configured.
Each disk group is associated with and managed by one device adapter (DA) and one of the
two ESS clusters. A disk group can only be assigned to one of the four LSSs (two CKD LSSs
and two FB LSSs) that are associated with that particular DA. Figure 8-2 shows an example
from a fully configured ESS. Any of the highlighted disk groups can be assigned to any of the
four LSSs, regardless of which loop the group is in. Since there is always an even number of
disk groups associated with a DA, you can divide them evenly between two LCUs if you plan
to define both LCUs, or in smaller configurations assign them all to one LCU.
Figure 8-2 Disk group to LSS association
16 06
Adapter 4
1E 0E
Cluster 2Cluster 1
LSS LSS
FB CKD CKD FB
03 13
Adapter 2
0B 1B
05 15
Adapter 3
0D 1D
07 17
Adapter 4
0F 1F
01 11
Adapter 1
09 19
A
A
A
A
A
A
A
A
B
B
B
B
B
B
B
B
Disk groups associated with Cluster 1, Adapter 2
Groups 2, 4, 6 Grou
p
s 1, 3, 5

Chapter 8. CKD storage configuration
105
All I/O operations to a disk group are performed by the cluster and DA it is associated with. In
the event of failure, the other cluster can take over all operations.
On the ESS Specialist
Storage Allocation
panel each rectangle in the middle represents a
disk group (Figure 8-3). The panel only displays configured disk groups. There may be disk
groups on the loop that have not yet been defined. Because there is always an even number
of disk groups on each loop, you know for certain that there are unconfigured disk groups on
the first and last loop in the figure. The panel displays both CKD and FB disk groups. The
highlighted groups shown in the figure are CKD storage. They are highlighted because we
clicked on the EsconNet host icon (not visible). The unhighlighted groups are configured as
FB. The figure also shows which disk groups belong to which cluster.
Figure 8-3 Disk groups in the Storage Allocation panel
The physical drives in a disk group typically reside in two 8-packs on a loop. When two
8-packs are installed on a loop, the physical disks drives form two disk groups, both
containing four drives from each 8-pack. Over time, the DDMs belonging to a disk group may
change due to disk sparing in RAID arrays.
See the
IBM TotalStorage Enterprise Storage Server,
SG24-5465 for more details on disk
groups and ranks.
Here is a summary of disk group attributes:

A disk group contains eight physical disk drives (DDMs).

A disk group is contained on a single SSA loop. It cannot span both loops on a device
adapter pair.

All DDMs in a disk group are of the same size. A loop may contain different size DDMs.

The number of disk groups on a loop is the same as the number of 8-packs installed on
the loop. That is, it can be 0, 2, 4, or 6.

The disk groups on a loop are numbered from 1 to 6. The even numbered disk groups are
managed by Cluster 1. The odd numbered groups are managed by Cluster 2.

A disk group can be configured in either of two ways: as a RAID-5 array or as non-RAID
disks (JBODs).
8.1.4 RAID arrays
The RAID-5 arrays on ESS will have one of two array configurations:

6+P+S
. This setup leaves one spare disk (S) in an SSA loop. The capacity of six physical
disks drives is used for storing data and one for storing the parity information (P). The first
two disk groups of any given DDM size on a loop will automatically have a spare physical
disk drive and the arrays will be 6+P+S.

7+P
. In this setup the capacity of seven disks is used for storing data, one for storing the
parity. The third and subsequent arrays of a given DDM size in a loop will be 7+P.

106

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Non-RAID ranks may also be used for CKD volumes, but we recommend the use of RAID
arrays due to the availability they offer. For this reason our examples mostly cover RAID
configurations.
When you configure a RAID array for CKD storage, you can optionally auto-allocate a group
of standard volumes on the array. As a result, two partitions, the interleaved partition and the
non-interleaved partition, are created on the array. The interleaved partition contains the
auto-allocated standard volumes. The non-interleaved partition will initially be empty.
Optionally, you may choose not to auto-allocate any volumes. In that case there will be no
interleaved partition, the whole array capacity will be used for the non-interleaved partition.
The auto-allocate option is available at RAID array configuration time, either when the array is
configured using the ESS Specialist, or when configured by the IBM Systems Support
Representative (IBM SSR) using the Batch Configuration tool (see 8.3, “Configuration
process” on page 114).
The interleaved structure has been designed to equalize the slight performance differences
that logical volumes residing towards the edges of physical disk drives in the RAID array have
due to longer average seek times. Interleaved logical volumes are divided into smaller stripes
which are distributed over the array to provide the volumes with uniform performance
characteristics. We recommend that you auto-allocate standard volumes, when possible.
The interleaved partition characteristics are described in the IBM Redbook
IBM TotalStorage
Enterprise Storage Server,
SG24-5465.
You cannot auto-allocate standard volumes on a 72.8GB disk array. The capacity of a 72.8GB
array is so large that it would not be feasible to allocate small standard volumes on it because
that would quickly exhaust the 256 device address limit of the LCU. Especially in multi array
LCUs you would not have enough room for alias devices to provide performance. (Refer to
Table 8-4 on page 111 for examples.) Large arrays are more suitable for large volumes with
PAV aliases. We recommend that you install the software support for Large Volumes and
allocate large custom volumes on 72.8GB arrays.
Here is a summary of RAID array attributes:

A RAID array is configured as either 6+P+S, or 7+P.

The first two disk groups of a given DDM size on an SSA loop have a spare disk drive, and
the RAID configuration is 6+P+S. This ensures that two spare disk drives of each DDM
size are available on the loop.

The third and subsequent arrays of a given DDM size on an SSA loop are 7+P.

A spare disk drive can be used for any RAID array of the same DDM size on that loop,
either CKD or FB.
8.1.5 CKD logical volumes
CKD logical volumes are mapped into a logical device map in a CKD LSS. The logical devices
represent the device address of the logical volume. It ranges from 0x00 to 0xFF. Because
each LSS is seen as a logical control unit, the zSeries systems will see it as a 3990-x with up
to 256 devices. The logical devices need not be mapped to the ESCON or FICON host
adapters, because the zSeries hosts have access to all the logical devices through any of the
ESCON or FICON connections available. The set of logical devices accessed by any zSeries
image is defined with the Hardware Configuration Definition (HCD) in the IODF file that the
operating system uses to recognize its hardware topology.

Chapter 8. CKD storage configuration
107
Logical volumes are allocated on ranks, either RAID ranks, or JBOD ranks. Logical volumes
cannot span multiple ranks. When a rank is configured, its track format is set as either 3390 or
3380. All logical volumes on the rank will be of the same track format. It is not possible to
intermix different track format volumes on a rank. However, different ranks in an LCU may
have different track formats.
An ESS
standard volume
is defined as a logical volume that is created by auto-allocation
when an array is configured, either during the initial Batch Configuration process, or using the
Configure Disk Groups
panel on the ESS Specialist. Standard volumes reside in the
interleaved partition of a RAID array, and are also referred to as interleaved volumes. You can
select to allocate one of the following standard volume types:

3390-2, 3390-3, or 3390-9 in 3390 track format

3390-2 or 3390-3 in 3380 track format
The number of standard volumes in an array is predefined, and depends on the volume type,
DDM size, and the array configuration. Refer to Table 8-1.
Custom volumes
are defined using the
Add Volumes
panel on the ESS Specialist. Custom
volumes reside in the non-interleaved partition of a RAID array, or on a JBOD rank. As the
name suggests, custom volumes have customized sizes. You can specify any number of
cylinders from 1 to 10017 when you define the logical volume. On systems that support large
volumes (LVS), you can specify up to 32760 cylinders. The volumes you define with the
Add
Volumes

panel are considered custom volumes, even when their size matches that of a
standard volume. Custom volumes are also called non-interleaved volumes.
You can define as many custom volumes as you like, within the limits of available capacity, up
to the maximum of 256 total devices per LCU.
Custom volumes are reported by the host system as:

3390-3 devices, when their size is between 1 to 3339 cylinders

3390-9 devices, when their size is more than 3339 cylinders
This implies that 3390-2 and 3390-2 (3380 track format) type devices can only be allocated as
interleaved standard devices. 3380 track format custom volumes can have maximum of 3339
cylinders, and are reported as 3390-3 devices.
Table 8-1 summarizes the capacities of different RAID arrays, and the number of logical
volumes that fit in the array. The number of volumes is indicated with an entry of format X+Y,
where X indicates the number of standard volumes in the interleaved partition, and Y
indicates the number of custom volumes of the same size that will fit in the non-interleaved
partition (NI-partition). Where standard volumes are not supported, only the number of
custom volumes is indicated. The number of standard volumes in each RAID array is fixed. If
you do not configure any standard volumes, you have the whole array capacity available for
custom volumes. Row LargeVol represents the largest possible custom volume, a 32760
cylinder 3390 volume.

108

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Table 8-1 Array capacities
Logical
device type
Logical
device
physical
capacity
(GB)
6+P+S array 7+P array
Physical
capacity
of array
(GB)
Physical
capacity of
NI-partition
(GB)
Number of
logical
volumes
Physical
capacity
of array
(GB)
Physical
capacity of
NI-partition
(GB)
Number of
logical
volumes
9.1 GB DDMs
3390-2 1.96 53.81 6.77 24 + 3 62.79 7.91 28 + 4
3390-3 2.94 53.81 6.77 16 + 2 62.79 15.75 16 + 5
3390-9 8.82 53.81 18.53 4 + 2 62.79 27.51 4 + 3
3390-2 (3380) 1.82 53.81 10.13 24 + 3 62.79 4.55 32 + 4
3390-3 (3380) 2.73 53.81 10.13 16 + 3 62.79 8.19 20 + 5
LargeVol 28.86 53.81 53.81 1 62.79 62.79 2
18.2 GB DDMs
3390-2 1.96 107.67 5.75 52 + 2 125.62 8.02 60 + 4
3390-3 2.94 107.67 13.59 32 + 4 125.62 8.02 40 + 2
3390-9 8.82 107.67 37.11 8 + 4 125.62 19.78 12 + 2
3390-2 (3380) 1.82 107.67 5.75 56 + 3 125.62 9.14 64 + 4
3390-3 (3380) 2.73 107.67 9.39 36 + 3 125.62 5.50 44 + 2
LargeVol 28.86 107.67 107.67 3 125.62 125.62 4
36.4 GB DDMs
3390-2 1.96 215.38 11.54 104 + 5 251.28 8.24 124 + 4
3390-3 2.94 215.38 15.46 68 + 5 251.28 16.08 80 + 5
3390-9 8.82 215.38 38.98 20 + 4 251.28 39.60 24 + 4
3390-2 (3380) 1.82 215.38 4.26 116 + 2 251.28 11.04 132 + 6
3390-3 (3380) 2.73 215.38 7.90 76 + 2 251.28 11.04 88 + 4
LargeVol 28.86 215.38 215.38 7 251.28 251.28 8
72.8 GB DDMs
3390-2 1.96 430.80 430.80 219 502.56 502.56 256
3390-3 2.94 430.80 430.80 146 502.56 502.56 170
3390-9 8.82 430.80 430.80 48 502.56 502.56 56
3390-2 (3380) 1.82 430.80 430.80 236 502.56 502.56 256 (1)
3390-3 (3380) 2.73 430.80 430.80 157 502.56 502.56 184
LargeVol 28.86 430.80 430.80 14 502.56 502.56 17
(1) 256 is maximum allowed. Full array capacity cannot be utilized using the volume type.

Chapter 8. CKD storage configuration
109
Table 8-2 summarizes the characteristics of CKD logical volumes.
Table 8-2 CKD logical device capacities
The physical capacity used is the amount of space required on the rank to configure the
logical volume. The difference between logical and physical capacity is mainly due to the fact
that data on the physical disks is stored in 524 byte sectors of which 512 bytes is available for
the logical volume.
You need the physical capacity information to calculate how many logical volumes you can
configure on a given ESS subsystem. Use the following formulae to
approximate
the physical
capacity of custom volumes. The formula do not completely reflect the algorithms of the ESS
as it allocates space for a logical volume.
The amount of physical capacity for 3390 devices can be approximated by:
Physical capacity = (((# Cylinders + 1) * Bytes per Cylinder * 524) / 512) * 1.013 bytes
The amount of physical capacity for 3380 devices can be approximated by:
Physical capacity = (((# Cylinders + 1) * Bytes per Cylinder * 524) / 512) * 1.122 bytes
The equations compensate for any overhead in the logical device such that the result is
always greater than or equal to the physical capacity required to configure the logical device.
Table 8-3 summarizes the number of volumes that will fit in the different RAID arrays. The
upper section of the table shows the figures for ESS standard volumes. This information is
essentially an extract from Table 8-1 on page 108. In the lower section we have included
selected custom volumes for your reference. Use the above formula to calculate the figures
for other custom volume sizes.
Logical device
type
Cylinders Bytes per
cylinder
Logical device
capacity (GB)
Physical
capacity used
(GB)
Standard volumes
3390-2 2226 849960 1.892 1.962
3390-3 3339 849960 2.838 2.943
3390-9 10017 849960 8.514 8.828
3390-2 (3380) 2226 712140 1.585 1.821
3390-3 (3380) 3339 712140 2.377 2.731
Custom volumes
3390-3 1-3339 849960 0.00085-2.838 0.00176-2.943
3390-9 3340-32760 849960 2.839-27.844 2.944-28.869
3390-3 (3380) 1-3339 712140 0.00071-2.377 0.00163-2.731

110

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Table 8-3 Number of volumes in a RAID array
Logical volume characteristics can be summarized as follows:

All logical volumes on a rank, both standard and custom volumes, have the same track
format, either 3390 or 3380. No intermixing is possible.

It is not possible to intermix CKD and open systems FB volumes on a rank.

A 3380 track format custom volume can have maximum 3339 cylinders.

JBOD ranks only contains custom volumes.
8.1.6 Parallel Access Volumes
The Parallel Access Volume (PAV) feature is performance feature that the IBM TotalStorage
Enterprise Storage Server brings specifically for the z/OS and OS/390 operating systems. It
allows your applications to access logical volumes in parallel. That is, your system can access
a single volume from a single host with multiple concurrent requests. This capability
represents a significant performance improvement by the ESS over traditional I/O processing
for the z/OS and OS/390 environments.
PAV is an optional priced feature of the ESS. In order to use PAVs, the feature must be
installed on the ESS. You then have to configure both your ESS and the MVS operating
system to use PAVs.

Use the ESS Specialist to define PAV alias devices for base devices on the ESS. This
creates a unit address relationship between the bases and aliases in the ESS hardware,
allowing concurrent I/O operations to a single volume by means of the multiple device
exposures.

Use HCD to define the corresponding device types on the host. You can specify in HCD
whether PAV bases and PAV aliases should be associated statically or dynamically,
resulting in static or dynamic PAVs.
The ESS does not distinguish between static and dynamic PAV - there are only PAVs.
Dynamic alias management is a function of the host software. When created, an alias device
must be assigned to a base device in the same LCU. WLM can then reassign the alias from
one base to another within the same LCU.
Volume type Vol
size
(cyl)
9 GB array 18 GB array 36 GB array 73 GB array
6+P 7+P 6+P 7+P 6+P 7+P 6+P 7+P
Standard volumes
3390-2 2226 27 32 54 64 109 128 219 256
3390-3 3339 18 21 36 42 73 85 146 170
3390-9 10017 6 7 12 14 24 28 48 56
3390-2 (3380) 2226 27 36 59 68 118 138 236 (1)
3390-3 (3380) 3339 19 25 39 46 78 92 157 184
Custom volumes
20000 3 3 6 7 12 14 24 28
30000 2 2 4 4 8 9 16 19
32760 1 2 3 4 7 8 14 17
(1) More than 256 volumes would be required to utilize full array capacity.

Chapter 8. CKD storage configuration
111
See the
IBM TotalStorage Enterprise Storage Server,
SG24-5465 for more details on PAV.
The ability to do multiple I/O requests to the same volume nearly eliminates IOSQ time, one
of the major components in response time. Traditionally, access to highly active volumes has
involved manual tuning, splitting data across multiple volumes, and more. With PAV and the
Workload Manager, your manual performance tuning efforts are minimized.
WLM Dynamic Alias Tuning will allow you to plan the I/O configuration at a coarser level of
granularity. Instead of needing to plan the exact relationship of PAV-alias volumes to PAV-base
volumes, and the number of PAV-alias volumes for each device, you just need to plan the total
size of the PAV-alias pool. This allows for a higher PAV base to PAV alias ratio, allowing you to
address more data behind a single ESS subsystem (some installations work well with one
dynamic alias for each three 3390-3s, and one alias per each two 3390-9s). Data set
placement and hot spot analysis will not be required as the operating system will determine
the PAV-alias to PAV-base relationships that best achieve the work load goals defined by you.
To achieve the benefits of PAVs, you need to design your ESS configuration so that there are
enough PAV alias devices available. The general rule of thumb is that you should have one
alias for a 3390-3 volume (2.8 GB) and three aliases for a 3390-9 volume (8.5 GB). This
translates to approximately 3 GB per alias. This applies to static PAVs. With dynamic PAV, you
will typically manage with less aliases, but you can keep 3GB per alias as a goal. The lower
the ratio is, the better. The larger the ratio gets, the more IOS queuing you can expect.
Larger volumes require more aliases to keep IOS queuing down, but also allow more aliases
to be defined for the LCU as there are fewer base devices. Consequently, they give better GB
per alias ratios, and typically provide better performance, particularly with dynamic PAV.
Table 8-4 shows the maximum number of PAV base and alias devices, and the GB per alias
ratio for different volume sizes in three different LCU configurations - an LCU with one 6+P+S
array (column 6P), a LCU with one 6+P+S and one 7+P array (column 6P+7P), and finally an
LCU with one 6+P+S and two 7+P arrays (column 6P+7P+7P). The GB per alias ratio is the
total capacity of logical volumes in the LCU divided by the number of aliases. The LargeVol
row represents arrays that are configured with 32760 cylinder volumes.
Table 8-4 PAV alias ratios
Ranks:6P 6P+7P 6P+7P+7P
Base
de-
vices
Alias
de-
vices
GB
per
alias
Base
de-
vices
Alias
de-
vices
GB
per
alias
Base
de-
vices
Alias
de-
vices
GB
per
alias
36.4 GB DDMs
3390-3 73 183 1.1 158 98 4.6 243 13 53.0
3390-9 24 232 0.9 52 204 2.2 80 176 3.9
LargeVol 8 248 0.9 17 239 2.0 26 230 3.1
72.8 GB DDMs
3390-3 146 110 3.8 256 0 (1) 256 0 (1)
3390-9 48 208 2.0 104 152 5.8 160 96 14.2
LargeVol 14 242 1.6 31 225 3.8 48 208 6.4
(1) More than the maximum 256 volumes would be required to utilize full LCU capacity

112

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Notice how larger volumes provide better GB per alias ratios. Notice also how in large
capacity configurations the 256 device limit per LCU becomes a constraint with smaller
volume sizes - not enough addresses are available for aliases. Larger volumes will relieve the
constraint, allow more PAV aliases, and consequently may provide better performance,
particularly with dynamic PAV. For 73GB arrays especially we recommend large volumes.
8.2 Designing the LCU address layout
In this section we give guidance on how to design your Logical Control Unit (LCU) address
layout, that is how to divide the LCU address range between base and alias devices. You
must define the configuration on the host system IODF and in the ESS and the two must
comply with each other, so it is important that you plan before starting the actual
implementation. All LCUs on an ESS will not necessarily be identical, the layout depends on
the capacity of each LCU and the type and number of volumes in each LCU.
The maximum number of devices in an LCU is 256. Each device has a unit address in the
range of 0x00-FF. This applies to both base devices and PAV alias devices. The devices are
assigned an address when they are configured on the ESS. The UNITADD parameter of the
CNTLUNIT macro can be defined with a 128 or 256 range value (Note: for planning purposes
consider that if you later have to increase the unitaddress range of the control unit, this will be
disruptive for the HCD).
When a logical volume is defined on the ESS, it is assigned the lowest available address in
the LCU address range. Thus, the base addresses typically occupy the low end of the LCU
address range (from 00 upwards). When alias devices are defined, they are assigned
addresses downwards from what is called the
Starting PAV
address (see 8.4.6, “Configuring
logical control units” on page 124). You can select the
Starting PAV
address when you define
the LCU with the ESS Specialist. The options are 0, 63 (0x3F), 127 (0x7F) and 255 (0xFF). If
all addresses below the
Starting
PAV address are already reserved, then the alias is assigned
the lowest available address above the starting address. (This is always the case when
Starting
PAV address is 0.) Thus, the alias devices typically occupy a contiguous address
range above the base device address range.
The S/390 Storage Requirements table in the
IBM TotalStorage Enterprise Storage Server
Configuration Planner,
SC26-7353 is useful for documenting your configuration. You may
want to add in the table rows to tie the ESS configuration with your IODF. See Table 8-10 on
page 140 for an example.
Layout 1: All 256 addresses are used
When PAVs are configured, you need to decide how many PAV aliases you can and want to
have, and where in the LCU address range they should be located. The base devices (logical
volumes) must be configured first and will occupy the low end of the address range as in the
first example. A straightforward approach is to use for aliases all the addresses left over from
the bases. In terms of performance this is a good solution as it will minimize IOS queuing
times. Assume the same configuration as in the previous example, but add the maximum
number of PAVs. We will have 52 bases and 204 aliases. See Table 8-5.
Table 8-5 Device address layout example 2 - 256 devices
Host I/O configuration ESS unit addresses
CNTLUNIT UNIT=2105,UNITADD=((00,256)),...
IODEVICE UNIT=3390B,ADDRESS=(9000,52),UNITADD=00,...
IODEVICE UNIT=3390A,ADDRESS=(9034,204),UNITADD=34,...
0x00-33 Base (52)
0x34-FF Alias (204)

Chapter 8. CKD storage configuration
113
When defining this configuration using the ESS Specialist, set Starting PAV Address for the
LCU to 255 (0xFF). Then define four aliases for 48 bases and three aliases for the remaining
four bases. Starting PAV Address of 0 would give you essentially the same result. The
addresses would be assigned in different order, but would occupy the same range.
PAVs are supported when the ESS has a PAV feature installed, and the host runs in
Exploitation mode.
Layout 2: Plan for future growth
So why would you want to define less than the maximum number of aliases? Each alias
device is assigned a UCB which requires some memory, but since the alias UCBs are located
above 16MB in virtual memory, that most often should not be an issue. One possible reason
could be, that you want to design your address layout in anticipation of future growth. For
example, assume you expect to upgrade the example subsystem to full 11TB capacity in the
future, and want to define your IODF already at this point so that you don’t have to change it
any more when the upgrade takes place. In a full configuration each LCU will hold 80 full size
3390-9 volumes which leaves 176 addresses for the aliases. See Table 8-6.
Table 8-6 Device address layout example 3 - non-contiguous address ranges
On the ESS Specialist, after defining the 52 base devices, select Starting PAV Address 255
(0xFF) for the LCU, then define three aliases for the 52 bases to come up with this
configuration. This leaves address range 0x34-63 undefined on the ESS. That’s where we
would later add 28 new logical volumes and 20 more aliases.
On the host we define the full address range in anticipation of the coming upgrade. Until the
upgrade takes place, addresses 0x34-4F appear to the host as non-existing devices, while
addresses 0x50-63 appear as unbound alias devices. (Refer to Example 8-1 on page 143 for
a sample D M=CHP command output.) The ESS Specialist would let us define alias devices
in range 0x34-4F, but we don’t want to do that because it would create an address mismatch
between the host IODF and the ESS. Addresses 0x34-4F are unusable at this point.
Layout 3: More connectivity
Another reason for defining less aliases is to reduce the LCU size. For example, an ESCON
port can address maximum 1024 devices, counting both base and alias devices. If you define
your LCUs with 256 addresses, one ESCON port can only access four LCUs. By limiting the
LCUs to 128 addresses, each ESCON port could access eight LCUs. By reducing the LCU
size you gain more connectivity. In our example of 52 base devices, an LCU size of 128 would
leave us with 76 alias addresses. See Table 8-7.
Table 8-7 Device address layout example 4 - 128 devices
On the ESS Specialist, select Starting PAV address 0 or 127 (0x7F) for the LCU and define
two aliases for 24 bases and one alias for the remaining 28 bases to come up with this
configuration.
Host I/O configuration ESS unit addresses
CNTLUNIT UNIT=2105,UNITADD=((00,256)),...
IODEVICE UNIT=3390B,ADDRESS=(9000,80),UNITADD=00,...
IODEVICE UNIT=3390A,ADDRESS=(9050,176),UNITADD=50,...
0x00-33 Base (52)
0x34-4F Future base (28)
0x50-63 Future alias (20)
0x64-FF Alias (156)
Host I/O configuration ESS unit addresses
CNTLUNIT UNIT=2105,UNITADD=((00,128)),...
IODEVICE UNIT=3390B,ADDRESS=(9000,52),UNITADD=00,...
IODEVICE UNIT=3390A,ADDRESS=(9034,76),UNITADD=34,...
0x00-33 Base (52)
0x34-7F Alias (76)

114

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
The 1024 address limitation is removed with FICON. A FICON channel can address up to
16384 devices.
Layout 4: Add aliases to improve performance
For a final example, assume you determine that, for performance reasons, you need two
aliases for all the base devices instead of what we had in the previous example. This will
make 52 bases and 104 aliases, for a total of 156 addresses. This will reduce connectivity,
each ESCON channel now being able to access six LCUs of that size. (Again, this is not the
case for FICON.) See Table 8-8.
Table 8-8 Device address layout example 5 - more aliases
Using the ESS Specialist with Starting PAV Address set to 0 or 127 (0x7F), add one new alias
for the 28 bases which in the previous example only had one. The ESS will automatically
assign the aliases from 0x80 upwards since all the lower addresses have been used.
8.3 Configuration process
There are two ways of configuring the ESS. The two methods are referred to as the
custom
configuration process
and the
standard configuration process
.
Custom configuration is the process of configuring the ESS using the ESS Specialist Web
interface. This method allows you to define fully customized configurations. In this redbook,
we describe the custom configuration process in detail.
The standard configuration process, also called the batch configuration process, is performed
by the IBM System Support Representative (SSR) on the ESS service console using a tool
called the ESS Batch Configuration Tool. It is a simplified configuration method, which
provides support for a set of common configurations and standard volume types. The SSR
may in some cases use the ESS Specialist to complete the standard configuration process.
For the IBM SSR to perform the standard configuration process, you need to provide him with
the necessary configuration information in the form of worksheets. See 8.3.2, “Configuration
worksheets” on page 116.
The Batch Configuration Tool is designed to perform the initial configuration of the ESS.
Subsequent changes to the configuration should be done with the ESS Specialist although
the ESS Batch Configuration Tool can be used for previously unallocated (unassigned)
8-packs on an SSA adapter card, for example when capacity is added to the ESS.
The ESS Batch Configuration Tool is only available to the IBM SSR. The ESS Specialist is
available for anyone to use.
8.3.1 Selecting the configuration method
The standard configuration process can be used if the target logical configuration consists of
the following elements, which are supported by the Batch Configuration Tool:

Disk groups defined as RAID-5 arrays

3990-6 type control unit images

Standard 3390-3 or 3390-9 volumes

3390 track format volumes only
Host I/O configuration ESS unit addresses
CNTLUNIT UNIT=2105,UNITADD=((00,156)),...
IODEVICE UNIT=3390B,ADDRESS=(9000,52),UNITADD=00,...
IODEVICE UNIT=3390A,ADDRESS=(9034,104),UNITADD=34,...
0x00-33 Base (52)
0x34-9B Alias (104)

Chapter 8. CKD storage configuration
115
The Batch Configuration Tool creates LCUs, assigns RAID arrays to the LCUs, and
auto-allocates standard volumes in the RAID arrays. All the volumes in a LCU will be of the
same type, while different LCUs may contain different volume types.
As part of the standard configuration process, the IBM SSR may use the ESS Specialist to
complete the configuration with items that the Batch Configuration tool does not support:

Add in each array full-size custom volumes of the same type as the standard volumes

Add a selected number of PAV alias addresses to each configured base device
Table 8-9 summarizes the type and number of logical volumes that are supported by the
standard configuration process. In a table entry of format X+Y, X is the number of standard
volumes allocated on the array, and Y the number custom volumes of the same size that may
optionally be allocated.
Table 8-9 Number of volumes supported by the standard configuration process
If your target configuration is supported by the Batch Configuration Tool, we recommend that
the ESS be configured using the standard configuration process. It saves you time and
creates a balanced configuration. If the standard process meets your requirements except for
the last two items, the custom volumes and PAVs, (for example, you want different size
custom volumes), you can still use the standard process - simply leave the custom volumes
and PAVs out of the batch configuration process, and later configure them yourself using the
ESS Specialist.
In this redbook we do not describe the Batch Configuration Tool in detail, because this tool is
only available to the IBM SSR. All you need to do is to complete configuration worksheets
which describe your planned configuration. The IBM SSR will then use the worksheets to
configure the ESS.
On the worksheets you essentially specify:

The number of LCUs to configure.

The volume type (3390-3 or 3390-9) on each LCU

Whether to add custom volumes or not

The number of PAV aliases per base
If the standard configuration process does not meet your configuration requirements, use the
custom configuration method instead. Custom configuration is required, for example, if you
want to configure:

Other than 3990-6 type control unit images

3390-2 volumes

3380 track format volumes

Different types of volumes on one LCU

Custom volumes of non-standard size

Large volumes with more than 10017 cylinders.
You can configure the ESS yourself using the ESS Specialist or have an IBM representative
do it for you. In this redbook we look in more detail on how to use the ESS Specialist to create
custom configurations.
9 GB disk array 18 GB disk array 36 GB disk array
6+P+S 7+P 6+P+S 7+P 6+P+S 7+P
3390-3 16+2 16+5 32+4 40+2 68+5 80+5
3390-9 4+2 4+3 8+4 12+2 20+4 24+4

116

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
8.3.2 Configuration worksheets
ESS configuration worksheets are intended to help you plan and document your ESS logical
configuration. The worksheets, together with instructions, are documented in the
IBM
TotalStorage Enterprise Storage Server Configuration Planner,
SC26-7353 which you can
download from the Internet at the ESS Web site:
http://www.storage.ibm.com/hardsoft/products/ess/ess.htm
Select
Reference information
at the bottom of the page (see 1.5.3, “ESS documentation
Web site” on page 10 for information on how to access the ESS documentation). Then refer to
the manual for information on how to fill in the worksheets.
The CKD storage worksheets are:

Common worksheets:
– S/390 Storage requirements worksheet
– S/390 Add custom volumes worksheet
Use these worksheets to identify your requirements for ESS storage. You should always fill
these in prior to the installation. Once you have completed the applicable common
worksheets, give them to either your IBM representative or to the help-line personnel.
They use the information recorded in the custom worksheets, and a copy of the ESS order
form to decide which configuration method to use.

Worksheets for standard configuration process:
– S/390 Batch Configuration worksheets for different capacities
Based on the information on the common worksheets, the IBM representative or IBM
help-line will generate Batch Configuration worksheets that are used by the IBM SSR
during the installation and configuration of the ESS. These worksheets are only required
for the standard configuration process.

Worksheets for custom configuration — these worksheets help you plan and perform an
ESS custom configuration:
– Define LCUs and PAVs: Use this worksheet to document the LCUs.
– Define Disk Groups: Use this worksheet to document RAID or non-RAID disk groups
and standard volumes.
– Add Custom Volumes: Optionally use this worksheet to document custom volumes.
– Configure Fibre Channel Ports: Optionally use this worksheet to document Fibre
Channel ports for FICON attachment.
You can also fill your configuration information online. Go to the above ESS documentation
Web site and select
ESS Configuration Worksheets.
On the Web page you can fill in
essentially the same information as on the common worksheets. Once complete, the
information is submitted to IBM Technical Support Help Line for verification.
Optionally, you can request the Help Line to contact you for assistance. The help line will
process the worksheet you submitted and return to you via e-mail a completed copy of S/390
Storage requirements table and a completed Batch Configuration worksheet that corresponds
to your ESS. Provide the completed worksheets to the IBM SSR for use during the installation
of the ESS.

Chapter 8. CKD storage configuration
117
8.3.3 Communication Resources worksheet
The Communication Resources worksheet allows you to define how various communication
functions within the ESS should be configured. IBM service personnel use the information
from the worksheet to establish these communication links during ESS initial installation. The
worksheet includes such configuration items as TCP/IP addresses for the ESS components,
call home and remote support options, modem settings, and telephone numbers.
The worksheet can be found in the
IBM TotalStorage Enterprise Storage Server Introduction
and Planning Guide,
GC26-7294
.
There are two items on the Communication Resources worksheet that are specific to zSeries
hosts, namely the level of Service Information Message (SIM) reporting, and the use of
Control Unit Initiated Reconfiguration
(
CUIR).

Service information messages (SIMs) for S/390:
You select the number of times that the ESS sends a SIM to S/390 or zSeries hosts for
entry in the Error Recording Data Set (ERDS). You also select the severity of SIMs that
are to display a message on the host console. The ESS sends SIMs to host consoles for
the following three types of SIMs (default reporting level shown in parenthesis):
DASD SIM
This SIM tracks disk drive module (DDM) failures and problems
(default NONE).
Media SIM
This SIM tracks data check problems on the media (default
ACUTE).
Storage server SIM
This SIM tracks control-unit, power-control, and other hardware
problems (default ACUTE).
You select for each SIM type the severity level of the SIMs the ESS sends to your console.
SIMs that are below the severity level that you select do not display a console message.
You also select the number of times that the ESS sends a SIM to the host. The ESS can
send a SIM 1-5 times at eight-hour intervals or not at all (0). The severity levels are:
Selection:Severity reported:
acute
A major subsystem resource is disabled. Performance may be
severely degraded. System and application outages may have
occurred.
serious
An unrecoverable error or a data check with loss of access to data.
moderate
Performance degradation is possible in a heavily loaded
environment. A primary subsystem resource is disabled.
Significant performance degradation is possible.
service
No system or application performance degradation is expected in
any environment.
none
None
See 11.6.3, “IEA480E SIM message” on page 194 for an example of the SIM message.

Allow CUIR to automatically vary paths off and on:
Check
Enable
to allow the ESS to initiate the reconfiguration for service. The ESS will
request the attached hosts to automatically vary paths offline for service, and back online
after the service is complete.
Check
Disable
(default) to set the paths to the ESS cluster offline for service. The system
operator has to manually vary the affected paths offline, and back online.

118

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
With CUIR enabled, the ESS can request that an operating system verify that one or more
subsystem resources can be taken offline for service. The ESS uses this process to
automatically vary channel paths offline and online to facilitate bay service or concurrent
code installation.
See 11.2.2, “Control Unit Initiated Reconfiguration (CUIR)” on page 178.
8.4 ESS custom configuration for CKD storage
In this section we describe ESS Specialist tasks and panels used for performing a custom
configuration of the ESS for CKD storage. For detailed information on how to use the ESS
Specialist, refer to
IBM TotalStorage Enterprise Storge Server Web Interface User’s Guide,
SC26-7346.
8.4.1 Tasks
Configuration of the ESS for use with zSeries systems involves the following main tasks which
are performed using the ESS Specialist:
1.Configure logical control units (LCUs)
Define the LCUs for your configuration. Select the control unit emulation mode, specify the
Subsystem Identifier (SSID), and enable or disable Parallel Access Volumes (PAVs) for
each LCU.
2.Configure disk groups for each LCU
Assign one or more disk groups to an LCU. Select the storage type (RAID-5 or non-RAID)
and track format (3390 or 3380) of the ranks. Optionally, auto-allocate a group standard
volumes on a RAID array.
3.Add logical volumes for each LCU
Allocate logical volumes on the disk ranks you defined previously. The volumes you add
are custom volumes, that is, you can specify the size of each volume individually.
4.Configure PAV alias devices for each LCU
If your system supports PAVs, configure PAV aliases for the logical volumes in the LCU.
You can configure multiple aliases for each base device.
5.Configure FICON ports
Configure FICON ports for use on zSeries systems. This task is not required for ESCON
ports.
During initial installation, tasks 1 to 4 are performed for each of the up to 16 LCUs, generally
in this order, although in large configurations you may need to repeat steps 2 and 3 multiple
times for an LCU. Task 5 needs to be done only once for the ESS. Later, you can modify the
configuration.
8.4.2 ESS Specialist panels for CKD storage allocation
The following ESS Specialist panels are used for configuring CKD storage on the ESS:
Storage Allocation
This panel gives you a graphical view of disk rank allocation status.
From this panel you enter either the S/390 Storage or the Open System
Storage configuration functions. Enter this panel from the Navigation
frame.

Chapter 8. CKD storage configuration
119
S/390 Storage
Use this panel to display your CKD configuration and to access the
following panels and their associated configuration tasks. This is the
main CKD storage configuration panel. From this, you enter the specific
task panels, which are used to:
Configure LCUs
Configure the LCUs.
Configure Disk Groups
Allocate physical disk capacity for an
LCU.
Add Volumes
Define custom volumes for an LCU.
Modify PAV Assignments
Define and delete PAV aliases.
Configure Host Adapter Ports
Configure FICON host adapter ports.
Figure 8-4 is a graphical presentation of the CKD storage configuration panels and the
functions you can select on them.
Figure 8-4 ESS Specialist: S/390 panels
The
Navigation frame
is visible on each panel. On most screen images in this chapter we do
not show the Navigation frame. It is shown in Figure 8-7 on page 122.
On each of the task panels, you enter information, set attributes, and make selections
required to define your configuration. When you have made the necessary definitions,
perform the actual configuration action by clicking the
Perform Configuration Update
button, or cancel it by clicking
Cancel Configuration Update
. In either case, once the action
is complete, you will return to the
S/390 Storage
panel, where you can select the next task.
To get back to the
Storage Allocation - Graphical View
panel, click the
Storage Allocation

button on the Navigation frame.
Status
Problem
Notification
Communications
Storage
Allocation
Users
Licensed
Internal
Code
Storage Allocation -- Graphical View
Open System Storage
S/390 Storage
S/390 Storage
Configure Disk Groups
Configure LCU
Add Volumes
Configure PAVs
Configure Host Adapter Ports
Configure LCU
Cancel Configuration Update
Perform Configuration Update
Configure Disk Groups
Cancel Configuration Update
Perform Configuration Update
Add Volumes
Cancel Configuration Update
Perform Configuration Update
Configure Host Adapter Ports
Cancel Configuration Update
Perform Configuration Update
Modify PAV Assignments
Cancel Configuration Update
Perform Configuration Update
Navigation Frame

120

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
8.4.3 ESS Specialist messages
During configuration procedures, the ESS Specialist will present messages in the Message
frame at the bottom of the screen. The two most commonly displayed messages are:

2108:
Retrieving the machine configuration data ... please wait.

1533:
Downlevel Data ... please wait a few minutes and try again.
ESS Specialist allows more than one user to be logged on simultaneously. Users may be
logged onto ESS Specialist in the other cluster. To ensure it has accurate configuration data,
ESS Specialist needs to read the current configuration of the machine. Message 2108 usually
appears when a function is selected on one of the panels.
You see message 1533 after a configuration change ends with a successful completion
message, and you click
OK
to return to the
S/390 Storage
panel. After message 1533 is
displayed, click the
Refresh Data
button on this panel; the panel will not refresh automatically.
After clicking the
Refresh Data
button, you will receive this message:

2122:
Retrieving the current configuration data ... please wait.
If this is again followed by message 1533, wait a few minutes while the ESS Specialist is
synchronizing configuration data in the two clusters, then re-try.
ESS Specialist also displays messages and status information in separate windows. When
you perform a configuration action that is expected to take longer than 30 seconds to
complete, the ESS Specialist displays a window showing the estimated remaining time for the
action (Figure 8-5):
Figure 8-5 ESS Specialist: Estimated time
If you click the
Continue
button, the window closes, and you can continue working with the
ESS Specialist. The actual configuration action continues in the background in the ESS.
However, you cannot initiate another configuration action until the current one is complete.
Note:
The ESS does not allow configuration changes to be performed using the ESS
Specialist while a service representative is logged on the service console. If you get error
message 1109 (see Figure 8-6) indicating that ESS components are in service mode, this
is probably the cause.

Chapter 8. CKD storage configuration
121
Figure 8-6 ESS Specialist: Error 1109
8.4.4 Storage Allocation panel
Begin ESS configuration by connecting to the ESS Specialist. When connecting, you may be
presented with a New Site Certificate window. View the series of certificates, then select ESS
Specialist on the Navigation frame. The Username and Password Required panel is
displayed. Proceed to enter a user ID with configuration authority and the password. Refer to
Chapter 5, “IBM TotalStorage Enterprise Storage Server Specialist” on page 79 for
instructions on how to start an ESS Specialist session.
Once connected, the
ESS Specialist Welcome
screen appears on a separate browser
window. Select
Storage Allocation
from the Navigation frame to start the storage
configuration function.
You should now see the
Storage Allocation - Graphical View
panel (Figure 8-7). It gives you
an overall view of the ESS configuration, showing defined host systems, installed host
adapters, disk ranks, and a storage allocation summary.
You do not deed to define zSeries hosts on the ESS Specialist. When ESCON or FICON host
adapters are installed on the ESS, a pre-defined EsconNet or FiconNet icon automatically
appears on the panel. The icons are pseudo-hosts representing all zSeries hosts attached
through ESCON or FICON adapters, respectively. The icons appear even if there is no
physical connection between the hosts and the ESS.
Unlike with FB storage, you do not need to assign CKD logical volumes to a particular port or
host. Every ESCON and FICON port in the ESS is potentially able to access any CKD logical
volume (up to 256 per LCU) in any of the defined LCUs (up to 16). Which LCUs and logical
volumes a particular host channel will actually access depends on the host I/O configuration.
The rectangles in the middle represent configured disk groups. If you click the EsconNet or
FiconNet icon, CKD disk groups display colored and with thick frames, and the ESCON or
FICON adapter ports highlight in yellow. This indicates that all installed ESCON/FICON ports
are automatically associated with the EsconNet/FiconNet host. The empty rectangles in this
case represent FB storage.

122

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Figure 8-7 ESS Specialist: Storage Allocation - Graphical View panel
On the
Storage Allocation
panel, you can select Tabular View. This displays a table showing
all the configured logical volumes.
On the
Storage Allocation
panel, click the
S/390 Storage
button to enter the storage
configuration functions for ESCON and FICON attached hosts. The
Open System Storage

button will take you to SCSI and FCP attached Fixed Block storage configuration functions.
8.4.5 S/390 Storage panel
The
S/390 Storage
panel (Figure 8-8) is the main panel for allocating CKD storage. From it,
you enter the specific configuration tasks, and once the task is complete, return back to it.
To display the
S/390 Storage
panel, click
S/390 Storage
on the
Storage Allocation - Graphical
View
panel.
The
Logical Control Units
table on the panel shows the LCUs for the ESS. Initially, LCUs 000
to 007 are visible, all undefined. The upper LCUs 008 to 00F are not shown until you have
assigned a disk group to the corresponding lower LCU on a device adapter.

Chapter 8. CKD storage configuration
123
Figure 8-8 shows a partially configured system, with some LCUs already configured, some
still undefined.
Figure 8-8 ESS Specialist: S/390 Storage panel
When you click a row on the
LCUs
table to select an LCU, the underneath
LCU Devices
table
will show the currently defined logical devices for the LCU (Figure 8-9). If no devices have yet
been configured for the LCU, the
LCU Devices
table is empty.
Figure 8-9 ESS Specialist: S/390 Storage panel with devices

124

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
In the
Logical Control Units (LCUs)
table:

The
Location
column shows the cluster and device adapter where the LCU resides. Refer
to Figure 8-1 on page 103 for the correspondence between LCUs, device adapters, and
clusters.

The
Storage Summary
column shows how many disk groups have been assigned to the
LCU, and how many unassigned groups are still available for the LCU.

The
Free Space
column shows the number of cylinders that are available in the LCU for
allocating new logical volumes.
In the
LCU Devices
table:

The
Device ID
column shows the unit address of the logical device. Once you have
initialized the logical volumes using ICKDSF, it will also display the volume serial number
(VOLSER).

The
Capacity
column shows the size of the logical volume (GB = 1,000,000,000 bytes).
The figure is slightly higher than the number of cylinders multiplied by the effective 3390 or
3380 cylinder size that is normally used in capacity calculations. The figure given here
includes fields of the CKD format track that do not account for the usable capacity.

The
Location
column shows the SSA loop and disk rank where the volume is allocated,
along with a sequential number within the rank. For alias devices, it shows where the
corresponding base device is located.
To start a configuration action, select in the
Logical Control Units (LCUs)
table the LCU that
you want to work with. The row will highlight in grey. Then click one of the function buttons in
the lower part of the panel. You need to select the LCU first for all other functions except for
the Configure Host Adapter Ports function.
If you want to return to the
Storage Allocation
panel from here, click the
Storage Allocation

button in the Navigation frame (not shown in the figure).
8.4.6 Configuring logical control units
Use the
Configure LCU
panel to define and modify Logical Control Units. On the
S/390
Storage
panel, select the LCU you want to work with. It will highlight in grey. Then click
Configure LCU
to display the
Configure LCU
panel for the selected LCU (Figure 8-10).

Chapter 8. CKD storage configuration
125
Figure 8-10 ESS Specialist: Configure LCU panel
Defining an LCU
These are the steps to be performed in defining an LCU:
1.Select the
Logical Control Unit Emulation Mode
. The options are:
– 3990-6
– 3990-3
– 3990-3 TPF
– 3990 (Siemens BS2000 Compatible).
Select 3990-6 unless you have a specific reason, such as a software requirement, to
select one of the other options. This is the only entry field displayed for an undefined LCU.
Once you select the mode, other fields appear on the panel.
2.Set the
Subsystem Identifier (SSID)
for the LCU. The SSID is a four digit hexadecimal
number in the range of 0x0001 to 0xFEFF (it must not have FF as the first two digits). The
SSID must be unique in the location, no other storage subsystem should have the same
SSID.
A practical convention is to use the first two digits to identify the ESS subsystem, and the
two last to indicate the logical subsystem (00-0F). This is useful for PPRC, for example, as
the commands to establish PPRC paths and pairs take the SSID and the two digit LSS
number as parameters.
Note:
the IBM TotalStorage Enterprise Storage Server is qualified for the Operating
System BS2000/OSD V4, allowing the integration of the ESS in solutions with
Fujitsu-Siemens servers. ESCON is the only attachment supported currently. An RPQ
must be submitted for getting IBM approval before attachment.

126

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
3.Set the
Parallel Access Volumes (PAV)
attribute to enabled or disabled. Select enabled if
your system supports Parallel Access Volumes. If the PAV feature is not installed, you can
only select disabled. If you add the PAV feature to the ESS later, you can change the PAV
attribute from disabled to enabled.
You must enable PAV here in order to be able to configure PAV aliases for the LCU later.
4.If you enabled PAV for the LCU, set the
Starting (highest) PAV Address
. The options are:
– 0 (0x00)
– 63 (0x3F)
– 127 (0x7F)
– 255 (0xFF).
The Starting PAV address controls where in the LCU address range the PAV alias devices
will be assigned when you configure them using the Configure PAVs function (discussed
later in 8.4.9, “Modify PAV assignments” on page 137).
Select starting address 0 to allocate the alias devices contiguously right after the base
devices. Select one of the other options only if you want to create an address layout where
the base and alias device address ranges are not contiguous.
PAVs are assigned the highest unused address available, beginning at the PAV starting
address, and proceeding in decreasing address order. Once all the addresses below the
Starting PAV address have been used, then addresses are assigned in ascending order
starting from the lowest unused address upwards, up to the limit of 256 addresses that are
available in an LCU. With the PAV starting address set to less than the total number of
bases and PAV aliases you will be defining, this results in a contiguous range of device
addresses for the LCU. Starting PAV address 0 always gives you a contiguous range.
You may refer to section 8.2, “Designing the LCU address layout” on page 112 for a
detailed discussion on PAV addresses.
5.When all entries are complete, click
Perform Configuration Update
, to apply the
configuration step for the LCU.
You do not have to define all the LCUs that are available. In smaller configurations you may
only need four or eight LCUs. If the storage capacity on some DAs is intended only for open
systems use, you do not need to configure the LCUs that correspond to those DAs.
Figure 8-11 shows an example of how you would configure an LCU.

Chapter 8. CKD storage configuration
127
Figure 8-11 ESS Specialist: Procedure to configure an LCU

128

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
After successful configuration of an LCU, you will return to the
S/390 Storage
panel. The
Logical Control Units (LCUs)
table now shows the LCU as Defined, with the attributes you
selected. The Free Space is zero, until you configure disk groups for the LCU.
Modifying an LCU
You can change the following LCU attributes:

You can change the Starting PAV Address for an LCU, at any time.

You can change the Parallel Access Volumes attribute from disabled to enabled for a
configured LCU, but you cannot change it back to disabled. To disable PAV, you must first
undefine, then redefine the LCU.

You can change the Subsystem Identifier. However, you must first terminate host functions
that depend on the SSID. Also all drives must be varied offline before changing the SSID
so software can correctly rebuild their control blocks. For example, PPRC path and pair
definitions use the SSID to identify the source and target LCUs. In order to change an
SSID, you must first delete all PPRC pairs and paths defined for the LCU. Figure 8-12
shows the warning message you will get when trying to change the SSID.
Figure 8-12 ESS Specialist: Warning 1554
You cannot change the Control Unit emulation mode once the LCU has been configured. To
change the emulation mode you would have to first undefine the LCU.
Undefining an LCU
You can undefine an existing configured LCU by setting its Logical Control Unit Emulation
Mode

to
Undefined
on the
Configure LCU
panel. When you click
Perform Configuration
Update
, the warning message shown in Figure 8-13 is displayed.
Figure 8-13 ESS Specialist: Warning 1550
Note:
The screen images on this figure may slightly differ from those that you see when
using a later level ESS Specialist.

Chapter 8. CKD storage configuration
129
You would need to undefine an LCU in order to change its emulation mode.
8.4.7 Configure disk groups for an LCU
Use the
Configure Disk Groups
panel to assign disk groups (that is, physical storage
capacity) to logical control units, and to set the disk group attributes. On the LCU table on the
S/390 Storage
panel, select the LCU you want to add disk groups to. The Storage Summary
column of the table must indicate that disk groups are available. The LCU row will highlight in
grey. Click
Configure Disk Groups
. Panel shown in Figure 8-14 will appear.
You cannot add disk groups to an undefined LCU. You must first configure the LCU.
Figure 8-14 ESS Specialist: Configure Disk Groups panel
Attention:
If you click
Yes
, the LCU definition is deleted, with its associated ranks and
logical volumes. All data on the logical volumes in that LCU is discarded. Make sure that
you have backed up all data on the volumes before you delete an LCU.

130

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Defining RAID Storage
1.Select an undefined disk group to configure by clicking on the corresponding row in the
Available Storage
table. The row highlights in grey. You can select a disk group either in
loop A, or loop B.
The
even
numbered LCUs (which are located in Cluster 1) are associated with the
even

numbered disk groups (2,4,6) on an SSA loop. The
odd
numbered LCUs (in Cluster 2)
manage the
odd
numbered disk groups (1,3,5).
2.Select the
Storage Type
in the
Disk Group Attributes
section. The options are:
– RAID Array
– non-RAID
– Undefined (to undefine an previously configured group).
Once you change an attribute, the selected row in the
Available Storage
table will be
updated. The Modification column now shows Defined indicating that you have changed a
disk group attribute.
3.Select the
Track Format
for the array. The options are:
– 3390 (3380 track mode)
– 3390 (3390 track mode)
All logical volumes on the rank will be of the track format you select here.
4.In the
Standard volumes to auto-allocate
field, select the type of standard volumes you
want to auto-allocate on the array. You cannot select the number of volumes, just the type.
The number is predefined, and depends on the array capacity, track format, and volume
type. The example pull-down in Figure 8-14 shows the options for an 18GB 7+P array. See
Table 8-1 on page 108 for a summary of the number of standard volumes in each array.
When you auto-allocate volumes, two partitions are created in the array. The interleaved
partition contains the auto-allocated standard volumes and takes up most of the capacity
in the array. The second, non-interleaved partition will initially be empty. You can later add
custom volumes in it using the
Add Volumes
panel.
If you select
0 volumes
, no standard volumes will be auto-allocated, just the track format
of the array will be set. An interleaved partition is not created, and the non-interleaved
partition will cover the whole array. This option allows you to fill the array with custom
volumes.
5.If you have more than one disk group to configure, you may repeat steps 1 to 4.
6.Click
Perform Configuration Update
to apply the configuration changes you defined.
Figure 8-15 shows an example of how you would configure a disk group.

Chapter 8. CKD storage configuration
131
Figure 8-15 ESS Specialist: Procedure to configure disk groups
Note:
The screen images on this figure may slightly differ from those that you see when
using a later level ESS Specialist.

132

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
The volumes auto-allocated in each array will not appear on the
LCU Devices
table until
formatting of the disk group is complete. This may take more than an hour depending on the
array size. Several disk groups can be formatting simultaneously.
After successful configuration of a disk group, the
S/390 Storage
panel appears as in
Figure 8-9 on page 123. The auto-allocated volumes now appear in the LCU Devices table.
The Storage Summary and Free Space columns of the
Logical Control Units (LCUs)
table
have been updated to reflect the new configuration.
Defining non-RAID storage
1.Select a disk group in the
Available Storage
table (Figure 8-14).
2.Select
non-RAID
as the
Storage Type
. When you select non-RAID, the row in the
Available Storage
table expands to eight rows, one for each of the drives in the disk group.
The Storage Type field in the table is non-RAID. The Capacity field has the unformatted
capacity of a single disk drive.
3.The rows for the non-RAID disks can be individually selected in the
Available Storage

table. Select a disk to configure.
4.Select
3390
or
3380
for the
Track Format
.
The Standard volumes to auto-allocate attribute is set to zero. You cannot change it. No
volumes can be auto-allocated on a non-RAID rank. You need to allocate the volumes
later using the Add Volumes panel.
5.Repeat steps 3 and 4 for the other non-RAID disks.
6.Click
Perform Configuration Update
.
A non-RAID disk group may contain both CKD and FB format disks in any combination. If a
disk group is defined as non-RAID using this panel, then any disks in the group that are not
formatted as CKD (3390 or 3380 track format) are available to be formatted in FB track format
using the Open Systems Storage function. Similarly, if a disk group is defined as non-RAID
using the Open Systems Storage function, any disks in the group not formatted as FB appear
in the
Configure Disk Groups
panel for the LCU and are available to be formatted as CKD.
Undefining a RAID array
You can undefine an previously configured array. This will delete its contents and make the
disk group eligible for configuration with a different logical volume setup, or as non-RAID
storage.
1.Select the array in the
Available Storage
table.
2.Set the Storage Type attribute to
Undefined
. If there are volumes allocated on the group,
warning message in Figure 8-16 is displayed.

Chapter 8. CKD storage configuration
133
Figure 8-16 ESS Specialist: Warning 1982
When you click
Yes
, the
Storage Type
changes to
Undefined
, the
Track Format

changes to
None
(unused disk), and the Modification column in the
Available Storage

table will show
Undefined
. At this point you can still cancel the change. The actual
reconfiguration is not applied until you perform the configuration update.
3.If you want to undefine more arrays in the LCU, repeat steps 1 and 2.
4.Click
Perform Configuration Update
.
The reconfiguration is a time consuming action. The physical disks in the array are
reformatted and this may take 30 to 60 minutes, even longer to complete. The disk group will
not appear on the Configure Disk Groups panel until the background formatting process has
completed. The progress of the reformatting process is listed with a % complete indication on
the Information window on the ESS Specialist Status panel.
Undefining non-RAID ranks
You undefine an individual non-RAID rank by changing its Track Format to
None (unused
disk)
on the Configure Disk Groups panel. Once complete, the disk is again available to be
formatted as either CKD or FB rank.
On this panel you can only undefine CKD format non-RAID ranks. Use the Open System
Storage Configure Disk Groups panel to undefine FB format ranks.
In order to undefine a non-RAID group, you first have to undefine all individual disks, both
CKD and FB format disks, in the group by changing their track format to None (unused disk).
Once you have undefined the individual disks in the group, you can proceed to undefine the
non-RAID group itself. The procedure is the same as for undefining RAID arrays. You can
undefine a non-RAID group using either this panel or the corresponding Open Systems
Storage Configure Disk Groups panel.
Attention:
When you click
Perform Configuration Update
, the disk group will be
reconfigured. It is deleted from the LCU, with all its associated logical volumes. Any data
on those volumes is discarded. Make sure you have backed up all data on the volumes
before you delete an disk group.
Attention:
When you change the track format of a non-RAID disk to None (unused disk),
all data on the disk will be lost. Make sure you have appropriate backups.

134

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
8.4.8 Add volumes to an LCU
To allocate custom volumes, select an LCU on the
S/390 Storage panel
, then click the
Add
Volumes
button. You should then see the
Add Volumes
panel (Figure 8-17).
Figure 8-17 ESS Specialist: Add Volumes panel
1.In the
Volume Attributes
section, select the
Storage Type
and
Track Format
of the
volumes you want to add. This will automatically select the corresponding row in the
Available Capacity, per Volume Type table. Alternatively, select a row in the Av
ailable
Capacity, per Volume Type

table

that corresponds to the type of volume you want to add.
This will automatically set the
Storage Type
and
Track Format
fields in the
Volume
Attributes
section.
The
Available Capacity, per Volume Type
table shows the capacity that is available for
adding different types of volumes in the LCU, classified into four categories:
– RAID Array in 3390 track format
– RAID Array in 3380 track format
– Non-RAID in 3390 track format
– Non-RAID in 3380 track format
The total capacity of the volumes you add cannot exceed the available capacity. If Max
Available Capacity is 0 cylinders, you cannot add any volumes of that type.
If you auto-allocated 3390 track format volumes when you configured the disk group using
the
Configure Disk Groups
panel, the non-interleaved partition is available as RAID - 3390
format, and you can allocate only 3390 format custom devices in the array. The same is
true for 3380 track format. This is because a CKD RAID array is formatted entirely in a
single track format, either 3390, or 3380.
2.Set the size of the volumes you want to add by entering the number of cylinders in the
Cylinders
field. The maximum number you can set is 32760 if the system supports Large
Volumes, otherwise 10017, or the number of cylinders in the Max Contiguous Capacity
field in the
Available Capacity
table, whichever is smaller. If you have selected a category
where no capacity is available, you cannot add new volumes.

Chapter 8. CKD storage configuration
135
The volumes you add here are custom volumes, that is, you can specify any number of
cylinders from 1 to 10017 (or 1 to 32760 if Large Volume Support is installed). If you want
to emulate the standard capacities for real 3390 volumes, specify:
1113 for 3390 Model 1
2226 for 3390 Model 2
3339 for 3390 Model 3
10017 for 3390 Model 9
For real 3380 track format volumes, the number of cylinders per volume are:
885 for 3380-STD, 3380-D, or 3380-J
1770 for 3380-E
2655 for 3380-K
3.Set the number of volumes you want to add in the
Number of Volumes
field.
The available space for the category may be distributed over multiple ranks. Because a
logical volume cannot span ranks, the number of volumes of a given size you can add may
be less than the Max Available Capacity divided by the size of the volume. The number is
dictated by the size of the contiguous segments on the ranks. The remaining space can be
used for allocating smaller volumes. You can add volumes of different sizes in one
category of available capacity, until no usable space remains.
4.When you have defined the attributes of the volumes, click the
ADD>>
button to add the
volumes to the
New Volumes
list. You can repeat this several times, with different volume
sizes and numbers.
If the total capacity of volumes you want to add is too large for the available space, the
ESS Specialist displays error message 1901 (Figure 8-18).
Figure 8-18 ESS Specialist: Error 1901
The volumes that fit in the available space are added to the
New Volumes
list.
The Max Available Capacity and Max Contiguous Capacity fields are updated. The
number of cylinders subtracted from the available capacity may slightly exceed the total
number of cylinders in the volumes added. This is because of overheads associated with
each volume, for example, metadata. The Cylinders attribute field is updated to show the
largest single volume that can be added in the remaining space.
5.Prior to performing the configuration update, you can remove a volume that has previously
been added to the
New Volumes
list by selecting its row in the table and clicking the

<<Remove
button.
6.When you have defined the desired volumes, click
Perform Configuration Update
to
apply the change. Volumes on the
New Volumes
list will be configured.
Attention:
After creation, the logical volumes cannot be individually deleted. The only way
to delete a logical volume, or to change its type is to undefine the whole disk group where
the volume resides. When a disk group is undefined using the
Configure Disk Groups

panel, all the logical volumes on the group are deleted, and any data on them is discarded.

136

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Figure 8-19 shows an example of how you would add volumes to a disk group.
Figure 8-19 ESS Specialist: Procedure to add volumes
Note:
The screen images on this figure may slightly differ from those that you see when
using a later level ESS Specialist

Chapter 8. CKD storage configuration
137
Once the configuration action completes, the
S/390 Storage
panel will be displayed. Select
your LCU; the
LCU Device
table now shows the custom volumes you added.
The new volumes are assigned available device IDs (unit addresses) in the order the volumes
are created. The lowest available device ID in the LCU is used first. The device IDs map to the
device numbers as seen by the zSeries operating system.
8.4.9 Modify PAV assignments
After you have defined an LCU, configured one or more disk groups for it, and created base
volumes, you can assign PAV aliases to the volumes in the LCU.
This task is not necessary
for an LCU for which PAV has been disabled (on the
Configure LCU panel).
To assign PAVs, select the LCU on the
S/390 Storage
panel, then click
Configure PAVs
to
display the
Modify PAV Assignments
panel (Figure 8-20).
Figure 8-20 ESS Specialist: Configuring PAV aliases
1.On the LCU Devices table, select the devices you want to modify. To select rows, use the
scroll bar to make the desired device visible in the
LCU Devices
list, then click the row to
select it. It will highlight in grey. Additional rows can be selected using CTRL- left click. A
range of rows can be selected by clicking at one end of the range, then SHIFT- left clicking
at the other end. A row can be deselected using CTRL- left click.
2.Once you have selected the devices, select the action:

Add PAVs to each selected volume
Specify

the
Number of PAVs to add
for each selected volume. Note that you cannot
enter a non-zero value here if PAVs are disabled for the LCU. The maximum number is
limited by the total of 256 devices an LCU can support.

Delete selected PAVs
This will delete the selected alias devices.

138

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
3.Click
Perform Configuration Update
to modify the PAVs.
You can select any combination of base devices and assign any number of PAVs to them,
provided the total number of base devices and aliases for the LCU does not exceed 256. To
add different number of aliases for different base devices, you need to repeat steps 1-3. You
can assign more aliases to a base which already has aliases assigned. A single base can
have up to 255 aliases assigned.
The upper part of the
LCU Devices
table in the example in Figure 8-20 shows base devices to
which aliases have already been assigned. The Base/Alias column indicates how many
aliases are currently assigned to a base. The lower part of the list shows existing alias device
entries. For the PAV aliases, the Base/Alias column indicates the base device to which the
alias is currently assigned to. The device table can be viewed in either this panel, or the
S/390
Storage
panel.
Aliases are assigned unit addresses in a specific order. See details in 8.4.6, “Configuring
logical control units” on page 124.
When dynamic PAV support is enabled, WLM can reassign PAV aliases from one base to
another. The alias devices retain their unit addresses, just the assignment to bases will
change. With dynamic PAV the initial assignment of aliases to bases on the
Modify PAV
Assignments
panel essentially only serves as a means of allocating a pool of aliases which
the WLM can then work with and reassign to achieve the best performance.
WLM can only reassign aliases to base devices within the same LCU. You have to configure
aliases for each LCU.
8.4.10 Configure FICON ports
The same ESS Fibre Channel host adapters are used for FICON attachments on S/390 and
zSeries hosts or for open systems Fibre Channel protocol (FCP) attachments, depending on
how they are configured. You must configure the ports for the intended attachment type using
the
Configure Host Adapter Ports
panel (Figure 8-21).
You do not need to configure ESCON ports.

Chapter 8. CKD storage configuration
139
Figure 8-21 ESS Specialist: Configure Host Adapter Ports panel
1.Select a port to configure, either from the
Host Adapter Port
pull-down, or by clicking its
icon.
2.The
Fibre Channel Topology
field shows the current Fibre Channel topology for the port
you selected. If it is undefined, set the topology to
Point to Point (Switched Fabric)
. This
is the only choice when the panel is accessed from the
S/390 Storage
panel.
3.The
Fibre Channel Protocol
field shows the current Fibre Channel protocol for the port
you selected. If the topology is undefined, set it to
FICON (System/390)
.

For unconfigured
ports, this is the only choice if you accessed this panel from the
S/390 Storage
panel.
If the topology is defined, you must first change the setting to
Undefined
before the ESS
can make an alternate setting available for configuration.
4.Select the next port and repeat the previous steps. At any time, you can click
Reset
Selected Port
to cancel any pending configuration changes made to the currently
selected port.
5.When all ports have been defined, click
Perform Configuration Update
to apply or
Cancel Configuration Update
to cancel the configuration change for all modified ports.
The
Fibre Channel Access Mode
field in the
Storage Server Attributes
box shows the
current Fibre Channel access mode for the ESS, either
Access_Any
or
Access_Restricted
.
The mode is initially set during installation of the ESS, and can only be changed by a service
representative during a service action. The mode only applies to open systems FCP
connections, it is not relevant for FICON.
The World-Wide Port Name field shows the WWPN of the ESS host adapter port.

140

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
8.4.11 A step-by-step example
In this section we provide an example of how to configure CKD storage on a newly installed
ESS subsystem using the ESS Specialist. Our example ESS is a mid-size FICON attached
subsystem with 28 36GB 8-packs for a 6.3 TB total capacity. In this configuration two SSA
loops only have two 8-packs, the others have four. As a result, some LCUs will have less
capacity than the others.
Design guidelines for the target logical configuration:

Configure 16 LCUs.

Configure disk groups as RAID arrays.

Assign two arrays to LCUs 0-B, one array to the remaining four LCUs.

Assign one smaller (6+P) and one larger (7+P) array for LCUs 0-B for them to have equal
capacities.

Fully populate the arrays with 3390-9 logical volumes..

Configure three aliases per each base.

Use Starting PAV Address of 255 to assign the aliases at the high end address range.
Before starting to work with the ESS Specialist, complete the S/390 Storage requirements
worksheet in the
IBM TotalStorage Enterprise Storage Server Configuration Planner,
SC26-7353, or fill in the information online at the ESS Web site. In the sample worksheet
(Table 8-10) we have added information for the host I/O configuration.
Table 8-10 S/390 storage requirements worksheet
Logical control unit 0 1 2 3 4 5 6 7
Cluster-Device Adapter 1-1 2-1 1-2 2-2 1-3 2-3 1-4 2-4
SSID 2200 2201 2202 2203 2204 2205 2206 2207
LCU emulation type 3990-6 3990-6 3990-6 3990-6 3990-6 3990-6 3990-6 3990-6
Track mode for LCU 3390 3390 3390 3390 3390 3390 3390 3390
Device type for LCU 3390-9 3390-9 3390-9 3390-9 3390-9 3390-9 3390-9 3390-9
Number of base devices 44+8 44+8 44+8 44+8 44+8 44+8 44+8 44+8
Number of PAV per base 3 3 3 3 3 3 3 3
Starting PAV address 255 255 255 255 255 255 255 255
Base device unit addr 0x00-33 0x00-33 0x00-33 0x00-33 0x00-33 0x00-33 0x00-33 0x00-33
Alias device unit addr 0x64-FF 0x64-FF 0x64-FF 0x64-FF 0x64-FF 0x64-FF 0x64-FF 0x64-FF
IOCP/HCD information
CUNUMBR 900 910 920 930 940 950 960 970
CUADD 0 1 2 3 4 5 6 7
ADDRESS (3390B) 9000,80 9100,80 9200,80 9300,80 9400,80 9500,80 9600,80 9700,80
ADDRESS (3390A) 9050,176 9150,176 9250,176 9350,176 9450,176 9550,176 9650,176 9750,176

Chapter 8. CKD storage configuration
141
Notice that on the host we define full 256 devices for each LCU, even though we do not
configure them all on the ESS. Figure 8-22 shows the layout of the disk groups. In the array
boxes, the first number is the number of logical 3390-9 volumes in the array, followed by their
unit addresses.
Figure 8-22 Sample configuration
Logical control units 8 9 A B C D E F
Cluster-Device Adapter 1-1 2-1 1-2 2-2 1-3 2-3 1-4 2-4
SSID 2208 2209 220A 220B 220C 220D 220E 220F
LCU emulation type 3990-6 3990-6 3990-6 3990-6 3990-6 3990-6 3990-6 3990-6
Track mode for LCU 3390 3390 3390 3390 3390 3390 3390 3390
Device type for LCU 3390-9 3390-9 3390-9 3390-9 3390-9 3390-9 3390-9 3390-9
Number of base devices 44+8 44+8 44+8 44+8 20+4 20+4 20+4 20+4
Number of PAV per base 3 3 3 3 3 3 3 3
Starting PAV address 255 255 255 255 255 255 255 255
Base device unit addr 0x00-33 0x00-33 0x00-33 0x00-33 0x00-17 0x00-17 0x00-17 0x00-17
Alias device unit addr 0x64-FF 0x64-FF 0x64-FF 0x64-FF 0xB8-FF 0xB8-FF 0xB8-FF 0xB8-FF
IOCP/HCD information
CUNUMBR 980 990 9A0 9B0 9C0 9D0 9E0 9F0
CUADD 8 9 A B C D E F
ADDRESS (3390B) 9800,80 9900,80 9A00,80 9B00,80 9C00,80 9D00,80 9E00,80 9F00,80
ADDRESS (3390A) 9850,176 9950,176 9A50,176 9B50,176 9C50,176 9D50,176 9E50,176 9F50,176
Cluster 1
1
A
B
08
2
A
B
0A
3
A
B
0C
4
A
B
0E
Array 2 Array 4
Cluster 2
1
A
B
09
2
A
B
0B
3
A
B
0D
4
A
B
0F
Array 3Array 1
00
02
04
06
01
03
05
07
24 x'00-17'
28 x'18-33'
LCU LCU
24 x'00-17'
24 x'00-17'
24 x'00-17'
24 x'00-17'
24 x'00-17'
24 x'00-17'
24 x'00-17'
28 x'18-33'
28 x'18-33'
28 x'18-33'
28 x'18-33'
28 x'18-33'
24 x'00-17'
24 x'00-17'
24 x'00-17'
24 x'00-17'
24 x'00-17'
24 x'00-17'
28 x'18-33'
28 x'18-33'
28 x'18-33'
28 x'18-33'
28 x'18-33'
24 x'00-17'
28 x'18-33'
24 x'00-17'

142

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
This is the procedure:
1.Configure LCUs 0 to 7.
Start by configuring the first eight LCUs using the
Configure LCU
panel. Initially only the
first eight LCU are displayed on the panel. Select 3990-6 emulation mode, give the SSID,
and select Starting PAV Address 255.
2.Configure the first disk group for LCUs 0 to 7.
Configure Disk Group 2 for the even LCUs (0,2,4,6) in Cluster 1, and Disk Group 1 for the
odd LCUs (1,3,5,7) in Cluster 2 using the
Configure Disk Groups
panel. Groups 1 and 2
have smaller capacities than Groups 3 and 4, which helps to locate and select them on the
panel. In each LCU you have two Group1s or Group2s to select from, one in each SSA
loop. Select the disk group for LCUs 0-3 from loop A, for LCUs 4-7 from loop B. Select a
group from loop B here first because we want LCUs 4-7 to have their two arrays from
different loops.
Configure the disk groups as RAID arrays, in 3390 track format, and auto-allocate 20 type
3390-9 standard volumes on each. These will all be 6+P+S arrays.
3.Configure LCUs 8 to F.
Once you have assigned the first disk group to LCUs 0-7, LCUs 8-F appear on the
Configure LCU
panel and become configurable. Define them as in step 1.
4.Configure the first disk group for LCUs 8 to F.
Configure Disk Group 2 for the even LCUs (8,A,C,E) in Cluster 1, and Disk Group 1 for the
odd LCUs (9,B,D,F) in Cluster 2 using the
Configure Disk Groups
panel. At this point you
only have one Group1 or Group2 to select from.
Configure the disk groups as RAID arrays, in 3390 track format, and auto-allocate 20 type
3390-9 standard volumes on each.
5.Add custom volumes for LCUs 0 to F.
One array has now been assigned for each LCU with 20 standard volumes on them. In
each array there is still room for four full-size 3390-9 custom volumes. Add them now in
each LCU using the
Add Volumes
panel. Select volume size of 10017 cylinders.
6.Configure the second disk group for LCUs 0 to 3.
Configure Disk Group 4 for LCUs 0 and 2 in Cluster 1, and Disk Group 3 for LCUs 1 and 3
in Cluster 2 using the
Configure Disk Groups
panel. In these LCUs you have two Group3s
or Group4s to select from, one in each SSA loop. For these LCUs, select the disk group
from loop B, that is, from the other loop than the first disk group in the LCU. This will
provide maximum SSA loop bandwidth for the LCU.
Configure the disk groups as RAID arrays, in 3390 track format, and auto-allocate 24 type
3390-9 standard volumes on each.
7.Configure the second disk group for LCUs 4 to B.
Configure Disk Group 4 for the even LCUs (4,6,8,A) in Cluster 1, and Disk Group 3 for the
odd LCUs (5,7,9,B) in Cluster 2 using the
Configure Disk Groups
panel. At this point you
only have one disk group in each LCU to select from.
Configure the disk groups as RAID arrays, in 3390 track format, and auto-allocate 24 type
3390-9 standard volumes on each.
All disk groups have now been configured. LCUs C-F will not get a second array assigned.

Chapter 8. CKD storage configuration
143
8.Add custom volumes for LCUs 0 to B.
In steps 6 and 7 you just added a second array with 24 standard volumes for LCUs 0-B. In
each of these arrays there is room for four full-size 3390-9 custom volumes. Add them now
using the
Add Volumes
panel. Select volume size of 10017 cylinders.
9.Configure PAVs for all LCUs.
All 16 LCUs are now fully configured with logical volumes. LCUs 0-B have 52 logical
volumes, LCUs C-F have 24 volumes. Now assign three PAV aliases to each base
address (there would be room for even more) using the
Modify PAV Assignments
panel.
10.Configure FICON ports.
Finally, if not already done so by the SSR, configure the FICON host adapter ports using
the
Configure Host Adapter Ports
panel.
Our experience is that you need roughly two working days to perform this configuration
process. Step 9 alone takes about four hours (about 10 PAVs per minute on an ESS model
F20 at G4 LIC level).
Except for the last step, this example is fully applicable to ESCON connected systems also.
However some considerations apply. If you were to use ESCON channels to connect this
system, you should preferably have at least 16 ESCON channels, configured in four path
groups with four LCUs on each channel. With less than 16 ESCON channels, you probably
need to configure more LCUs on each channel and reduce the number of alias devices in
each LCU to keep the total number of devices on a channel below 1024. For example, if each
channel was to address eight LCUs, you would have 360 base devices (6 x 52 + 2 x 24) on a
channel, leaving 664 addresses available for aliases, less than two aliases per base.
Example 8-1 shows part of the
DISPLAY M=CHP
command output for this configuration.
Example 8-1 DISPLAY M=CHP output
D M=CHP(91)
IEE174I 11.07.04 DISPLAY M 790
CHPID 91: TYPE=1A, DESC=FICON POINT TO POINT, ONLINE
DEVICE STATUS FOR CHANNEL PATH C1
0 1 2 3 4 5 6 7 8 9 A B C D E F
900 + + + + + + + + + + + + + + + +
901 + + + + + + + + + + + + + + + +
902 + + + + + + + + + + + + + + + +
903 + + + + $@ $@ $@ $@ $@ $@ $@ $@ $@ $@ $@ $@
904 $@ $@ $@ $@ $@ $@ $@ $@ $@ $@ $@ $@ $@ $@ $@ $@
905 UL UL UL UL UL UL UL UL UL UL UL UL UL UL UL UL
906 UL UL UL UL AL AL AL AL AL AL AL AL AL AL AL AL
907 AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL
908 AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL
909 AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL
90A AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL
90B AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL
90C AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL
90D AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL
90E AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL
90F AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL
************************ SYMBOL EXPLANATIONS ************************
+ ONLINE @ PATH NOT VALIDATED - OFFLINE . DOES NOT EXIST
* PHYSICALLY ONLINE $ PATH NOT OPERATIONAL
BX DEVICE IS BOXED SN SUBCHANNEL NOT AVAILABLE
DN DEVICE NOT AVAILABLE PE SUBCHANNEL IN PERMANENT ERROR
AL DEVICE IS AN ALIAS UL DEVICE IS AN UNBOUND ALIAS

144

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment

© Copyright IBM Corp. 2002
145
Chapter 9.
S/390 and zSeries host setup
tasks
This chapter examines the S/390 and zSeries hosts configuration tasks that you will need to
perform to successfully implement the ESS. We describe the IOCP/HCD definitions required
to set up the host. We also discuss such topics as how to implement the dynamic alias
management function and the ESS Large Volume Support (LVS).
9

146

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
9.1 Preparation and considerations
It is recommended that before commencing these tasks the internal ESS logical setup, or a
logical configuration plan, be completed since many of the parameter values required to
complete the host definitions are dependent on the ESS logical configuration. The IOCP/HCD
definitions and ESS logical configuration must match each other. See section Section 8.2,
“Designing the LCU address layout” on page 112 for a discussion on how to design the ESS
configuration.
It may be an advantage to have network access to the ESS and an ESS Specialist user id
with “View” authority so you can more easily cross check the ESS internal configuration
settings with your I/O definitions.
When coding your I/O definitions, you will be defining to the host multiple logical control units
(LCUs) and logical devices rather than defining the physical 2105 machine configuration.
Due to the hardware internal switching functions, the enhanced recoverability, and internal
logical structure within the ESS, you only need to code in HCD/IOCP one 2105/3990
CNTLUNIT statement per LCU rather than the 3990 convention of defining two CU images.
Even though the ESS can support 4096 devices, 16 CU x 256 devices, be aware that an
ESCON port can only support 1024 devices, the equivalent of 4 CU x 256, or 8 CU x 128. A
FICON port, on the other hand can address 16,384 devices, making it possible for a FICON
channel to address all the 4096 devices you can have on an ESS. This allows you to logically
“daisy chain” all 16 LCUs in the ESS on one FICON channel.
Depending on your zSeries operating system and the levels of software, you may be
restricted to a subset of available new functions. See Chapter 7, “zSeries systems support”
on page 95 to review the exploitation levels.
9.2 OS/390 software configuration
This section describes how to implement the ESS in the z/OS and OS/390 systems.
9.2.1 IOCP/HCD
For the currently supported releases of OS/390 and z/OS the ESS runs in what is called
Exploitation mode (See Section 7.1.1, “OS/390 and z/OS support” on page 96).
In Exploitation mode the z/OS I/O subsystem recognizes the new ESS control unit type 2105.
Up to 16 logical control units (CNTLUNIT UNIT=2105) can be defined per ESS. The host
recognizes the ESS as a 2105 control unit and thus can exploit the PAV capability.
The new Parallel Access Volume (PAV) device types of base IODEVICE UNIT=3390B (or
3380B) and alias UNIT=3390A (or 3380A) are fully supported. The alias 3390A/3380A UCBs
are built above the 16 MB line (See Section 9.2.4, “WLM dynamic alias management” on
page 154 for discussion on Static and Dynamic PAVs).
The total number of devices defined (3390, 3380, 3390B, 3380B, 3390A, and 3380A) must
not exceed 256 per CNTLUNIT statement.

Chapter 9. S/390 and zSeries host setup tasks
147
Here are some additional considerations:

Make sure the IOCP/HCD control unit CUADD, device type UNIT and device quantities
match with the logical control unit and volume/device definitions within the ESS.

Mismatches between the logical hardware configuration and device definition in HCD are
reported by DFSMS device support message
IEA435I PHYSICAL DEVICE INCONSISTENT
WITH LOGICAL DEFINITION
. This message is generated during the vary on-line process for
the device.

The Multiple Allegiance function does not require specific definitions within OS/390 or
z/OS, since it is provided by ESS internal logic.

Do not define esoteric unit names for PAV alias devices. Alias devices are not used for
device allocation.
If you run an MVS in Transparency mode, then define the control unit in IOCP/HCD as UNIT
type 3990. You only need one CNTLUNIT statement for each LCU in IOCP/HCD rather than
the 3990 convention of defining two CU images. No Parallel Access Volume (PAV) device
types, 3390B, 3390A, 3380B, 3380A are permitted in Transparency mode.
If the OS/390 I/O subsystem recognizes the new ESS control unit type 2105 (3990-6
emulation), but no ESS native functions are supported, then you are running the ESS in
Toleration mode. The new Parallel Access Volume (PAV) device types of base (IODEVICE
UNIT=3390B or 3390B) and alias (UNIT=3390A or 3380A) are recognized. However, UCBs
are only built for the base 3390B/3380B and non-PAV device types (3390, 3380).
The total number of devices defined (3390, 3380, base 3390B/3380B, and alias
3390A/3380A) must not exceed 256 per CNTLUNIT statement.
9.2.2 FICON host connectivity
The ESS now supports FICON channels. FICON channel connectivity brings some
differences and provides benefits over the ESCON channel connectivity. These are among
the benefits you should consider:

Increased channel device-address support. FICON channel supports 16,384 devices on
one channel, compared to 1,024 devices on an ESCON channel. This makes possible to
any FICON channel connected to the ESS, to address all the 4096 devices you can have
within the ESS. This extra flexibility will simplify your configuration setup and management.

FICON provides an increased number of channel-to-control unit concurrent I/O
connections. ESCON supports one I/O connection at any one time while FICON channels
support multiple concurrent I/O concurrent connections. While an ESCON channel can
have only one I/O operation at a time, the FICON channel can have I/O operations to
multiple LCUs at the same time, even to the same LCU, by using the FICON protocol
frame multiplexing.

FICON has more than 4 times the effective channel bandwidth for the initial
implementation (70 MB/sec, compared to 17 MB/sec for ESCON).

Reduced number of channels and required fibers with increased bandwidth and I/O rate.

Greater un-repeated fiber link distances (from 3 km. for ESCON to up to 10 km., or 20 km.
with an RPQ, for FICON).
Note:
Previous releases of OS/390 and DFSMS/MVS, now not supported, will have to run
the ESS in either Transparency or Toleration mode. These modes of operating the ESS
must be considered as transitional because the recommendation is to have all your
systems at the currently supported levels of software and therefore capable to run the ESS
in Exploitation mode.

148

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
All these factors allow you to lay out more simple redundant configurations using FICON,
accessing more data with better performance than what is possible with ESCON.
Table 9-1 summarizes the maximum number of ports, devices, and logical paths supported by
FICON and ESCON attachments.
Table 9-1 FICON and ESCON comparison
The ESS limitation of 128 logical paths per LCU limits the total number of logical paths per
ESS to 2048 for both FICON and ESCON.
The intermixing of ESCON and FICON channels to the same LCU from the same operating
system image is only supported for migration purposes. It is not a recommended
configuration for the production environment. The coexistence is useful during the transition
period from ESCON to FICON channels. The mixture allows you to dynamically add FICON
native channel paths to a control unit while keeping its devices operational. A second dynamic
I/O configuration change can then remove the ESCON channels while keeping devices
operational.
For more information on ESS FICON support, see the
IBM TotalStorage Enterprise Storage
Server,
SG24-5465. You may also refer to the document at
http://www.storage.ibm.com/hardsoft/products/ess/support/essficonwp.pdf
for further
considerations on FICON system attachment.
9.2.3 IOCP/HCD examples
Following are two IOCP/HCD examples.
Example 1: ESCON with ESCON Director, 8 LCUs
Sample IOCDS
Example 9-1 shows a sample IOCDS listing which defines a 2105/3990 logical control unit
within an ESS. The devices defined are 16 PAV base devices 3390B and 112 PAV alias
devices 3390A.
Example 9-1 IOCDS for ESCON attached ESS
CHPID PATH=(8A,9A,A0,B0,C0,D0,E2,E6),SWITCH=30,TYPE=CNC
CNTLUNIT CUNUMBR=1A00,PATH=(8A,E2,A0,C0,9A,B0,D0,E6),UNIT=2105,*
LINK=(C0,C1,C2,C3,C0,C1,C2,C3),UNITADD=((00,128)),CUADD=0
CNTLUNIT CUNUMBR=1A80,PATH=(8A,E2,A0,C0,9A,B0,D0,E6),UNIT=2105, *
LINK=(C0,C1,C2,C3,C0,C1,C2,C3),UNITADD=((00,128)),CUADD=1
CNTLUNIT CUNUMBR=1B00,PATH=(8A,E2,A0,C0,9A,B0,D0,E6),UNIT=2105, *
LINK=(C0,C1,C2,C3,C0,C1,C2,C3),UNITADD=((00,128)),CUADD=2
CNTLUNIT CUNUMBR=1B80,PATH=(8A,E2,A0,C0,9A,B0,D0,E6),UNIT=2105, *
LINK=(C0,C1,C3,C3,C0,C1,C2,C3),UNITADD=((00,128)),CUADD=3
CNTLUNIT CUNUMBR=1C00,PATH=(8A,E2,A0,C0,9A,B0,D0,E6),UNIT=2105, *
LINK=(C0,C1,C2,C3,C0,C1,C2,C3),UNITADD=((00,128)),CUADD=4
CNTLUNIT CUNUMBR=1C80,PATH=(8A,E2,A0,C0,9A,B0,D0,E6),UNIT=2105, *
LINK=(C0,C1,C2,C3,C0,C1,C2,C3),UNITADD=((00,128)),CUADD=5
CNTLUNIT CUNUMBR=1D00,PATH=(8A,E2,A0,C0,9A,B0,D0,E6),UNIT=2105, *
FICON ESCON
Host adapter ports per ESS 16 32
Devices per channel 16384 1024
Logical paths per host adapter port 256 64
Logical paths per LCU (ESS limitation) 128 128

Chapter 9. S/390 and zSeries host setup tasks
149
LINK=(C0,C1,C2,C3,C0,C1,C2,C3),UNITADD=((00,128)),CUADD=6
CNTLUNIT CUNUMBR=1D80,PATH=(8A,E2,A0,C0,9A,B0,D0,E6),UNIT=2105, *
LINK=(C0,C1,C2,C3,C0,C1,C2,C3),UNITADD=((00,128)),CUADD=7
IODEVICE ADDRESS=(1A00,016),CUNUMBR=(1A00),STADET=Y,UNIT=3390B
IODEVICE ADDRESS=(1A10,112),CUNUMBR=(1A00),STADET=Y,UNIT=3390A
IODEVICE ADDRESS=(1A80,016),CUNUMBR=(1A80),STADET=Y,UNIT=3390B, *
UNITADD=00
IODEVICE ADDRESS=(1A90,112),CUNUMBR=(1A80),STADET=Y,UNIT=3390A, *
UNITADD=10
IODEVICE ADDRESS=(1B00,016),CUNUMBR=(1B00),STADET=Y,UNIT=3390B
IODEVICE ADDRESS=(1B10,112),CUNUMBR=(1B00),STADET=Y,UNIT=3390A
IODEVICE ADDRESS=(1B80,016),CUNUMBR=(1B80),STADET=Y,UNIT=3390B, *
UNITADD=00
IODEVICE ADDRESS=(1B90,112),CUNUMBR=(1B80),STADET=Y,UNIT=3390A, *
UNITADD=10
IODEVICE ADDRESS=(1C00,016),CUNUMBR=(1C00),STADET=Y,UNIT=3390B
IODEVICE ADDRESS=(1C10,112),CUNUMBR=(1C00),STADET=Y,UNIT=3390A
IODEVICE ADDRESS=(1C80,016),CUNUMBR=(1C80),STADET=Y,UNIT=3390B, *
UNITADD=00
IODEVICE ADDRESS=(1C90,112),CUNUMBR=(1C80),STADET=Y,UNIT=3390A, *
UNITADD=10
IODEVICE ADDRESS=(1D00,016),CUNUMBR=(1D00),STADET=Y,UNIT=3390B
IODEVICE ADDRESS=(1D10,112),CUNUMBR=(1D00),STADET=Y,UNIT=3390A
IODEVICE ADDRESS=(1D80,016),CUNUMBR=(1D80),STADET=Y,UNIT=3390B, *
UNITADD=00
IODEVICE ADDRESS=(1D90,112),CUNUMBR=(1D80),STADET=Y,UNIT=3390A, *
UNITADD=10
If the LCU will be shared with another system image, then FEATURE=SHARED must be
specified for the IODEVICE macro.
HCD panels
The following are the HCD panels used to generate the IOCDS shown in Example 9-1. The
devices defined are 16 PAV base devices (unit type 3390B) and 112 PAV aliases (unit type
3390A), a total of 128 devices on the LCU.
Begin by defining the 2105 logical control unit as shown in Figure 9-1.
Figure 9-1 HCD panel control unit 2105
Specify or revise the following values.
Control unit number . . . . 1A00 +
Control unit type . . . . . 2105______ +
Serial number . . . . . . . FCA50_____
Description . . . . . . . . SSID 0501________________
Connected to switches . . . 30 30 30 30 70 70 70 70 +
Ports . . . . . . . . . . . c0 c1 c2 c3 c0 c1 c2 c3 +
If connected to a switch, select whether to have CHPIDs/link
addresses, and unit address range proposed.
Auto-assign . . . . . . . . 2 1. Yes

150

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
The description fields SSID and serial number should be completed as a means of cross
referencing. The ESS serial number is shown on the ESS Specialist
Welcome
screen (see
Figure 5-3 on page 84). The control unit descriptions are shown in the ESS
Specialist S/390

Storage panel (Figure 9-2). The LCU number (CUADD), logical control unit emulation type
and the SSID assigned are also detailed. Native 2105 support is shown as 3990-6 emulation.
Figure 9-2 ESS Specialist LCU and device details
The CUADD (see Figure 9-3) must match the LCU number specified in the ESS logical
configuration. CUADD=0 associates the CU image with ESS LCU 0, CUADD=1 with LCU 1,
and so on up to CUADD=F.
Figure 9-3 HCD panel CUADD
Figure Note:
Figure 9-2 is for illustration purposes only; the values on the figure do not
match the example configuration.
ssssssssssssssssssssss Select Processor / Control Unit sssssssssssssssssssss
Row 33 of 40 More: >
Command ===> ___________________________________________ Scroll ===> 0016
Select processors to change CU/processor parameters, then press Enter.
Control unit number . . : 1A00 Control unit type . . . : 2105
Log. Addr. -------Channel Path ID . Link Address + -------
/ Proc. ID Att. (CUADD) + 1---- 2---- 3---- 4---- 5---- 6---- 7---- 8----
_ SYS9X9 0_ 8A.C0 E2.C1 A0.C2 C0.C3 9A.C0 E6.C1 B0.C2 D0.C3
_ SYS911 __ _____ _____ _____ _____ _____ _____ _____ _____
_ SYS967 __ _____ _____ _____ _____ _____ _____ _____ _____

Chapter 9. S/390 and zSeries host setup tasks
151
Select the device address and range required, as shown in the HCD panel in Figure 9-4.
If you are defining PAV devices, the device range match the aliases defined on the ESS.
Figure 9-4 HCD panel address range
The number of volumes allocated to the ESS logical control unit and the device types
emulated internally within the logical control unit are shown on the
LCU Devices
table on the
ESS Specialist
S/390 Storage
panel (Figure 9-2). These are “real” base volumes.
Start at the lowest device address available and define your PAV base and non-PAV (standard
3390) devices.
Figure 9-5 shows the definition for 16 PAV base devices of type 3390B. Valid devices for
bases are 3390B and 3380B.
sssssssssssssssssssss Select Processor / Control Unit ssssssssssssssssssssss
Row 35 of 40 More: < >
Command ===> ___________________________________________ Scroll ===> 0016
Select processors to change CU/processor parameters, then press Enter.
Control unit number . . : 1A00 Control unit type . . . : 2105
--------------Unit Address . Unit Range + --------------
/ Proc. ID Att. 1-------- 2------- 3----- 4----- 5----- 6----- 7----- 8-----
_ SYS9X9 00.128 ________ _____ ______ ______ ______ ______ ______
_ SYS911 ________ ________ ______ ______ ______ ______ ______ ______
_ SYS967 ________ ________ ______ ______ ______ ______ ______ ______

152

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Figure 9-5 HCD Add Device panel for 16 base devices
The new WLMPAV parameter is shown on the
Define Device Parameters / Features
panel in
Figure 9-6. It should be set to
NO
for Static PAVs or non-PAV devices, and
YES
for dynamic
PAVs. See Section 9.2.4, “WLM dynamic alias management” on page 154 for more details.
Figure 9-6 HCD Define Device Parameters panel with WLMPAV parameter
ssssssssssssssssssssssssssssssss Add Device sssssssssssssssssssssssssssssss
Specify or revise the following values.
Device number . . . . . . . . 1A00 (0000 - FFFF)
Number of devices . . . . . . 16__
Device type . . . . . . . . . 3390B________ +
Serial number . . . . . . . . __________
Description . . . . . . . . . ________________________________
Volume serial number . . . . . ______ (for DASD)
Connected to CUs . . 1A00 ____ ____ ____ ____ ____ ____ ____ +
Row 1 of 1
Command ===> _____________________________________ Scroll ===> CSR
Select processors to change device/processor definitions, then press
Enter.
Device number . . : 1A00 Number of devices . : 16
Device type . . . : 3390B
Preferred Explicit Device
/ Processor ID UA + Time-Out STADET CHPID + Candidate List
_ SYS9X9 __ No Yes __ No
Row 1 of 6
Command ===> ___________________________________________ Scroll ===> CSR
Specify or revise the values below.
Configuration ID . : MVS02 DB2 Perf
Device number . . : 1A00 Number of devices : 16
Device type . . . : 3390B
Parameter/
Feature Value P Req. Description
OFFLINE No Device considered online or offline at IPL
DYNAMIC Yes Device supports dynamic configuration
LOCANY No UCB can reside in 31 bit storage
WLMPAV Yes Device supports work load manager
SHARED Yes Device shared with other systems
SHAREDUP No Shared when system physically partitioned

Chapter 9. S/390 and zSeries host setup tasks
153
Figure 9-7 shows the HCD Define Device Parameters / Features panel for alias device
definition. Notice that the only device attribute you can set on an alias device is WLMPAV.
Valid devices for aliases are 3390A or 3380A.
Figure 9-7 HCD panel define alias device
Example 2: Direct FICON attachment, 16 LCUs
Example 9-2 shows the IOCP statements for a fully configured FICON attached ESS with 16
LCUs. We have four FICON channels directly connected. All channels address all 16 LCUs.
Each LCU is defined with 80 base devices and 176 alias devices. The number of base
volumes, 80, is the number of full-size 3390-9 volumes in an LCU in a fully configured 11.2TB
ESS (36GB DDMs).
A native FICON channel is defined as TYPE=FC. A FICON channel connected through a
FICON Bridge card on the ESCON director is defined as TYPE=FCV.
The IOCP in Example 9-2 corresponds to the sample configuration in Section 8.4.11, “A
step-by-step example” on page 140.
ssssssssssssssssssssssssssssssss Add Device sssssssssssssssssssssssssssssss
Specify or revise the following values.
Device number . . . . . . . . 1A10 (0000 - FFFF)
Number of devices . . . . . . 112_
Device type . . . . . . . . . 3390A________ +
Serial number . . . . . . . . __________
Description . . . . . . . . . ________________________________
Volume serial number . . . . . ______ (for DASD)
Connected to CUs . . 1A00 ____ ____ ____ ____ ____ ____ ____ +
Row 1 of 1
Command ===> ___________________________________________ Scroll ===> CSR
Specify or revise the values below.
Configuration ID . : MVS02 DB2 Perf
Device number . . : 1A10 Number of devices : 112
Device type . . . : 3390A
Parameter/
Feature Value P Req. Description
WLMPAV Yes Device supports work load manager

154

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Example 9-2 IOCP for FICON attached ESS
CHPID PATH=(91,92,93,94),TYPE=FC,SHARED
*=================================================================
* 16 LCU’S
CNTLUNIT CUNUMBR=900,PATH=(91,92,93,94),UNITADD=((00,256)), *
UNIT=2105,CUADD=0
CNTLUNIT CUNUMBR=910,PATH=(91,92,93,94),UNITADD=((00,256)), *
UNIT=2105,CUADD=1
:
CNTLUNIT CUNUMBR=9F0,PATH=(91,92,93,94),UNITADD=((00,256)), *
UNIT=2105,CUADD=F
*=================================================================
* 80 BASE + 176 ALIAS DEVICES PER LCU
IODEVICE ADDRESS=(9000,080),CUNUMBR=(900),STADET=Y,UNIT=3390B, *
FEATURE=SHARED,UNITADD=00
IODEVICE ADDRESS=(9050,176),CUNUMBR=(900),STADET=Y,UNIT=3390A, *
FEATURE=SHARED,UNITADD=50
IODEVICE ADDRESS=(9100,080),CUNUMBR=(910),STADET=Y,UNIT=3390B, *
FEATURE=SHARED,UNITADD=00
IODEVICE ADDRESS=(9150,176),CUNUMBR=(910),STADET=Y,UNIT=3390A, *
FEATURE=SHARED,UNITADD=50
:
IODEVICE ADDRESS=(9F00,080),CUNUMBR=(9F0),STADET=Y,UNIT=3390B, *
FEATURE=SHARED,UNITADD=00
IODEVICE ADDRESS=(9F50,176),CUNUMBR=(9F0),STADET=Y,UNIT=3390A, *
FEATURE=SHARED,UNITADD=50
9.2.4 WLM dynamic alias management
The IBM TotalStorage Enterprise Storage Server's implementation of PAVs introduces the
concept of base addresses (base devices) and alias addresses (alias devices).
Base address
This is the actual unit address of the volume. There is only one base address
for any volume.
Alias address
Alias addresses are mapped back to a base device address. I/O scheduled
for an alias is physically performed against the base by the ESS. No physical
disk space is associated with an alias address, however, they do occupy
operating system storage. Alias UCBs are stored above the 16 MB line.
The link between base address and alias addresses occurs during IPL processing. A base
address and its aliases must exist in the same ESS logical subsystem (LSS). Together the
base and alias addressed are termed exposures. The number of exposures for a volume is
base plus number of current aliases.
While the base address is the actual unit address of a given volume, there can by many alias
addresses assigned to a base address, and any or all of those alias addresses can be
reassigned to a different base address.
Alias devices assigned to a base device allow the host operating system to start multiple I/O
requests to the base device in parallel and thus reduce UCB queueing in IOS. Depending on
the system configuration, the total number of aliases can be limited by addressing or other
constraints. Therefore, aliases can be a relatively scarce resource that needs to be managed
to ensure the aliases are assigned to the base devices that need them at the moment.

Chapter 9. S/390 and zSeries host setup tasks
155
For best I/O management, the aliases should be assigned to the busiest base devices and
those handling the most important work. You can reassign aliases manually using the ESS
Specialist. However, this process is labor intensive and, to achieve optimum results, must be
repeated whenever there are workload changes. The process of alias reassignment needs to
be automatic and be sensitive to the importance and level of I/O activity on the device.
The OS/390 and z/OS Workload Manager (WLM) and I/O Supervisor (IOS) components
allow the automatic management of aliases. This function is called
dynamic alias
management.
We also refer to this as
dynamic PAV
support. With dynamic alias
management, WLM can automatically perform alias device reassignments from one base
device to another to help work meet its goals and to minimize IOS queueing as workload
requirements change. By default, the function is turned off, and must be explicitly activated for
the sysplex through an option in the WLM service definition. A device-level option is available
through HCD.
Dynamic PAV is supported on systems running OS/390 2.7 and DFSMS 1.5 with enabling
maintenance. To exploit the use of dynamic PAVs, you need to be running your sysplex WLM
in
goal mode
.
Static PAV
support is available on systems running OS/390 1.3 and DFSMS 1.3 and above.
With static PAV, alias devices remain assigned to the base devices you have assigned them
using the ESS Specialist. Alias reassignment can be done manually using the ESS Specialist.
PAVs are static if you are running your sysplex WLM in compatibility mode.
Controls for dynamic workload management
The following three options control the use of WLM dynamic alias management:

Use the
Dynamic alias management
option on the
WLM Service Coefficient/Service
Definition Options
panel (Figure 9-8) to enable dynamic alias management globally in a
sysplex. The default value is
NO
(dynamic alias management is disabled).

The
I/O priority management
option on the
WLM Service Coefficient/Service Definition
Options
panel has on effect on which dynamic alias management algorithms will be used.
The default value is
NO
.

Use the
WLMPAV
parameter in a device's HCD definition to individually enable or disable
dynamic alias management on a given device. The default is
YES
.
Figure 9-8 Activate WLM dynamic alias management
Coefficients/Options Notes Options Help
-------------------------------------------------------------
Service Coefficients/Service Definition Options
Enter or change the Service Coefficients:
CPU. . . . . . . . . . . . . _______ (0.0-99.9)
IOC. . . . . . . . . . . . . _______ (0.0-99.9)
MSO. . . . . . . . . . . . . _______ (0.0000-99.9999)
SRB. . . . . . . . . . . . . _______ (0.0-99.9)
Enter or change the service definition options:
I/O priority management . . . . . . . . YES (Yes or No)
Dynamic alias management . . . . . . . YES (Yes or No)

156

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
When you specify
Yes
for the Dynamic alias management option, you enable dynamic alias
management globally throughout the sysplex. WLM will keep track of the devices used by
different workloads and broadcast this information to other systems in the sysplex. If WLM
determines that a workload is not meeting its goal due to IOS queue time, then WLM attempts
to find alias devices that can be moved to help that workload achieve its goal. Even if all work
is meeting its goals, WLM will attempt to move aliases to the busiest devices to minimize
overall queueing.
Note that the WLM Dynamic alias management option is a global setting for the entire
sysplex. If any system in the sysplex specifies YES for Dynamic alias management, it is
enabled for all systems in the sysplex. There is no consistency checking for this parameter. It
is an installation’s responsibility to coordinate definitions consistently across a sysplex. WLM
will not attempt to enforce a consistent setting of this option.
WLM Service definition options for dynamic alias management
There are two algorithms controlling the distribution of aliases within a subsystem - the goal
algorithm and the efficiency algorithm. The goal algorithm recognizes when a service class is
missing its goal due to IOS queue delay and moves aliases to the base devices serving the
most important work that needs help. The efficiency algorithm simply moves aliases to the
base devices that have the greatest IOS queueing without regard to the importance or goals
of the work. These two algorithms work together to first help service classes meet goals and
beyond that to minimize overall IOS queueing in the subsystem.
You control the activation of these algorithms through the I/O priority management and
Dynamic alias management options in the WLM service definition as shown in Table 9-2:
Table 9-2 Dynamic alias management algorithms
If you enable dynamic alias management, you should also enable I/O priority management.
So you need to specify YES for both of these options on the panel. If I/O priority management
is set to NO, you will get only the efficiency part of dynamic alias management and not the
goal-oriented management. This means that WLM will make alias moves that minimize
overall IOS queueing, but these moves will not take service class goals into consideration.
(Note that you still need to be in WLM goal mode, even if you wish to use the efficiency
function only.)
Device eligibility for dynamic management
While you can globally enable or disable dynamic alias management on the WLM ISPF panel,
you can also individually enable or disable dynamic alias management on a given device via
HCD. You can do this by specifying WLMPAV=YES or NO in that device's definition on the
HCD Define Device Parameters / Features
panel. See examples in section Section 9.2.3,
“IOCP/HCD examples” on page 148.
The relationship of the WLM Dynamic alias management option and the HCD WLMPAV
parameter is following:

If the WLM Dynamic alias management option for the sysplex is set to NO, then all PAVs
are static regardless of their WLMPAV setting.
I/O priority management Dynamic alias management Dynamic alias algorithm in effect
YES YES both efficiency and goal
NO YES efficiency only
YES NO none
NO NO none

Chapter 9. S/390 and zSeries host setup tasks
157

If the WLM Dynamic alias management option for the sysplex is set to YES, then PAVs are
dynamic or static depending on their WLMPAV setting.
You can specify WLMPAV=YES or NO on both base and alias devices in HCD. The WLMPAV
settings on an alias device, however, is only meaningful when the alias device is assigned to
a base device that is offline, as follows:

If the base device is
offline
, then only alias devices with WLMPAV set to YES are eligible to
be reassigned to other base devices.
The WLMPAV setting on the base device itself is irrelevant when the base device is offline,
for either “giving” or “receiving” aliases. (Even if WLMPAV was set to YES on the base
device, it cannot have new aliases assigned to it, as it is offline.)

If the base device is
online
, then the WLMPAV settings on the aliases are ignored, and
WLM can reassign aliases as follows:
– If WLMPAV is set to YES on the base device, then the aliases can be reassigned
regardless of their WLMPAV settings.
– If WLMPAV is set to NO on the base device, then no aliases can be reassigned,
regardless of their WLMPAV settings.
For a WLMPAV=YES base device, the aliases initially assigned to it should be allowed to
default to WLMPAV=YES. The only situation where you might want to change an alias to
WLMPAV=NO is if the alias is initially assigned to a WLMPAV=NO base device. Because the
base is set to NO, the aliases initially assigned to it will not be moved to other bases by WLM.
Then, because the aliases are set to NO, if the base is ever varied offline, the aliases remain
assigned to that base and cannot be reassigned by WLM to other bases. Certain
combinations of WLMPAV settings are not recommended, as described in Table 9-3:
Table 9-3
Effects of WLMPAV settings on base and alias devices in HCD
Base
Device
WLMPAV
Setting
Alias
Device
WLMPAV
Setting
Effects/Recommendations
YES YES
If base is online
: Base is WLM-managed. Aliases can be freely
moved to and from the base device by WLM.
If base is offline
: Aliases become unbound and are available to
WLM to assign to other WLM-managed bases.
YES NO
Not recommended
. If base is WLM-managed, then it is not
predictable which aliases will remain assigned to that base when
the base goes offline. If the base device is set to YES, then you
should set the aliases to YES as well. (See previous option.)
NO YES
Not recommended
. If the base is not WLM-managed, then you
risk losing all of its aliases when the device goes offline. (See next
option.)
NO NO
If base is online
: Base is not WLM-managed. The initial aliases
assigned to this base remain there.
If base is offline
: Aliases remain bound to the offline base device
and are not available to WLM for reassignment. When the base
comes back online, it retains its initial alias assignments.

158

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Aliases of an offline device are considered unbound and WLM will consider unbound aliases
as the best donor devices. If you have a device offline to one system and online to others, you
should make the device ineligible for dynamic alias management.
Scope of management
WLM’s scope of management is sysplex wide. WLM's decisions about reassigning aliases is
based on sysplex-wide awareness of workload and whether the workload is meeting its goals.
It is recommended not to use dynamic alias management for a device unless all systems
sharing that device have dynamic alias management enabled. Otherwise, WLM will be
attempting to manage alias assignments without taking into account the activity from the
non-participating systems. If systems in the sysplex are at levels that do not support dynamic
PAV, but do support static PAVs, a potential performance problem may exist. Static PAV
systems will recognize PAV reassignment instituted by dynamic PAV systems, however, the
I/O contributed by the back level systems will not be considered by WLM during alias
reassignment.
Dynamic PAVs should not be defined for volumes that are shared outside the sysplex as WLM
has insufficient information about volume usage to correctly manage these devices. In a
configuration where two sysplexes share dynamic PAVs, the WLM on sysplex1 would make
alias/base decisions based on the workload on sysplex1, and the WLM on sysplex2 would
make alias/base decisions based on the workload on sysplex2. As neither WLM knows
anything about the workload on the other sysplex, the two could end up moving the same
alias back and forth between different bases, just causing overhead and not really helping
either sysplex.
Preparing for dynamic alias management
Following is a checklist of the steps required to prepare for using dynamic alias management
and the steps to activate it:
1.
Upgrade OS/390:
Install OS/390 V2R7 or higher on all systems that will be sharing the
ESS volumes. The support is integrated in z/OS.
2.
Switch to goal mode:
Dynamic alias management is only active in goal mode.
3.
Turn on I/O priority management:
Set I/O priority management to YES in the WLM
service definition. This function is necessary to provide the I/O delay data needed for
dynamic alias management. Check velocity goals before switching to I/O priority
management because this function factors I/O using and delay samples into the velocity
calculation. To aid in this migration step, the RMF
Workload Activity
report shows what the
velocities will be when I/O priority management is set to YES.
4.
Configure devices on the ESS:
Using the ESS Specialist, define the logical volumes.
They will be assigned base unit addresses. Then configure PAV aliases for the bases.
When you initially define the aliases, you assign them to a base device. WLM can then
dynamically reassign them. Refer to Section 8.4, “ESS custom configuration for CKD
storage” on page 118.
Our recommendation is to configure one alias for a 3390-2/3 type volume, three for a
3390-9 type volume. You may configure as many aliases as your addressing constraints
allow. Base devices plus aliases cannot exceed 256 for each logical subsystem.
In order for dynamic alias management to be most effective from the beginning, try to
spread out your aliases in the initial definition. WLM will only take one PAV alias at a time
from a given base. If the aliases are spread across bases, WLM can make multiple moves
at one time. If aliases are assigned to the same base, WLM will still move them to the busy
devices, but it will take some more time for WLM to reassign the aliases appropriately.

Chapter 9. S/390 and zSeries host setup tasks
159
5.
Define device addresses on the host:
Using HCD, define base and alias device
addresses to z/OS to match the unit address specified in ESS Specialist. Ensure the
IODFs for the base devices and aliases are the same across the sysplex. Leave the
WLMPAV parameter set to YES (the default) for the bases and aliases. This assumes a
sysplex where all the systems using the device are capable of dynamic alias
management.
Enabling dynamic alias management
1.
Install service:
Install enabling function for IOS and WLM together. These are the PTFs
for APARs OW39854 for WLM and OW40669 for IOS (please check the latest PSP bucket
information). They are corequisites. Roll across all OS/390 V2R7 and V2R8 systems in the
sysplex. Support is included in OS/390 2.10 and in z/OS. The default for dynamic alias
management in the sysplex is NO.
2.
Turn on SMF Type 99 recording if desired:
This data can be used by IBM for analysis of
alias management activity if that becomes necessary.
3.
Activate function:
Switch on alias management for the sysplex by editing the WLM
service definition and changing
Dynamic alias management
and
I/O priority
management
to YES. Install the modified service definition and activate a policy.
The dynamic alias management option is NO by default to allow the enabling service to be
installed across the sysplex before activating the function. Overriding the default changes
the functionality level of the service definition to LEVEL008. This means that once the
option is set to YES, you cannot edit, install, activate, or extract the service definition from
a pre-R7 system in the sysplex.
This is the usual functionality level restriction in a mixed-release sysplex. There are also
restrictions when sharing service definitions between sysplexes. For a complete
discussion of migration considerations when using a new functionality level, see Chapter
16: “Workload Management Migration” in
z/OS MVS Planning: Workload Management
,
SA22-7602.
After you install the service definition for dynamic alias management, DO NOT manipulate
the service definition with a back-level R7 or R8 system that does not have the service
installed. The results will be unpredictable.
4.
Monitor results:
Monitor dynamic alias management using operator commands and the
RMF device reports. The command
D M=DEV(devnum-devnum)
displays the current number
of exposures (base + aliases) for a base device. The new PAV column in the RMF device
reports gives the number of exposures observed for a device at the end of the RMF
interval. An asterisk (*) next to the value in the PAV column indicates that the number of
exposures changed either up or down during the last interval.
Additional considerations

To invoke the Parallel Access Volume (PAV) function, you first need to have ordered the
appropriate feature on your ESS. This feature enables multiple I/Os to the same volume or
device address from the same system image.

You can mix non-PAV device types (3390 or 3380) with PAV base devices (3390B/3380B)
and alias devices (3390A/3380A) on the same LCU.

The number of base and aliases may vary in different logical control units. On the HCD
you must ensure that the base and alias definitions for each LCU are consistent with the
ESS logical configuration.

Non-PAV device types do not have access to the alias device pool.

PAVs are static if you are running your sysplex WLM in compatibility mode. You must be
sure to set WLMPAV=NO in your IOCP/HCD device attributes panel.

160

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment

Once WLM has assigned an alias to a base, the alias will remain associated with that new
base until WLM decides to move it again.
Displaying PAV information
MVS system commands
DISPLAY M=DEV
and
DISPLAY M=CHP(xx)
provide new PAV related
information.
The
DEVSERV QPAVS
(
DS QP
) command can be used to display the status information on a
device, group of devices, or an LCU, including PAV devices.
Figure 9-9 shows a
DEVSERV QPAVS
command and resulting display. The
UNIT NUM
heading is
the host device addresses and the
UNIT ADDR
heading details the ESS internal device
address. The
TYPE
heading shows
BASE
,
ALIAS
, or
NON-PAV
as the device type. It also displays
base and alias affiliations. The
STATUS
heading contains information to highlight discrepancies
between the IODF and the ESS configuration, for example an invalid alias
INV-ALIAS
.
Figure 9-9 DEVSERV QPAVS command
See 11.5, “MVS system commands” on page 187 for more examples.
9.2.5 Custom volumes
Custom volumes are defined as 3380 or 3390 format devices types, usually with a
non-standard number of cylinders. There are no special parameters to set when defining
them to the host I/O Subsystem. Prior to using the volume, you must first format the volume
as all other volumes via ICKDSF minimal INIT. A sample job is included in Example 9-3 on
page 165.
When you initialize custom volumes, adjust the VTOC and VTOC Index size to correspond to
the volume size. The same also applies to the VVDS.
DS QP,D222,VOLUME
IEE459I 08.20.32 DEVSERV QPAVS 591
HOST SUBSYSTEM
CONFIGURATION CONFIGURATION
------------- --------------------
UNIT UNIT UA
NUM. UA TYPE STATUS SSID ADDR. TYPE
---- -- ---- ------ ---- ---- ------------
D222 22 BASE 0102 22 BASE
D2FE FE ALIAS-D222 0102 FE ALIAS-22
D2FF FF ALIAS-D222 0102 FF ALIAS-22
**** 3 DEVICE(S) MET THE SELECTION CRITERIA

Chapter 9. S/390 and zSeries host setup tasks
161
9.2.6 Large volumes
The ESS initially supported custom volumes up to 10017 cylinders, the size of the largest
standard volume, the 3390 model 9. This was the limit set by the operating system software.
The ESS Large Volume Support (LVS) enhancement has now increased the upper limit to
32760 cylinders, approximately 27.8 GB. This enhancement is provided as a combination of
ESS Licensed Internal Code (LIC) changes and system software changes, available for z/OS,
OS/390, and z/VM. Refer to Section 7.4, “Large Volume Support” on page 99 for details on
software support.
Benefits of using large volumes can be briefly summarized as follows:

They reduce storage management tasks by allowing you to define and manage smaller
configurations.

They reduce the number of multivolume data sets you have to manage.

They relieve architectural constraints by allowing you to address more data within the
existing 64K subchannel number limit.
Large volume considerations
The size of the logical volume defined does not have an impact on the performance of the
ESS subsystem. The ESS does not serialize I/O on the basis of logical devices so an
increase in the logical volume size does not affect the ESS backend performance. Host
operating systems, on the other hand, serialize I/Os against devices. As more data sets
reside on a single volume, there will be greater I/O contention accessing the device. With
Large Volume Support (LVS) it is more important than ever to try to minimize contention on
the logical device level. To avoid potential I/O bottlenecks on devices

Exploit the use of Parallel Access Volumes to reduce IOS queuing on the system level.

Eliminate unnecessary reserves by using WLM in GOAL mode.

Multiple allegiance will automatically reduce queuing on sharing systems.
Parallel Access Volume (PAV) support is of key importance when implementing large
volumes. PAV enables one MVS system to initiate multiple I/Os to a device concurrently. This
keeps IOSQ times down and performance up even with many active data sets on the same
volume. PAV is a practical “must” with large volumes. We discourage you from using large
volumes without PAV. In particular, we recommend the use of dynamic PAV.
As the volume sizes grow larger, more data and data sets will reside on a single S/390 device
address. Thus, the larger the volume, the greater the multi-system performance impact will be
of serializing volumes with RESERVE processing. You need to exploit a GRS Star
Configuration and convert all RESERVE's possible into system ENQ requests.
Implementation steps
When planning for your logical volume configuration, you will notice that you cannot fully
utilize the array capacity with all 32760 cylinder volumes. For example, a 6+P+S array of
72.8GB disks will contain 14 maximum size Large Volumes, but some 26.6 GB, three times
the capacity of a 3390-9 will be left over. To use all the space, configure one or more smaller
custom volumes to fill the array. Alternatively, if you prefer all of your volumes to be of the
same size, select a smaller uniform volume size that will better utilize the array capacity.

On the ESS, you configure large volumes like any other custom volume using the ESS
Specialist
Add Volumes
panel.

On the host IOCP/HCD, define a large volume as 3390, or 3390B like any other 3390 type
volume.

162

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment

A large volume is reported by the host system as a 3390-9 device containing up to 32760
cylinders. They have 3390 track format. The maximum size for 3380 track format volumes
is 3339 cylinders.
The large volume coexistence support for DFSMS/MVS 1.4 and 1.5 will allow these system
levels to coexist in the same Sysplex with LVS systems. You must install the support in order
to prevent unpredictable results that may arise from systems without Large Volume Support
(LVS) accessing volumes that have more than 10017 cylinders. The coexistence support will

prevent a device with more than10017 cylinders from being varied online to the system

prevent a device from coming online during an IPL if it is configured with more than 10017
cylinder.
On an MVS system with coexistence support installed, the following message will be issued if
you try to vary online a device with more than 10017 cylinders:
IEA434I DEVICE ONLINE IS NOT ALLOWED, GREATER THAN 10017 CYL
In IOCP/HCD, define large volumes as UNIT=3390 type devices, or preferably as PAV base
devices with UNIT=3390B.
Product specific considerations
ICKDSF
ICKDSF has been enhanced to support larger volume sizes. With the necessary support
installed, ICKDSF is able to initialize large volumes.
When you initialize a large volume, be sure to define the VTOC size in proportion to the
volume size. The VTOC is to reside within the first 65,535 tracks of the logical volume as it
does today. It is highly recommended that you use Indexed VTOC on large volumes.
JES spool data sets
JES spool data set maximum size is 65535 tracks or 4369 cylinders as before. On 3390-3
volumes, the spool could occupy the whole volume, but as you move to larger volumes you
will have room on the spool volumes for other data sets. You need to decide whether you want
to allow data set allocations on large spool volumes. Review your SMS constructs to
allow/disallow these allocations. If you do allow them, check what implications this has on
your operational procedures. For example, if your spool volumes are excluded from
DFSMShsm migration and incremental backup processing, you need to modify the relevant
definitions to have them included.
Note that the spool data set must be allocated within the first 65535 tracks of a volume. With
z/OS 1.2 JES spool can reside anywhere on the volume.
DFSMSdss
Large Volume support does not degrade the performance of previously existing functions
when the amount of tracks, data sets, or extents being processed is equivalent. For those
DFSMSdss functions that process all of the tracks on a volume, physical
FULL
volume dump
for example, all of the data sets on a volume, logical copy with
DS(INC(**))
specified for
example, or all of the extents on a volume,
DEFRAG
for example, the processing time will
increase due to the larger number of tracks, data sets, or extents on large volumes.
Due to the larger number of control blocks that DFSMSdss will need to build in order to
represent the larger number of tracks, data sets, and extents on large volumes, the amount of
virtual memory that is required will increase whenever DFSMSdss processes any of the
following:

Chapter 9. S/390 and zSeries host setup tasks
163

All of the tracks on a volume, as in a physical
FULL
volume
DUMP

All of the data sets on a volume, as in a logical
COPY
with
DS(INC(**))

All of the extents on a volume, as in a
DEFRAG
.
Stand-alone Restore
The DFSMSdss Stand-alone Restore program has been enhanced to support large volumes.
You need generate with the DFSMSdss
BUILDSA
command a new copy of the Stand-alone
Restore program (probably on tape) to have the appropriate support included in it. Your old
versions of Stand-alone Restore do not support large volumes.
DFSMShsm
DFSMShsm space-management, backup, ABARS, and reporting functions recognize and
support the use of large volumes. DFSMShsm supports large volumes for

SMS storage groups

Level-0 Non-SMS volumes

Level-1 (ML1) volumes

Level-2 (ML2) DASD volumes

Backup DASD volumes

Control data sets (MCDS, BCDS, OCDS)

Journal
DFSMShsm control data set record formats have been updated to include new larger fields
for the larger volumes. Since an HSMplex can involve several DFSMShsm releases sharing
one MCDS, BCDS, OCDS, and Journal, coexistence PTFs will be required by low level
DFSMShsm systems to share control data sets with DFSMShsm systems that support large
volumes:

If a large volume is used for any CDS or the Journal, all systems must have full Large
Volume Support (LVS) installed

Any older releases need the coexistence support to recognize the new Volume (V) record
in the MCDS, for purposes of the
LIST
and
REPORT
commands

The coexistence support will prevent any large volume used for level-0 DASD to be online
to older releases.
Large volume, with its larger number of tracks and potentially larger number of data sets, can
extend the processing time needed for:

Volume space management (migration)

Volume backup and recover

Volume dump and restore

AUDIT MEDIACONTROLS
However, total migration, backup, and dump times of your data storage should not change
much, as it depends more on the amount of data being processed, and less on the number of
volumes.
If planning to use large volumes as ML1 volumes, consider that DFSMShsm recommends
using at least one more ML1 volume than the maximum number of concurrent migration
tasks. (In the case of Small Dataset Packing, DFSMShsm provides for one SDSP data set per
ML1 volume.) If one large volume replaces some number of smaller ML1 volumes, and the
resulting number of ML1 volumes is now fewer than the number of migration tasks,
performance contention could occur.

164

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Other considerations
The large volume enhancement does not change the size limitations that today exist on
certain data set types. Thus, for example:

The DADSM component of DFSMS/MVS will continue to limit the number of tracks for a
data set to 65535 per volume. This applies for example to SAM, PAM (PDS), IMS GSAM,
and BDAM data sets. This does not apply to VSAM, extended format, PDSE and HFS
data sets.

DFSORT supports large volumes for input, output, and work devices. The number of
records in any DFSORT SORTWK data set cannot exceed 1,000,000 records. The
restriction does not apply to DFSORT input and output data sets.

PDSE data sets can only be single volume. With LVS, you can now allocate larger PDSE
than before.

PPRC, FlashCopy, and XRC support large volumes.

Small custom volumes may be required for those applications that must still use hardware
reserve/release processing, if these are not converted. These custom volumes can be
defined in the non-interleaved partition of the rank.
Performance
The performance of large volumes when implemented with PAVs should allow equivalent
performance to existing volumes sizes today. With larger volumes configured more PAV alias
devices become available for use by z/OS WLM. WLM functions will help to eliminate hot
spots on logical volumes by dynamically managing PAV-alias's in order to meet the specified
work load goals. ESS functions of Multiple Allegiance, I/O Priority Queuing and striping data
across raid ranks will grow in importance in maintaining subsystem performance.
Larger volume sizes will require longer backup and restore times per volume. Physical volume
dumps to tape for large volumes will need to be restored to a volume that can accommodate
the large volume capacity.
Implementation tasks
We can now review the tasks that are required to implement and exploit large volumes:

Plan PAV alias pool size.

Configure PAV-alias and PAV-base devices in HCD and ESS.

Configure large volumes in ESS.

Implement WLM Goal Mode.

Enable WLM Dynamic Alias Tuning.

Assess the impact of having large volumes in your disaster recovery / contingency plan.

Install required vendor products support for large volumes.

Install coexistence support on pre OS/390 R10 systems.

Install large volume software support on all sharing systems in the Sysplex.

Initialize large volumes for use with ICKDSF INIT function.

Migrate data to the ESS large volumes based on your storage need.

Monitor configuration performance.
9.2.7 MIH interval
The recommended Missing Interrupt Handler (MIH) value for ESS logical volumes is 30
seconds.

Chapter 9. S/390 and zSeries host setup tasks
165
The ESS devices are self describing so entries are not required for base addresses. The ESS
will supply a recommended interval of 30 seconds as part of the Read Configuration Data
during IPL. This applies for both ESCON and FICON attachment. OS/390 and z/OS will use
this information to automatically set its MIH value.
If you need to set a value, ensure you set the same MIH value for all images in your Sysplex
that have access to the ESS. The MIH is defined in the IECIOSxx PARMLIB member. The
setting should appear as follows:
MIH DEV=(1A00-1A5F),TIME=30.
The operator command to
achieve the same is
SETIOS MIH,DEV=(1A00-1A5F),TIME=30
.
If any value is defined, this will override the system’s determined value. Make sure you don’t
have in your IECIOS member a global entry such as MIH DASD=mm:ss which applies to all
DASD devices.
MIH values are not supported for alias UCBs. Setting an MIH value will cause an error
message to be generated:
IOS090I dev IS AN INCORRECT DEVICE
9.2.8 DFSMS
There are no new or unique parameters that need to be set in DFSMS while installing the
ESS, apart from including the ESS volumes into existing storage group constructs.
You may wish to consider allocating ESS volumes to a new storage group, with associated
ACS routines, to take advantage of the higher performance functions available to high activity
volumes (Multiple Allegiance and PAV).
DFSMSdss will now invoke the FlashCopy function as described in Section 9.2.10, “ESS
Copy Services” on page 166.
The DFSMS Media Manager will vary the Define Extent Range processing for DB2, IMS Fast
Path, VSAM, and PDSE to optimize the Multiple Allegiance and PAV performance.
9.2.9 Volume initialization
ESS logical volumes need to be initialized with an ICKDSF minimal initialization. A sample job
is included in Example 9-3.
Example 9-3 ICKDSF minimal init
//INITVOL1 JOB account info...
//*----------------------------------------------------------*
//* THIS JOB WILL INIT A VOLUME IN AN OFFLINE MODE *
//* THIS IS A MINIMAL INIT *
//*----------------------------------------------------------*
//MININIT EXEC PGM=ICKDSF,REGION=2M
//SYSPRINT DD SYSOUT=*
//SYSIN DD *
INIT UNITADDRESS(D123) VOLID(MVS203) OWNER(SYSPROG) -
VTOC(0,1,29) INDEX(2,0,15) NOVALIDATE NOCHECK
//
A PAV alias cannot be initialized. Aliases are only known to the I/O Supervisor routines of the
z/OS.
When you initialize custom volumes, adjust the VTOC and VTOC Index size to correspond to
the volume size. You can check the volume size on the ESS Specialist
S/390 Storage
panel
(Figure 9-2). The same also applies to the VVDS.

166

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
9.2.10 ESS Copy Services
In this section we briefly describe OS/390 and z/OS support for ESS Copy Services. ESS
Copy Services implementation tasks are described to a greater extent in Chapter 12, “ESS
Copy Services for S/390” on page 197 and in the IBM Redbook
Implementing ESS Copy
Services on S/390,
SG24-5680.
FlashCopy
The DFSMSdss
COPY FULL
command will automatically invoke the ESS FlashCopy service.
DFSMSdss performs the copy using FlashCopy service if the following conditions are met:

The source and target volumes support the FlashCopy service.

The source and target volumes reside on the same LSS.

The source and target volumes must be online.

The source and target volumes are the same device type.

The source volume does not have a FlashCopy relationship with another volume.

The target volume is not in a FlashCopy session with another volume.

The target volume is the same size or larger than the source volume.

The target volume is not an XRC primary volume.

The target volume is not a PPRC primary or secondary volume.

The target volume is not the source of an in-progress, concurrent copy operation.
If these conditions are not met, but the CC keyword is specified, DFSMSdss provides the
standard concurrent copy service. If these conditions are not met and the CC keyword is not
specified, DFSMSdss uses the standard I/O service to copy the volume.
A sample job stream to invoke FlashCopy is shown in Example 9-4.
Example 9-4 DFSMSdss COPY FULL job
//COPYFULL JOB . . . . .
//*
//INSTIMG EXEC PGM=ADRDSSU
//SYSPRINT DD SYSOUT=*
//SRCVOL DD
//TRGVOL DD
//SYSIN DD *
COPY FULL INDYNAM((SRCVOL)) OUTDYNAM((TRGVOL)) COPYVOLID
/*
PPRC and XRC
The PPRC and XRC functions are defined and operated much the same way as was done
with the previous IBM 3990 Model 6 and RAMAC storage controllers. There are some
additional features to consider, for example, XRC support of the suspend/resume commands.
To invoke the PPRC or XRC function, you need to have ordered the appropriate features
(XRC FC 181x, PPRC FC 182x) on your ESS.
The establishing and managing of the PPRC and FlashCopy functions is done using TSO
commands or using the ESS Copy Services Web interface.
Concurrent copy
The concurrent copy function is defined and operated in the same way as is currently done
with the IBM 3990 Model 6.

Chapter 9. S/390 and zSeries host setup tasks
167
There is no separate feature to order for enabling Concurrent Copy. However, if you have the
FlashCopy feature installed and the conditions for using the FlashCopy service are met,
FlashCopy will be automatically invoked instead of concurrent copy.
9.3 VM/ESA software configuration
VM/ESA operates in Transparency mode with guest-only support of some exploitation
functions. VM/ESA therefore sees the ESS as a group of up to 16 IBM 3990 Model 6 with up
to 256 devices per logical subsystem (LSS).
Therefore, the IOCP definitions are as they would be if you were to define a group of “real”
IBM 3990 Model 6 with up to 256 devices including 3390 and 3380. Guests must also define
the ESS internal logical control units as IBM 3990 Model 6 and not native 2105.
The Multiple Allegiance function requires no specific definitions within VM/ESA as it is
provided by ESS internal logic.
You must ensure the IOCP control unit CUADDR, device type UNIT and device quantities
match, or are within the maximum ranges as set in the logical control unit (LCU) and
volume/device definitions within the ESS. A sample IOCP deck is shown in Example 9-1 on
page 148.
Mismatched definitions between the ESS internal logical configuration and VM/ESA can be
displayed via the
QUERY PAV
command.
9.3.1 Parallel access volumes (PAV)
PAV support is available for guests only and the PAV volumes, base 3390B and alias 3390A,
must be dedicated to a guest. For an explanation on how to code PAVs, see Section,
“Additional considerations” on page 159.
A PAV base (3390B or 3380B) and non-PAV devices (3390 or 3380) can be attached to the
SYSTEM, but alias devices (3390A or 3380A) can only be attached to a guest. If a PAV base
device is attached to the SYSTEM, then the alias pool of devices cannot be attached to a
guest. If a PAV base device is attached to a guest, then the alias pool of devices must also be
attached to the same guest.
Displaying the host view of PAVs
The VM/ESA
QUERY PAV
command can be used to display the base and alias affiliation details
for PAV devices. This command is available for class B authorized users.
The command is in the format
QUERY PAV,ALL
or
QUERY PAV,rdev
.
An example of a response to the command issued against a base device is:
Device 01A2 is a base Parallel Access Volume device with the following aliases: 01FE
01FF
An example of a response to the command issued against an Alias device is:
Device 01F3 is an alias Parallel Access Volume device whose base device is: 01A4

168

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
9.3.2 ESS Copy Services
To invoke the PPRC or FlashCopy function, you need to have ordered the appropriate
features (PPRC FC 182x, FlashCopy FC 183x) on your ESS.
The PPRC functions are managed in the ESS in the same way as with previous storage
controllers using the ICKDSF utility. With the ESS, in a VM/ESA environment you can also
establish and manage the FlashCopy and PPRC functions using the ESS Copy Services Web
interface.
Native FlashCopy is available to guests only and, therefore, the source and target volumes,
dedicated full-pack minidisks, must both be ATTACHed to the VM/ESA guest. The guest
invokes the function, not VM/ESA.
9.4 VSE/ESA software configuration
VSE/ESA operates in Transparency mode only and, therefore, sees the ESS as a group of up
to 16 IBM 3990 Model 6 with up to 256 devices per logical subsystem.
Therefore, the I/O definitions are as they would be if you were to define a group of “real” IBM
3990 Model 6 with up to 256 devices including 3390 and 3380.
The Multiple Allegiance function requires no specific definitions within VSE/ESA as it is
provided by ESS internal logic.
You must ensure the I/O definitions for your control unit CUADDR, device type UNIT and
device quantities match, or are within the maximum ranges as set in the Logical Control Unit
and volume/device definitions within the ESS.
9.4.1 ESS Copy Services
To invoke the PPRC or FlashCopy function, you need to have ordered the appropriate
features (PPRC FC 182x, FlashCopy FC 183x) on your ESS.
The PPRC functions are managed in the ESS in the same way as with previous storage
controllers using the ICKDSF utility. With the ESS, in a VSE/ESA environment you can also
establish and manage the FlashCopy and PPRC functions using the ESS Copy Services Web
interface.
9.5 TPF software configuration
TPF operates in Transparency mode (discounting enhanced CCW support) and therefore
sees the ESS as a group of up to 16 IBM 3990 Model 3 TPF control units with up to 256
devices per logical subsystem (LSS).
Therefore, the I/O definitions are as they would be if you were to define a group of “real” IBM
3990 Model 3 TPF with up to 256 devices.
The Multipath locking facility is supported as on IBM 3990 control units for TPF environments.
The Multiple Allegiance function requires no specific definitions within TPF as it is provided by
ESS internal logic. However, it is only available as an RPQ.

Chapter 9. S/390 and zSeries host setup tasks
169
You must ensure the I/O definitions for your control unit CUADDR, device type UNIT and
device quantities match, or are within the maximum ranges as set in the logical control unit
(LCU) and volume/device definitions within the ESS.
9.5.1 PPRC and FlashCopy
PPRC and FlashCopy are optional features on the ESS Models F10 and F20 and require
ESS LIC level 1.5.0. or later. These functions are optional features on the ESS (FC 182x for
PPRC and 183x for FlashCopy). TPF must be Version 4 Release 1.
In a TPF environment the FlashCopy and PPRC functions are established and managed
using the ESS Specialist Copy Services Web interface.
9.6 Linux software configuration
On zSeries 900 and on S/390 servers you can run Linux natively as a stand-alone or as a
logical partition (LPAR). In addition, the S/390 Virtual Image Facility for Linux, and z/VM V4
enable you to run more Linux images than can be deployed using LPARs and provide
capabilities to help create and manage these images.
Some basic considerations about the different configurations that can be used when running
Linux on a zSeries server are shown in Table 9-4.
Table 9-4 Linux implementations in zSeries servers
Current Linux for S/390 distributions supported with the ESS are SuSE Linux Enterprise
Server for S/390 and Turbo Linux Server 6 for zSeries and S/390. For the most current
information on the Linux for S/390 releases and distributions that are supported with the ESS,
refer to:
http://www.storage.ibm.com/hardsoft/products/ess/supserver.htm
Native hardware
Linux can be run natively on a zSeries server, or a S/390 server that has
Relative
and
Immediate
instruction support. This is not the most recommend because only one Linux
system can be run at any one time.
Linux for S/390
implementation
Reasonable number
of Linux images
Considerations
Native 1 With the exception of the smaller platforms
(as the P/390) this is not economically
feasible
LPAR 15 Useful if a very small number of Linux
images is needed for testing and
development
VIF Tens to hundreds Management of Linux images becomes
difficult when greater than a few tens
z/VM Hundreds to
thousands
Probably the most cost-effective way to run
and manage Linux on the mainframe

170

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Logical partition
Linux images can be run in S/390 logical partitions (LPARs) — up to 15 different logical
partitions. This configuration has the ability to run and manage up to 15 individual Linux
systems. Perhaps more compelling is the ability to run a number of Linux partitions
concurrent with z/OS or VSE/ESA partitions providing high value services such as
transactions and access to existing applications and data.
z/VM guest support
Linux runs as a guest operating system in one or more VM Virtual Machines. z/VM uses
architectural hardware functions in S/390 to virtualize the S/390 instruction set, making each
guest think it has its own dedicated server processor. Any number of Linux or other guests,
including z/OS, VSE/ESA and z/VM CMS Interactive System can run concurrently. The
number of concurrent guests is limited by the resources available to virtualize them.
This configuration has the ability to use devices, both virtual and real, supported by z/VM and
not directly supported by Linux, and this is a plus. You will be able to share more efficiently the
resources among the z/VM users instead of having to dedicate them. For communication
between guest virtual machines, this can be done at extremely high speed.
Running IBM's industry leading z/VM virtualization technology, the zSeries offering for Linux
offers an ideal platform for server consolidation, utilizing the mainframe's ability to create as
few as 20 and up to hundreds of virtual Linux servers on a single physical box, saving server
users substantially on energy, floor space, and maintenance expense.
Virtual Image Facility (VIF)
The S/390 Virtual Image Facility for Linux is an entry level product that offers a complete
server environment for multiple Linux systems on a single zSeries (or S/390) server.The
Virtual Image Facility is an easy-to-use, high-performance environment that operates within a
logical partition or in native S/390 mode and provides the capability to create a significant
number of Linux images. An internal network provides high-speed communication among
Linux images.
Integrated Facility for Linux (IFL)
z/VM V4R2 users can make use of the IBM Integrated Facility for Linux (IFL) to run multiple
Linux images on an IFL-configured z/VM system without affecting the software fees of their
existing zSeries 900 or S/390 environment.
The heart of zSeries and S/390 hardware is the Multi-Chip Module (MCM) which contains up
to 20 Processing Units (PU), commonly referred to on other platforms as CPUs or engines. All
PUs are identical, but can take on different roles in the system. The IFL engine is a new type
of PU which allows additional engines to be dedicated to z/VM or Linux workloads. It is lower
priced than standard engines and allows traditional S/390 software charges to remain the
same. IFL is based on PR/SM LPAR technology and it can be dedicated to a single logical
partition or shared by multiple logical partitions.
For detailed step-by-step guidelines to implement Linux for S/390 we recommend that you
read the following IBM publications:
Implementing Linux with IBM Disk Storage,
SG24-6261;
Linux for S/390,
SG24-4987; and
Linux for IBM zSeries and S/390: Distributions,
SG24-6264.
Note:
Installations should note that the Linux workspace enabled by the IFL facility will not
support any of the S/390 traditional operating systems (z/OS, OS/390, TPF, VSE, or VM).
Only Linux applications or Linux operating in conjunction with z/VM or VIF are supported
by the IBM Integrated Facility for Linux.

© Copyright IBM Corp. 2002
171
Chapter 10.
Data migration in the zSeries
environment
This chapter gives you the information necessary to plan the methods and tools your
installation will be using when doing the migration of the existing data into the IBM
TotalStorage Enterprise Storage Server. The information presented here is basically oriented
to S/390 users who must move data from non-ESS storage subsystems into a ESS storage
subsystems.
10

172

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
10.1 Migration planning
Migration planning is a key factor to the success of ESS implementation. Today, because
business is very dependent on data processing, interruptions are not accepted at all. The
environmental changes have to be made completely transparent to the customer. Therefore,
storage administrators have to plan for non-disruptive methods of migration rather than the
disruptive ones. In addition, it does not matter what method will be used for the migration, it
must guarantee total data integrity.
Storage consolidation
The ESS is a very scalable subsystem that ranges from 420 GB (eight arrays of 9.1 GB
DDMs) to 22.4 Terabytes (48 Arrays of 72.8 GB DDMs). The ESS can simulate 16 IBM 3990
Model 6 controllers and it can address up to 4,096 logical volumes. Therefore, data can be
migrated from one or several storage subsystems to a single central repository of data (ESS).
The whole migration process can take hours or days to be completed and can be done in one
or more steps according to the environmental complexity.
Migration methods
Data migration is the final installation process, when the data begins to populate the ESS.
The data migration from the previous storage subsystems to ESS can be done by using
standard host utilities, ESS Copy Services, or using IBM migration services offerings.
10.2 Data migration for MVS environments
There are two major considerations when planning data migration for MVS environments.
Systems-Managed Storage (SMS) environments
SMS implements storage policies and handles most of your storage management tasks. The
data migration for ESS is greatly simplified in the SMS environment compared to non-SMS.
You can create a new storage group for ESS and alter the ACS routines to reflect the
changes. You can also add ESS volumes to an existing storage group.
Non-SMS environments
If your environment is non-SMS, you can migrate data from non-ESS volumes to ESS
volumes using standard utilities like DFSMSdss. To avoid changes in JCL, you should copy
the volume data by using DFSMSdss
COPY
with
COPYVOLID
.
10.2.1 Migration methods
Among several migration methods, you can choose one method - or combine more than one
method - to migrate data that best fits your environment.
DFSMSdss
DFSMSdss is the faster data mover product that helps you to migrate data between
like
and
unlike
devices.
You can use the DFSMSdss
COPY
command to perform a physical volume copy. Full-volume
COPY
can move data only between like devices of equal or greater capacity (for example, from
a single-capacity 3390-3 model to a triple-capacity 3390-9 model). If you specify
COPYVOLID
,
the volume serial number of the source volume is copied to the target volume.

Chapter 10. Data migration in the zSeries environment
173
Logical volume copy is a data-set oriented method that allows you to move data between
unlike devices (for example, from a 3380 to a 3390). To move a volume logically, use the
DFSMSdss
COPY DATASET
command.
DFSMSdss
DUMP
and
RESTORE
commands move data from source volumes to tape cartridges
and then to the target volumes. The advantage of this approach is that you do not need to
attach to host all storage subsystems involved in a migration at the same time. You can also
use
DUMP/RESTORE
to bring ESS data from storage subsystems installed in different buildings.
You can choose to dump and restore physically or logically by using or not using the
DSNAME

keyword in conjunction with DFSMSdss
DUMP
and
RESTORE
command.
Others host utilities
You can have on hand the following alternatives to DFSMSdss to move data sets:

IDCAMS
EXPORT/IMPORT
(VSAM)

IDCAMS
REPRO
(VSAM, SAM, BDAM)

IEBCOPY (PDSs - including load module libraries - and PDSEs)

ICEGENER (SAM) - part of DFSORT

IEBGENER (SAM)

Specialized data base utilities (such as for CICS, DB2 or IMS)
DFSMShsm
DFSMShsm invokes DFSMSdss to perform a data migration. So, it is a similar method to the
DFSMSdss described above. DFSMShsm has control data sets and journal to help you to
control the entire migration process. You can use DFSMShsm
DUMP
or
MIGRATE
commands to
move data.
See the
z/OS DFSMShsm Storage Administration Guide,
SC35-0421, and
z/OS DFSMShsm
Storage Administration Reference,
SC35-0422.
DFSMShsm ABARS
Although its primary purpose is disaster recovery, you can also use DFSMShsm ABARS to
migrate volumes to cartridges and restore them in an ESS. In the Aggregate Group you have
to define the data sets you intend to migrate to the ESS.
For reference, see
DFSMShsm ABARS and Mainstar Solutions,
SG24-5089.
XRC
XRC is a non-disruptive alternative to migrating data to ESS. Migration takes place while the
application is still running. Source volumes have to reside on IBM 3990 controller or non-IBM
controller with XRC support. The target volumes are those defined on an ESS. In an XRC
session, System Data Mover (SDM) manages the process of copying data from source
volume to the corresponding target volume. During the copy, updates on a source volume will
be made on a target volume in the same order, thus maintaining update sequence
consistency.
When all pairs are synchronized, you can stop the application systems, stop the XRC,
recover the volumes, and restart the applications using the new volumes. The XRC
XRECOVER

command re-labels the target volumes with the source volume serials.
For details about XRC characteristics and implementation, see
Planning for IBM Remote
Copy,
SG24-2595 and
Implementing ESS Copy Services on S/390,
SG24-5680

174

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
10.3 Data migration for VM environments
DFSMS/VM contains services which include a data mover, an automated move process, and
interactive user interface:

DASD Dump Restore (DDR) is a service utility shipped with VM that you can use to dump
data from disk to tape, restore data from tape to disk, and copy data between like disk
drive volumes. You cannot use DDR to copy data between disk devices with different track
formats.
CMDISK
is a DIRMAINT command you can use to move minidisks from any device type
supported by VM to any other type.
COPYFILE
is a CMS command you can use to copy files or minidisks between devices with
the same or different track formats.
PTAPE
is a CP command you can use to dump spool files to tape and to load them from
tape to disk.
10.4 Data migration for VSE environments
You can use several dialogs in the VSE interactive interface to set up the jobs to move data.
You can reorganize your data and eliminate space fragmentation by using the backup/restore
dialogs:
1.Export and import VSAM files.
2.Back up and restore VSAM files.
3.Back up and restore ICCF libraries.
4.Back up and restore the system history file.
5.Bac kup and restore the system residence library.
6.Create a loadable tape with the system residence library and system history file ready to
restore.
You can also use these capabilities:

VSE/FASTCOPY to move volumes and files between devices with identical track formats.

VSE/DITTO to copy files.

VSE/POWER commands to transfer the SPOOL queue from one device to another.

VSE/VSAM to move any VSAM data set using either
REPRO
or
EXPORT/IMPORT
functions.

Other vendors also provide utilities to move data from one device to another.
10.5 Data migration across ESSs
The information in the previous sections is oriented to the data migration scenarios where
data from non-ESS storage subsystems is moved (or copied) to ESS storage subsystems.
If your data migration has to do with moving data from one ESS to another ESS you still can
use the techniques described so far. But for this ESS to ESS migration scenarios you also
have the additional alternative of using the ESS Copy Services functions. More precisely you
can take advantage of the optimized ESTABLISH and RESYNCH operations over PPRC
pairs of logical volumes at long distances using the INRANGE Fibre Channel Directors (see
Section, “INRANGE” on page 22).

Chapter 10. Data migration in the zSeries environment
175
10.6 IBM migration services
This is the easiest way to migrate data, because IBM will assist you throughout the complete
migration process. In several countries IBM offers a migration service. Check with your IBM
sales representative about migration services for your specific environment and needs.

176

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment

© Copyright IBM Corp. 2002
177
Chapter 11.
Managing and monitoring
the ESS
In this chapter we present the tools you will be using for managing and monitoring the IBM
TotalStorage Enterprise Storage Server in the S/390 storage environments. Because for
S/390 environments, the ESS emulates 3990 storage controls with 3390 (or 3380) storage
devices attached, the basic management and monitoring processes and tools that have been
used with previous 3990 and RAMAC storage controllers are also applicable to the ESS.
However, there are some changes relating to the specific characteristics and functions of the
ESS which are covered in this chapter.
11

178

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
11.1 General considerations
The following general items should be checked regarding systems management discipline:

Have the system management procedures been updated to reflect the new hardware?

Does a firewall need to be created to protect against unauthorized logins?

Has a process to maintain the SNMP alert e-mail addresses been established?

Have you set up a TCP/IP address on your local network for the ESS?

Has a firewall been set up to keep unauthorized users away from the ESS?

Have the user IDs and addresses been established for configuring, administering and
viewing the ESS?
11.2 Managing tools
Typically the management of the ESS consist of activities like configuration changes, storage
capacity re-allocations, enablement-disablement of ESS resources and functions. The IBM
TotalStorage Enterprise Storage Server Specialist (ESS Specialist) program and the Control
Unit Initiated Reconfiguration (CUIR) function will assist you in these activities.
11.2.1 IBM TotalStorage Enterprise Storage Server Specialist
The tool you will be using most of the time for managing the ESS is the IBM TotalStorage
Enterprise Storage Server Specialist (ESS Specialist). Any configuration related activity, like
changes related to server attachment, or changes related to storage capacity assignment, will
need the use of the ESS Specialist. Chapter 5, “IBM TotalStorage Enterprise Storage Server
Specialist” on page 79, describes this management tool.
11.2.2 Control Unit Initiated Reconfiguration (CUIR)
The Control Unit Initiated Reconfiguration (CUIR) function automates channel path
quiesce/resume actions to minimize manual operator intervention during selected ESS
service or upgrade actions.
CUIR allows the IBM Service Support Representative (IBM SSR) to request that all attached
system images set all paths associated with a particular service action offline. System images
with the appropriate level of software support will respond to such requests by varying off the
affected paths, and either notifying the ESS subsystem that the paths are offline, or that it
cannot take the paths offline. CUIR reduces manual operator intervention and the possibility
for human error during maintenance actions, at the same time reducing the time required for
the maintenance. This is particularly useful in environments where there are many systems
attached to an ESS.
CUIR is supported on zSeries and S/390 platforms and requires ESS Licensed Internal Code
(LIC) level 1.5.0, or later. CUIR support is not available for open systems. Refer toSection 7.5,
“CUIR support” on page 100 for information regarding CUIR software support.
The CUIR function can be enabled/disabled on each ESS subsystem selectively. The IBM
SSR will enable or disable CUIR during the initial ESS install sequence based on information
provided by the customer on the Communication Resources work sheet (option
Allow CUIR
to automatically vary paths off and on
). See Section 8.3.3, “Communication Resources
worksheet” on page 117. The options are:

Enable:
The ESS will initiate the reconfiguration for service.

Disable:
The operator will set the paths offline for service.

Chapter 11. Managing and monitoring the ESS
179
The IBM SSR can later change the original configuration option setting.
When CUIR is enabled, the ESS invokes CUIR as part of selected service processes. When
the service operation reaches a stage where it becomes necessary to take a disruptive action
that would involve paths being disabled, the IBM SSR can request a quiesce of the channel
paths from the ESS subsystem. ESS sends a reconfiguration request to all attached
operating systems requesting that they take the appropriate paths offline.
When an operating system receives the request from the ESS, it determines the paths
affected by the request, waits until in-process I/O operations are complete, and then takes the
appropriate paths offline, if not so already, and marks them as in use by CUIR. CUIR will not
take the last path offline from any processor image to an online device.
The following MVS messages are issued in response to a successful quiesce request (see
Figure 11-1).
Figure 11-1 MVS messages for quiesce request
When the paths are offline, the host system acknowledges to the ESS that the paths have
now been quiesced and may be disabled for the service action. After ESS has received
positive response from all the affected systems, the IBM SSR can perform the service action.
If the quiesce request was unsuccessful, the process notifies the SSR to request the system
operator to issue the host commands to vary the necessary channel paths offline. Possible
reasons for a quiesce request to fail are:

No response from the host system

The host does not support CUIR

The request would take the last path offline to a device.

The vary path process failed
When the paths are no longer required for the service action, the IBM SSR resumes the I/O
components that were quiesced. The ESS sends a resume request to all attached host
systems to bring the paths back online. The operating system determines which channel
paths are affected and varies them online automatically.
MVS issues the following messages in response to a successful resume request (see
Figure 11-2):
Figure 11-2 MVS messages for resume request
A path quiesced by CUIR is logically offline to the host. Use the
DEVSERV PATHS
command to
display path status. Figure 11-4 on page 188 shows the output of the
DSERV PATHS
command.
IOS275I C.U.I.R. REQUEST TO QUIESCE THE FOLLOWING PATH(S):
CHPID A0 TO DEVICE(S) 2E00-2E3F
IOS281I C.U.I.R. REQUEST SUCCESSFUL
IOS278I C.U.I.R. REQUEST TO RESUME THE FOLLOWING PATH(S):
CHPID A0 TO DEVICE(S) 2E00-2E3F
IOS281I C.U.I.R. REQUEST SUCCESSFUL

180

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
You cannot vary online channel paths that are taken offline by CUIR. In MVS, an attempt will
fail with message IEE169I (see Figure 11-3).
Figure 11-3 Vary path online rejected
If you know that the path is no longer being serviced, use MVS command
VARY PATH((dev1-dev2),xx),ONLINE,FORCE
to vary the path online. It is possible that the path
was never varied back online by the CUIR service after it completed a service action involving
the path. This situation could occur if the service representative neglected to resume the path
after finishing the service action or if the resume failed.
CUIR supports both ESCON and FICON channel paths.
CUIR does not have automatic quiesce support for ESCON links used for PPRC. If a service
action requires a PPRC port to be disabled, the service representative will disable it manually
using the service terminal. PPRC will continue to operate with the remaining paths. When the
service action is complete, the IBM SSR will enable the port and PPRC restarts using the link.
If you only have a few PPRC paths defined, quiescing one of them may impact performance.
If possible, add a temporary link with the PPRC
ESTPATH
command in place of the affected
link. Otherwise, schedule the service action at a time of low write activity.
Example 11-1 is a list of MVS messages related to CUIR. For details, please see
z/OS MVS
System Messages, Vol 9,
SA22-7639.
Example 11-1 MVS messages related to CUIR
IOS275I C.U.I.R. REQUEST TO QUIESCE THE FOLLOWING PATH(S): CHPID xx TO
DEVICE(S) dev,dev1-dev2, ... CHPID yy TO DEVICE(S) dev,dev1-dev2, ...
IOS278I C.U.I.R. REQUEST TO RESUME THE FOLLOWING PATH(S): CHPID xx TO
DEVICE(S) dev,dev1-dev2, ... CHPID yy TO DEVICE(S) dev,dev1-dev2, ...
IOS281I C.U.I.R. REQUEST SUCCESSFUL
IOS283I C.U.I.R. VARY PATH(dev,xx) REJECTED, text
IOS284I C.U.I.R. REQUEST REJECTED - VARY COMMAND PROCESSOR FAILED
IOS290I C.U.I.R. REQUEST UNSUCCESSFUL
11.3 Monitoring tools
Several tools help you in the activity of monitoring the ESS. As already mentioned, most of
these tools work basically in the same manner as they did with the previous IBM disk storage
controllers (IBM 3990s, RAMACs). In the following sections we describe these tools,
highlighting the ESS specifics:

RMF

CRR and DFSMS Optimizer

SMF

SIM and EREP

IDCAMS

MVS system commands
VARY PATH(2E00,A4),ONLINE
IEE169I VARY REJECTED, PATH(2E00,A4) OFFLINE DUE TO C.U.I.R.

Chapter 11. Managing and monitoring the ESS
181
IBM StorWatch Enterprise Storage Server Expert
The IBM StorWatch Enterprise Storage Server Expert (ESS Expert) is a managing and
monitoring tool specifically oriented for the ESS administration. The ESS Expert is covered in
Chapter 6, “IBM StorWatch Enterprise Storage Server Expert” on page 87 of this redbook.
11.3.1 RMF
RMF reports regarding ESS performance are almost the same as for previous IBM 3990
storage controllers. Enhancements have been provided for PAV reporting, and FICON
channel reporting.
Device Activity report
Response, connect, disconnect, PEND and IOSQ times are reported for a corresponding
device address. Alias addresses for PAVs are not reported separately, but RMF will report the
number of aliases that have been used by a device.
Example 11-2 shows an extract from an RMF Device Activity report. The PAV column shows
the number of PAV
exposures
(base plus aliases) for the device at the end of the reporting
interval. If the number has been changed during the reporting interval, it is followed by an
asterisk (*).
Example 11-2 RMF Device Activity report
D I R E C T A C C E S S D E V I C E A C T I V I T Y
z/OS V1R2 SYSTEM ID XXXX START 11/13/2001-09.00.00
RPT VERSION V1R2 RMF END 11/13/2001-09.30.00
TOTAL SAMPLES = 3,600 IODF = 02 CR-DATE: 09/20/2001 CR-TIME: 15.48.19 ACT: POR
DEVICE AVG AVG AVG AVG AVG AVG AVG AVG
STORAGE DEV DEVICE VOLUME PAV LCU ACTIVITY RESP IOSQ DPB CUB DB PEND DISC CONN
GROUP NUM TYPE SERIAL RATE TIME TIME DLY DLY DLY TIME TIME TIME
SYSTEM 9000 33909 SYS201 4 001C 4.887 1.5 0.0 0.0 0.0 0.3 0.2 1.0
GROUP01 9001 33909 PUB201 4 001C 0.661 1.3 0.0 0.0 0.0 0.3 0.3 0.7
GROUP01 9002 33909 PUB202 4 001C 0.206 1.5 0.0 0.0 0.0 0.3 0.5 0.7
GROUP01 9003 33909 PUB203 4 001C 11.675 1.8 0.0 0.0 0.0 0.3 0.8 0.7
GROUP01 9004 33909 PUB204 4 001C 6.689 1.1 0.0 0.0 0.0 0.3 0.3 0.5
GROUP01 9005 33909 PUB205 4 001C 0.132 2.4 0.0 0.0 0.0 0.3 1.3 0.8
GROUP01 9006 33909 PUB206 4 001C 0.066 4.3 0.0 0.0 0.0 0.3 2.3 1.7
WORK 9007 33909 WRK201 4 001C 0.001 0.8 0.0 0.0 0.0 0.4 0.0 0.4
GROUP01 9008 33909 PUB207 4 001C 0.359 7.6 0.0 0.0 0.0 0.4 3.4 3.8
GROUP01 9009 33909 PUB208 4 001C 0.004 3.3 0.0 0.0 0.0 0.3 2.7 0.4
GROUP01 900A 33909 PUB209 4 001C 0.214 2.8 0.0 0.0 0.0 0.3 2.0 0.5
Additional information about usage of alias addresses can be found in SMF record, type 74,
subtype 5. RMF does not use this information.

182

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Besides the Device Activity report showing the number of PAV exposures in its PAV column,
also the following RMF reports have an additional column to show the number of aliases:

Mon I / Postprocessor: DASD Activity report, Shared DASD Activity report

Mon II: DEV report, DEVV report

Mon III: DEVR report, DSNV report
Workload Activity report
The RMF Workload Activity report has been enhanced to show the I/O Priority Management
status and WLM Dynamic Alias Management status on the Service Policy Page (see
Example 11-3).
Example 11-3 RMF Workload Activity report
W O R K L O A D A C T I V I T Y
z/OS V1R1 SYSPLEX PLEX2 START 11/14/2001-09.00.00 INTERVAL 000.30.01 MODE = GOAL
CONVERTED TO z/OS V1R2 RMF END 11/14/2001-09.30.00
POLICY ACTIVATION DATE/TIME 11/09/2001 12.15.48
- SERVICE POLICY PAGE -
SERVICE DEFINITION: servdef1 WLM definitions -SERVICE DEFINITION COEFFICIENTS-
INSTALL DATE: 11/09/2001 12.15.35 INSTALLED BY: XXXXX IOC CPU SRB MSO
POLICY: FULLTIME Standard WLM definition
I/O PRIORITY MANAGEMENT: YES 0.5 1.0 1.0 0.0000
DYNAMIC ALIAS MANAGEMENT: YES
FICON Channel Path reporting
RMF support for FICON channels (both FCV and FC) includes the reporting of five new
measurements:

Bus utilization

Percentage of bus cycles the bus has been found busy for this channel in relation to the
theoretical limit.

Read bandwidth for an image in MB/sec

Data transfer rates from the control unit to the channel for this individual logical partition.

Total read bandwidth (for all images on the system) in MB/sec

Data transfer rates from the control unit to the channel for the entire system.

Write bandwidth for an image in MB/sec

Data transfer rates from the channel to the control unit for this individual logical partition.

Total write bandwidth (for all images on the system) in MB/sec

Data transfer rates from the channel to the control unit for the entire system.
Sample RMF Channel Path Activity report information is shown in Example 11-4.

Chapter 11. Managing and monitoring the ESS
183
Example 11-4 RMF Channel Path Activity report
RMF V1R2 CHANNEL PATH ACTIVITY LINE 25 OF 47
COMMAND ===> SCROLL ===> CSR
SAMPLES: 119 SYSTEM: XXXX DATE: 11/12/01 TIME: 12.44.00 RANGE: 120 SEC
CHANNEL PATH UTILIZATION(%) READ(B/S) WRITE(B/S) MSG MSG SEND RECV
ID NO G TYPE S PART TOT BUS PART TOT PART TOT RATE SIZE FAIL FAIL
90 FC Y 3.4 7.0 15.4 872K 2M 88K 406K
91 FC Y 3.5 7.0 15.3 870K 2M 96K 415K
92 FC Y 3.4 7.0 15.4 875K 2M 94K 412K
A0 FC Y 2.6 7.0 15.7 386K 2M 93K 635K
A1 FC Y 2.6 7.2 15.8 430K 2M 95K 632K
A2 FC Y 2.6 7.2 15.8 426K 2M 93K 639K
A3 CNC_P Y 1.1 1.1
B0 FC Y 0.2 1.7 13.5 48K 988K 56K 235K
B1 FC Y 0.2 1.7 13.5 56K 995K 54K 223K
B2 FC Y 0.2 1.6 13.5 56K 1M 42K 213K
C0 FC Y 0.1 1.1 13.0 21K 665K 27K 115K
C1 FC Y 0.1 1.0 13.0 23K 654K 17K 73K
C2 FC Y 0.1 1.1 13.0 19K 665K 27K 84K
11.3.2 CRR and DFSMS Optimizer
Two tools also have available for monitoring the ESS performance are the Cache RMF
Reporter (CRR) that collects data and produces its own reports, and the DFSMS Optimizer.
CRR also collects certain type of records for DFSMS Optimizer reporting.
Cache RMF Reporter
The Cache RMF Reporter (CRR), that is part of RMF, has been enhanced to produce a new
RAID Rank Activity report, which shows I/O statistics for the RAID ranks in the ESS. The
sample output in Example 11-5 shows measurements for two RAID ranks, 0000 and 0001.
Example 11-5 CRR RAID Rank Activity report
------------------------------------------------------------------------------------------
RAID RANK ACTIVITY
------------------------------------------------------------------------------------------
ID RAID DA HDD -------- READ REQ ------- ------- WRITE REQ -------
TYPE RATE AVG MB MB/S RTIME RATE AVG MB MB/S RTIME
*ALL 14 94 0.037 3.5 17 31 0.023 0.7 101
0000 RAID-5 01 7 85 0.039 3.3 18 20 0.025 0.5 108
0001 RAID-5 01 7 9 0.018 0.2 12 11 0.020 0.2 87
DFSMS Optimizer
The DFSMS Optimizer collects and stores historical system activity data in a database. The
extraction process captures data from SMF, RMF, HSM, and the cache RMF reporter, or CRR
(the data collection component of CRR is distributed with the DFSMS Optimizer), and
compresses this data by 80% to 90% into the DFSMS Optimizer database, allowing you to
store months of data at a time. The database is used for DFSMS Optimizer analyses and
reports.

184

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
A major capability of the DFSMS Optimizer is its ability to analyze and simulate your storage
environment based on its database. The analyses possible are:

Hardware performance analysis

Data set performance analysis

DFSMShsm performance analysis

Management class analysis
The two first reports can help you when analyzing the storage subsystem performance:

Hardware performance analysis:
The DFSMS Optimizer provides performance
information about your storage subsystem in the form of reports and charts to help you
isolate the cause of hardware performance problems such as device skew, I/O
bottlenecks, or I/O intensive periods. This information is used to balance the I/O workloads
across your subsystem.

Data set performance analysis:
The DFSMS Optimizer uses SMF and RMF data to
provide data-set-level statistics on the use of cache. Using cost components that you
provide, it reports the cost of cache resources associated with the storage class policies
and the data sets to which those policies apply.
11.3.3 SMF
SMF record type 99 subtype 7 contains summary information on alias usage for an ESS with
Parallel Access Volume (PAV) definitions enabled. A subtype 7 record is written every third
policy interval. Turn on SMF type 99 recording if desired; this data can be used for analysis of
alias management activity if it becomes necessary. Refer to the
z/OS System Managemet
Facilities (SMF),
SA22-7630 for a description of the record layout.
11.3.4 SIM and EREP
The Environmental Recording, Editing, and Printing (EREP) program provides problem
incident reports for the new device type 2105. However, the ESS maintenance strategy does
not rely on the analysis of data in EREP reports. Service Information Messages (SIMs) and
problem records in the ESS internal problem log contain information necessary for the IBM
SSR to start a service action.
Sense data records for some ESS temporary and all permanent errors are sent from the ESS
to the system to give information necessary to perform needed system error recovery
procedures. The ESS sense data is logged in the error-recording data set (ERDS) in the
system, but it is not used for ESS problem determination by the IBM SSR. The ESS
generates and send to the host SIMs whenever ESS service is needed. The SIMs summarize
the service information necessary to isolate and repair ESS error conditions. SIMs are
presented as host system console messages and are logged in the ERDS. The SIM ID is the
same as the problem number in the ESS problem log.
Even if EREP reports are not generally required for ESS maintenance, you may still want to
run the EREP Service Information Message report (see Example 11-6) on a regular basis as
it gives a compact summary of SIMs generated during the reporting period and serves as a
checklist of problem incidents. Other system exception reports might contain ESS
information. The other reports may occasionally be required by the IBM SSR so you should
be prepared to run the reports if necessary.

Chapter 11. Managing and monitoring the ESS
185
Example 11-6 Service Information Messages report
SERVICE INFORMATION MESSAGES REPORT DATE 024 99
PERIOD FROM 021 99
TO 022 99
******************************************************************************************
COUNT FIRST OCCURRENCE LAST OCCURRENCE
1 021/99 17:44:27:78 021/99 17:44:27:78
MODERATE ALERT 2105-E20 S/N 0113-10473 REFCODE C211-1060-A00A ID=03
DASD EXCEPTION ON SSID 0011
ADDITIONAL ANALYSIS REQUIRED TO DETERMINE REPAIR IMPACT.
SEE PROBLEM NUMBER 03 FOR DETAILS
2 021/99 19:24:19:56 021/99 19:24:19:56
SERVICE ALERT 2105-E20 S/N 0113-30224 REFCODE 4320-0000-5284 ID=06
MEDIA EXCEPTION ON SSID 00D2, VOLSER 380050 DEV 0E12, 0D
REFERENCE MEDIA MAINTENANCE PROCEDURE 2
3 021/99 19:24:04:67 022/99 03:29:01:65
SERIOUS ALERT 2105-E20 S/N 0113-10473 REFCODE C211-1060-A00A ID=09
DASD EXCEPTION ON SSID 00D2
ADDITIONAL ANALYSIS REQUIRED TO DETERMINE REPAIR IMPACT.
SEE PROBLEM NUMBER 09 FOR DETAILS
11.4 Logical subsystem resource usage
LISTDATA
and
SETCACHE
are AMS commands used to display logical subsystem resource
usage and to control caching functions of IBM storage controllers. Both commands are
changed to support ESS.
The
LISTDATA
command now generates a new RAID Rank Counters report and a modified
Subsystem Counters report.
SETCACHE
, on the other hand, has been updated to produce
messages indicating that ESS does not support the 3990 Dual Copy function or the disabling
of Cache or DASD Fast Write (DFW).
11.4.1 IDCAMS LISTDATA
IDCAMS
LISTDATA
command generates reports on activity within all IBM caching models of
3880, 3990, and RAMAC storage controllers.
With the ESS you can also use IDCAMS
LISTDATA
to request activity information. You will
receive a modified Subsystem Counters report and a new RAID Rank Counters report when
you issue the
LISTDATA COUNTS SUBSYSTEM
or the
LISTDATA COUNTS ALL
command.
LISTDATA
COUNTS DEVICE
command

generates the

Subsystem Counters report, but not the RAID Rank
Counters report.
IDCAMS Subsystem Counters report
The Subsystem Counters report is changed to include the RAID RANK ID of a logical volume.
This identifier is required for a cross reference to the new RAID Rank Counters report.
Example 11-7 presents a Subsystem Counters report including this identifier. The command
used in this case was
LISTDATA COUNTS VOLUME(volser) UNIT(3390) SUBSYSTEM.

186

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Example 11-7 IDCAMS Subsystem Counters report
IDCAMS SYSTEM SERVICES
2105 STORAGE CONTROL
SUBSYSTEM COUNTERS REPORT
VOLUME SYSXXX DEVICE ID X'20'
SUBSYSTEM ID X'5100'
CHANNEL OPERATIONS
......SEARCH/READ..... ..............WRITE...............
TOTAL CACHE READ TOTAL DASDFW CACHE WRITE
REQUESTS
NORMAL 5882213 5811266 17158 17058 17058
SEQUENTIAL 4370514 4364583 194108 194050 194050
CACHE FAST WRITE 0 0 0 N/A 0
TOTALS 10252727 10175849 211266 211108 211108
REQUESTS CHANNEL OPERATIONS
INHIBIT CACHE LOADING 0
BYPASS CACHE 0
TRANSFER OPERATIONS DASD/CACHE CACHE/DASD
NORMAL 101440 851628
SEQUENTIAL 926668 N/A
DASD FAST WRITE RETRIES 0
DEVICE STATUS CACHING: ACTIVE
DASD FAST WRITE: ACTIVE
DUPLEX PAIR: NOT ESTABLISHED
RAID RANK ID X'0001'
LEGEND
SUBSYSTEM COUNTERS LEGEND
VOLUME - VOLUME SERIAL NUMBER FOR WHICH THE DATA IS GATHERED
DEVICE ID - CHANNEL CONNECTION ADDRESS OF THE DEVICE ON WHICH
THE I/O WAS DONE
SUBSYSTEM ID - SUBSYSTEM TO WHICH THE DEVICE IS ATTACHED
:
: (Legend removed)
:
RAID RANK ID - THE RAID RANK ID ASSOCIATED WITH THIS VOLUME
IDCAMS RAID Rank Counters report
The RAID Rank Counters report is produced to support ESS. You can generate this report by
issuing the
LISTDATA COUNTS SUBSYSTEM
or the
LISTDATA COUNTS ALL
command. Generated
after the Subsystem Counters report, the RAID Rank Counters report contains summary data
for the whole RAID rank. Example 11-8 shows a RAID Rank Counters report (
Note
: due to
hardware restrictions, RAID Rank data is only available when the device with the lowest unit
address in a RAID rank is online).
Example 11-8 IDCAMS RAID Rank Counters report
IDCAMS SYSTEM SERVICES
2105 STORAGE CONTROL
SUBSYSTEM COUNTERS REPORT
RAID RANK COUNTERS
SUBSYSTEM ID X'5100'
RAID RANK ID X'0001'
DEVICE ADAPTER ID X'01'
NUMBER OF HDDS IN RAID RANK 8
HDD SECTOR SIZE 524
RAID RANK OPERATIONS
REQUESTS I/O REQUESTS RESP/TIME SECTOR REQUESTS

Chapter 11. Managing and monitoring the ESS
187
READ 25770871 0.012 2342659927
WRITE 7982700 0.111 1315104518
LEGEND
RAID RANK COUNTERS LEGEND
SUBSYSTEM ID - SUBSYSTEM TO WHICH THE RAID RANK IS ATTACHED
NUMBER OF HDDS - NUMBER OF HARD DISK DRIVES IN RAID RANK
HDD SECTOR SIZE - SIZE IN BYTES OF A PHYSICAL HDD IO REQUEST
RANK OPERATIONS - OPERATIONS ASSOCIATED WITH A RAID RANK
I/O REQUESTS - NUMBER OF I/O REQUESTS
RESP/TIME - AVERAGE RESPONSE TIME (MS)
SECTOR REQUESTS - NUMBER OF PHYSICAL HDD IO REQUESTS
READ - OPERATIONS CONTAINING AT LEAST ONE SEARCH OR READ
COMMAND BUT NO WRITE COMMANDS
WRITE - OPERATIONS CONTAINING AT LEAST ONE WRITE COMMAND
11.4.2 IDCAMS SETCACHE
Traditionally, you use the IDCAMS
SETCACHE
command to allow or prohibit access to IBM
caching models of 3880 and 3990 storage controllers. For example, you can use
SETCACHE
to
set Dual Copy and Cache/DFW.
However, on the ESS Cache and DFW are ON by default, and you are not allowed to modify
them. In addition, ESS does not support the Dual Copy function. You will receive message
IDC31562I with return code 12 indicating that a parameter is not available if you issue any of
the following Cache/DFW or Dual Copy commands:
SETCACHE SETSECONDARY
SETCACHE SUSPENDPRIMARY
SETCACHE SUSPENDSECONDARY
SETCACHE RESETTODUPLEX
SETCACHE REESTABLISHDUPLEX
SETCACHE RESETTOSIMPLEX
SETCACHE DFW OFF
SETCACHE NVS OFF
SETCACHE DEVICE OFF
SETCACHE SUBSYSTEM OFF
Example 11-9 displays the message given in response to the
SETCACHE SETSECONDARY

command.
Example 11-9 SETCACHE response message
SETCACHE SETSECONDARY(D00E) FILE(FILEX)
IDC31562I THE SETSECONDARY PARAMETER IS NOT AVAILABLE FOR THE SPECIFIED
IDC31562I SUBSYSTEM OR DEVICE
IDC3003I FUNCTION TERMINATED. CONDITION CODE IS 12
11.5 MVS system commands
MVS system commands
DEVSERV
and
DISPLAY
have been enhanced in support of ESS.

188

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
11.5.1 DEVSERV command
In z/OS, the
DEVSERV
command can be used to obtain status information about DASD and
tape subsystems. The command response is a display of basic status information about a
device, a group of devices, or storage control units, and can optionally include a broad range
of additional information. With the new parameter, QPAVS, you can display base volumes and
their aliases.
DEVSERV PATHS
The
DEVSERV PATHS
command (
DS P
) displays useful information on the state of a device and
the paths used to access it from the host. Figure 11-4 shows the output (IEE459I message)
for the
DEVSERV PATHS
command for three devices in an ESS, beginning at address 2E00
(
DS P,2E00,3
).
The information returned includes the subsystem ID (SSID), the VOLSER for each device,
and indication of the result for each channel path of an attempt to communicate with the
device over that path. The plus (+) sign for a path in the example indicates successful
communication. The dynamic legend at the bottom defines the symbols that appear in the
output.
Figure 11-4 DEVSERV PATHS output
An important attribute of the
DEVSERV PATHS
command is that the state of the subsystem is
obtained by performing I/O operations at the time of command execution. Apart from the
VOLSER field (which is obtained from the UCB), the information is read directly from the disk
subsystem. For example, if a channel path has a plus (+) sign, you know that the path was
working at the time you entered the
DS P
command.
The
DEVSERV PATHS
command can be performed to both online and offline devices. It will not
display the VOLSER for an offline device (the VOLSER is not in the UCB).
DEVSERV QUERY DASD
The
DEVSERV QUERY DASD
command (
DS QD
) is used to obtain information about a DASD
device. Figure 11-5 shows the output of
a DS QD
to an ESS subsystem.
DS P,2E00,3
IEE459I 16.00.12 DEVSERV PATHS 340
UNIT DTYPE M CNT VOLSER CHPID=PATH STATUS
RTYPE SSID CFW TC DFW PIN DC-STATE CCA DDC ALT CU-TYPE
2E00,33903 ,O,000,MC2E00,A4=* A5=+ A6=+ A7=+ F8=+ F9=< FA=+ FB=+
2105 6290 Y YY. YY. N SIMPLEX 00 00 2105
2E01,33903 ,O,000,MC2E01,A4=* A5=+ A6=+ A7=+ F8=+ F9=< FA=+ FB=+
2105 6290 Y YY. YY. N SIMPLEX 01 01 2105
2E02,33903 ,A,000,MC2E02,A4=* A5=+ A6=+ A7=+ F8=+ F9=< FA=+ FB=+
2105 6290 Y YY. YY. N SIMPLEX 01 01 2105
************************ SYMBOL DEFINITIONS ************************
A = ALLOCATED O = ONLINE
+ = PATH AVAILABLE * = LOGICALLY OFF, PHYSICALLY ON
< = PHYSICALLY UNAVAILABLE

Chapter 11. Managing and monitoring the ESS
189
Figure 11-5 DEVSERV QUERY DASD output
The
DEVSERV QUERY DASD
command with parameter UCB displays the device UCB as shown
in Figure 11-6.
Figure 11-6 DEVSERV QUERY DASD UCB output
Examples of other data areas that can be displayed using the
DEVSERV QUERY DASD
command
are DCE, DCPT, and SSSCB. Other
QD
parameters such as RDC (read device
characteristics), RCD (read configuration data), and SNSS (sense subsystem status) return
information directly from the hardware. This can assist in determining the hardware
configuration of a logical device.
DEVSERV QPAVS
DEVSERV
is a z/OS system command used to request basic status information on a device, a
group of devices, or storage control units. You can use
DEVSERV
with the
QPAVS
parameter to:

Describe how a logical subsystem configuration is defined to z/OS

Highlight the inconsistencies, if any, between IODF and the LSS configuration

Display unbound alias device types with the UCB parameter and, if necessary, unbox a
boxed alias device with the
UNBOX
parameter

Show information on both a PAV-base address and its PAV-aliases by specifying the
VOLUME
parameter

Display information on devices
DS QD,9000,3
IEE459I 13.31.27 DEVSERV QDASD 461
UNIT VOLSER SCUTYPE DEVTYPE CYL SSID SCU-SERIAL DEV-SERIAL EF-CHK
9000 SYS201 2105F20 2105 10017 2100 0175-12345 0175-12345 **OK**
9001 PUB201 2105F20 2105 10017 2100 0175-12345 0175-12345 **OK**
9002 PUB202 2105F20 2105 10017 2100 0175-12345 0175-12345 **OK**
**** 3 DEVICE(S) MET THE SELECTION CRITERIA
**** 0 DEVICE(S) FAILED EXTENDED FUNCTION CHECKING
DS QD,9000,UCB
IEE459I 13.32.00 DEVSERV QDASD 505
UNIT VOLSER SCUTYPE DEVTYPE CYL SSID SCU-SERIAL DEV-SERIAL EF-CHK
9000 SYS201 2105F20 2105 10017 2100 0175-12345 0175-12345 **OK**
UCB AT V00E58FA0
00A8FF8C90000000 0000000008E4C3C2 3030200F00E58F78 000F0100E2E8E2F2
F0F1104800A00197 00E58DA0020209C8 0080030300000000
UCB PREFIX AT V0211BDB0
000E804000000000 0000000000010583 289C001FE00080E0 9091920000000000
0108000000000001
UCB COMMON EXTENSION AT V00E58F78
0000094020AA0008 0211BDB00000011D 0000000000FBE93C 00E58F4000000000
**** 1 DEVICE(S) MET THE SELECTION CRITERIA
**** 0 DEVICE(S) FAILED EXTENDED FUNCTION CHECKING

190

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Based on your service needs, you can issue the
DEVSERV QPAVS
command in the following
forms:
DEVSERV QPAVS,dddd,nn
DEVSERV QPAVS,dddd,tttt
DEVSERV QPAVS,SSID=ssss
The parameters in these commands have the following values:
QPAVS
is a required positional parameter
dddd
is a 3 or 4-hex-digit device number
nn
is a decimal number from 1-256 with 1 as the default
tttt
can be
UCB
,
VOLUME
, or
UNBOX
ssss
is the specified SSID value
You can refer to the
z/OS MVS System Commands
for additional information on the
DEVSERV

command. Also the
IBM TotalStorage DFSMS Software Support Reference
, SC26-7440
manual can be referenced for additional information on the
DEVSREV
command.
Figure 11-7 and Figure 11-8 show the command output for two different
DEVSERV QPAV

commands. The first displays the requested device and all its base and alias affiliations. The
second displays the requested devices in order.
The form of the command,
DS QP,SSID=ssid
displays the information for all devices in the
requested LCU. The
SSID=
parameter can be a long running command as it issues I/O to
every DASD drive in the host’s hardware configuration and then only prints the drives with the
SSID. This allows for checking of duplications of the SSID which must be unique to each LCU.
Figure 11-7 DEVSERV QPAVS VOLUME output
Explanation of the output fields:
UNIT NUM
.
MVS device number (address)
UA
Unit address from host configuration
TYPE
Device type from host configuration: BASE, ALIAS-dddd, or NON-PAV
SSID
SSID where the device belongs to
UNIT ADDR
Unit address (device ID) from ESS subsystem configuration
UA TYPE
Device type from subsystem configuration: BASE, ALIAS-bb, or NC (not a base
not an alias)
DS QP,D222,VOLUME
IEE459I 08.20.32 DEVSERV QPAVS 591
HOST SUBSYSTEM
CONFIGURATION CONFIGURATION
------------- --------------------
UNIT UNIT UA
NUM. UA TYPE STATUS SSID ADDR. TYPE
---- -- ---- ------ ---- ---- ------------
D222 22 BASE 0102 22 BASE
D2FE FE ALIAS-D222 0102 FE ALIAS-22
D2FF FF ALIAS-D222 0102 FF ALIAS-22
**** 3 DEVICE(S) MET THE SELECTION CRITERIA

Chapter 11. Managing and monitoring the ESS
191
STATUS
Indicates any discrepancies between the host and the ESS subsystem
configurations. Possible values are:
INV-ALIAS
On the host side, the unit is defined as an ALIAS whose BASE is
different from the one on the subsystem side.
NOT-BASE
On the host side, the unit is a BASE, while on the subsystem side,
it is not.
NOT-ALIAS
On the host side, the unit is an ALIAS, while on the subsystem
side, it is not.
NON-NPAV
On the host side, the unit is not a BASE nor an ALIAS, while on the
subsystem side, it is an ALIAS.
Figure 11-8 DEVSERV QPAVS output
Figure 11-9 shows a
DEVSERV QPAVS
output with devices for which the device information was
not listed, and the corresponding reason code.
Figure 11-9 Unlisted devices
DS QP,D123,3
IEE459I 08.20.32 DEVSERV QPAVS 591
HOST SUBSYSTEM
CONFIGURATION CONFIGURATION
------------- --------------------
UNIT UNIT UA
NUM. UA TYPE STATUS SSID ADDR. TYPE
---- -- ---- ------ ---- ---- ------------
D123 23 BASE 0101 23 BASE
D124 24 BASE 0101 24 BASE
D124 25 BASE 0101 25 BASE
**** 3 DEVICE(S) MET THE SELECTION CRITERIA
DS QPAVS,C15F,4
IEE459I 09.18.55 DEVSERV QPAVS 844
HOST SUBSYSTEM
CONFIGURATION CONFIGURATION
------------- --------------------
UNIT UNIT UA
NUM. UA TYPE STATUS SSID ADDR. TYPE
---- -- ---- ------ ---- ---- ------------
**** UNLISTED DEVICE(S) AND REASON CODES :
C15F(07) C160(0A) C161(0A) C162(0A)
**** (07) - DEVICE I/O ERROR
**** (0A) - DEVICE IS AN UNBOUND PAV-ALIAS
**** 0 DEVICE(S) MET THE SELECTION CRITERIA

192

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
The following list shows the reason codes for unlisted devices:
(01) - DEVICE NOT CONFIGURED,UCB NOT FOUND
(02) - UCB NOT CONNECTED
(03) - DEVICE UNAVAILABLE,SCP ROUTINE IN CONTROL
(04) - SUBCHANNEL ERROR
(05) - DEVICE BOXED
(06) - UCB NOT A DASD
(07) - DEVICE I/O ERROR
(08) - DEVICE IS NOT A DASD
(09) - DSE-1 CCW BUILD FAILED
(0A) - DEVICE IS AN UNBOUND PAV-ALIAS
11.5.2 DISPLAY MATRIX command
An enhancement to the MVS
DISPLAY MATRIX
(
D M
) command causes information on PAV
aliases to be displayed for an ESS.
DISPLAY M=CHP
The
DISPLAY M=CHP
command displays the online and offline status of channel paths. Refer to
Figure 11-10 for a sample output from
DISPLAY MATRIX
command for a channel
(
D M=CHP(80)
), showing the devices configured on the channel, and the aliases.
Figure 11-10 Display Channel path status
Symbol
AL
indicates an
active
alias device. Symbol
UL
indicates an
unbound
alias device. The
alias device is inactive, not bound to any base device. For example, when a base device is
varied offline, its alias devices become unbound.
D M=CHP(80)
IEE174I 10.05.24 DISPLAY M 779
CHPID 80: TYPE=05, DESC=ESCON SWITCHED POINT TO POINT
DEVICE STATUS FOR CHANNEL PATH 80
0 1 2 3 4 5 6 7 8 9 A B C D E F
680 + + + + + + + + + + + + + + + +
681 + + + + + + + + + + + + + + + +
682 + + + + + + + + + + + + + + + +
683 + + + + + + + + + + + + + + + +
684 AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL
685 AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL
686 AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL
687 AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL AL
688 + + + + + + + + + + + + + + + +
689 + + + + + + + + + + + + + + + +
68A + + + + + + + + + + + + + + + +
68B + + + + + + + + + + + + + + + +
68C UL UL UL UL UL UL UL UL UL UL UL UL UL UL UL UL
68D UL UL UL UL UL UL AL AL AL AL AL AL AL AL AL AL
68E UL UL UL UL UL UL UL UL UL UL UL UL UL UL UL UL
68F UL UL UL UL UL UL UL UL UL UL UL UL UL UL UL UL
************************ SYMBOL EXPLANATIONS ************************
+ ONLINE @ PATH NOT VALIDATED - OFFLINE . DOES NOT EXIST
* PHYSICALLY ONLINE $ PATH NOT OPERATIONAL
BX DEVICE IS BOXED SN SUBCHANNEL NOT AVAILABLE
DN DEVICE NOT AVAILABLE PE SUBCHANNEL IN PERMANENT ERROR
AL DEVICE IS AN ALIAS UL DEVICE IS AN UNBOUND ALIAS

Chapter 11. Managing and monitoring the ESS
193
DISPLAY M=DEV
The
DISPLAY M=DEV
command displays the number of online channel paths to devices (see
Figure 11-11).
Figure 11-11 D M=DEV - Base device
The
PAV BASE AND ALIASES
field shows the number of exposures for an active PAV base
device. The number of exposures is obtained by adding the base device plus all aliases that
are currently bound to the base device.
When issued for an alias device, the command shows the base device the alias is currently
bound to (see Figure 11-12)
Figure 11-12 D M=DEV - Alias device
11.6 Messages and codes
This section documents new and changed z/OS messages and codes in support of ESS. This
is not an all inclusive list of ESS related messages.
11.6.1 IEA434I
IEA434I DEVICE ONLINE IS NOT ALLOWED, GREATER THAN 10017 CYL
Explanation
:
During
VARY ONLINE
command processing, the device service exit determined that the
device is not allowed to become online for the reason specified.
This message is displayed if you try to vary online a device with more than 10017 cylinders on
a z/OS system without Large Volume Support (LVS) coexistence support installed.
DISPLAY M=DEV(C100)
IEE174I 11.30.43 DISPLAY M 704
DEVICE C100 STATUS=ONLINE
CHP C0 C1 C2
DEST LINK ADDRESS 0D 0D 0D
DEST LOGICAL ADDRESS 01 01 01
PATH ONLINE Y Y Y
CHP PHYSICALLY ONLINE Y Y Y
PATH OPERATIONAL Y Y Y
ND = 002105. .IBM.75.000000012345
DEVICE NED = 2105. .IBM.75.000000012345
PAV BASE AND ALIASES 4
D M=DEV(D2FF)
IEE174I 21.02.20 DISPLAY M 504
DEVICE D2FF STATUS=ALIAS OF BASE D205

194

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Also, devices with more than 32760 configured cylinders will get the message:
IEA434I DEVICE ONLINE IS NOT ALLOWED, GREATER THAN 32760 CYL
11.6.2 IEA435I
The existing message IEA435I generated by DFSMS Device Support is updated to identify
the following error condition:
IEA435I PHYSICAL DEVICE INCONSISTENT WITH LOGICAL DEFINITION,
[PHY=
devtype
LOG=
devtype]
Explanation:

Mismatches between the hardware configuration and device definition in HCD exist.
Problem Determination:

During the processing of VARY,ONLINE command, system services determined that the
physical device type was not compatible with the logical device type defined in the I/O
configuration. The physical (PHY) and logical (LOG) device type are provided.
System Programmer Response:
Correct the HCD device definitions to match the physical device type attached to the
system. Issuing the DSERV QDASD command in this situation will provide information
from the hardware which can be cross checked with the HCD definitions.
11.6.3 IEA480E SIM message
The IEA480E Service Information Message (SIM) console message is a summary message
prepared and sent to a host asynchronous to any given error event or events. It reflects the
result of error event collection and analysis in the storage control subsystem and indicates
that some kind of action needs to be taken, that is, a service action point has been reached.
The SIM message contains a description of the problem the subsystem has identified (the
impact of failure) and a description of what resources the repair action affects (the impact of
repair). For media SIMs, a description of what action should be taken to resolve the problem
is given as well.
The SIM message is not new to the ESS. It is used with other products such as the 3990 or
the Magstar 3590 tape subsystem, but its role as a service notification facility for the ESS has
changed, as documented in the
IBM TotalStorage Enterprise Storage Server Service Guide
2105 Models E10, E20, F10 and F20, and Expansion Enclosure, Volume 1,
SY27-7605-08:
“SIM generation by the ESS family of products is not intended to be the primary
notification for service, as it was for the 3390, 3990, 9340, and 9390 product families. SIM
generation for ESS is a complement to the existing problem notification process, and is
used to support previous system attachments to S/390 hosts. The strategy for SIM
presentation differs from previous products. Instead of directing a SIM to the failing device
and system, hardware SIMs will be presented to all S/390 hosts attached to the storage
subsystem. Exception Class 0 and Media SIMs will still be off-loaded against the failing
device and system.”

Chapter 11. Managing and monitoring the ESS
195
IEA480E
yyyy
,
{SCU | CACHE | DASD | MEDIA }, {SERVICE | MODERATE | SERIOUS |
ACUTE} ALERT, MT=
machine type/model
, SER=MMPP-SSSSS, REFCODE=

nnnn-nnnn-nnnn
, VOLSER=
volser
, ID=
id
, CCHH=x
'cccc hhhh'
,
REPEATED
[CONTINUATION OF IEA480E SNS=
sense
]
Explanation
:
IEA480E is a Service Information Message (SIM) console message. The specified device
or storage control has detected an abnormal condition that requires operator or service
attention.
The subsystem component affected by the failure is categorized as
SCU
(storage control
hardware),
CACHE
(storage control cache or nonvolatile storage),
DASD
(storage device
hardware), or
MEDIA
(device data storage media).
The severity level of the SIM event being reported is classified as
SERVICE, MODERATE,
SERIOUS,
or
ACUTE
. ACUTE is the most severe and SERVICE is the least severe.
You can select the severity level of SIM messages being reported. The IBM service
representative will set the reporting level at ESS installation time based on information you
provide on the Communication Resources work sheet (option
Service information
messages (SIMs) for S/390
). See 8.3.2, “Configuration worksheets” on page 116.
Figure 11-13 shows a SIM message that was issued on the system console to notify the
operator of a DDM failure on a RAID array.
Figure 11-13 Service Information Message (SIM)
11.6.4 IEE169I
IEE169I VARY REJECTED, PATH
(
dddd,xx
)
OFFLINE DUE TO C.U.I.R.
Explanation:

A request was made to place online a path to a device through a channel path. However,
the path is being kept offline by a system service.
This message is issued if you try to vary online a channel path that has been taken offline
by CUIR.
11.6.5 IOS090I
IOS090I
{
xx,rrrr
.
| SETIOS}

dev

IS AN INVALID DEVICE
Explanation
:
An error occurred when the operator entered the
SET IOS=xx
or
SETIOS MIH,..
command.
This message is issued if the IECIOSxx PARMLIB member sets a MIH value to a PAV
alias device, or the operator enters the
SETIOS MIH,...
command to set a MIH value to a
PAV alias device.
*IEA480E A70D,DASD,SERIOUS ALERT,MT=2105,
SER=0175-12345,REFCODE=E100-0191-0909,VOLSER=/UNKN/,ID=09
CONTINUATION OF IEA480E SNS=00101B4000358F09
8F80400191090904E50049C95201E10005100200FE000000

196

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
11.6.6 CUIR messages
Refer to 11.2.2, “Control Unit Initiated Reconfiguration (CUIR)” on page 178 for a summary of
CUIR initiated messages.
11.7 IBM StorWatch Enterprise Storage Server Expert
The IBM StorWatch Enterprise Storage Server Expert (ESS Expert) is an optional software
product which provides storage resource monitoring and management functions for the ESS
and IBM tape libraries. The ESS Expert provides asset management, capacity management,
and performance management capabilities.
The performance reporting function for the ESS provides cluster, device adapter, disk groups,
and logical volume level statistics. The information includes number of I/O requests, number
of bytes transferred, read and write response times, and cache use statistics reported on an
hourly basis. The reports can be viewed in tabular and graphical format. With this information,
you can make informed decisions about volume placement and capacity planning, as well as
isolate I/O performance bottlenecks.
The ESS Expert is described in Chapter 6, “IBM StorWatch Enterprise Storage Server
Expert” on page 87. For detailed information on how to use the ESS Expert, you can consult
the IBM redbook
IBM StorWatch Enterprise Storage Server Expert Hands-on Usage Guide,
SG24-6102.

© Copyright IBM Corp. 2002
197
Chapter 12.
ESS Copy Services for S/390
This chapter overviews the ESS Copy Services available for the S/390 environments. The
copy services functions include Peer-to-peer Remote Copy, FlashCopy, Extended Remote
Copy, and Concurrent Copy.
12

198

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
12.1 S/390 Copy Services: introduction
ESS Copy Services for S/390 is composed of four functions which serve as tools for disaster
recovery, data migration and data duplication. These functions are the following:

Peer-to-Peer Remote Copy (PPRC)

FlashCopy

Extended Remote Copy (XRC)

Concurrent Copy (CC)
Extended Remote Copy and Concurrent Copy rely on a high speed data engine, the System
Data Mover (SDM). Peer-to-Peer Remote Copy and FlashCopy are functions of the ESS
which do not rely on the SDM for data movement.
An entire IBM Redbook,
Implementing ESS Copy Services for S/390
, SG24-5680, is devoted
to topics related to implementing these four functions. In this chapter, we do not attempt to go
into that same level of detail regarding ESS Copy Services. Rather, we simply remind you
about the various planning considerations involved.
12.1.1 Peer-to-Peer Remote Copy (PPRC)
PPRC is designed to maintain a current copy of your application data at a remote site. PPRC
is a hardware-based disaster recover solution. PPRC is a feature of the ESS which maintains
copies of the data using a synchronous method.
PPRC is a hardware solution that enables the shadowing of application system data from one
site and its associated volumes to a second system at another site. Updates made on the
primary volumes are synchronously shadowed to the secondary volumes. The update is only
completed when the data being updated is secured on the secondary site.
12.1.2 FlashCopy
FlashCopy is a function of the ESS which provides a point-in-time copy of the data for
application usage such as backup and recovery operations. FlashCopy enables you to copy
or dump the data while applications are updating the data. Both the source and target
volumes must reside on the same LSS.
DFSMSdss automatically invokes FlashCopy when you issue the COPY Full command to a
subsystem that supports FlashCopy functions.
12.1.3 Extended Remote Copy (XRC)
XRC is the asynchronous remote copy solution available for the ESS. It is a combination of
hardware and software which offers the highest levels of data integrity and data availability in
a disaster recover, workload movement and disk migration environment.
The DFSMSdfp component called the System Data Mover (SDM) will copy the writes issued
to the primary volumes to the secondary devices. The SDM in conjunction with the XRC
support on the ESS are required for an XRC implementation.

Chapter 12. ESS Copy Services for S/390
199
12.1.4 Concurrent Copy (CC)
Concurrent Copy (CC) is both an extended function of the ESS and a component of
DFSMSdss. CC enables you to copy or dump data while applications are updating the data.
Both IMS/ESA and DB2 can use CC for their backups. CC delivers a copy of the data, in a
consistent form, as it existed before the updates took place.
Similar to FlashCopy, CC creates a T0 copy of the source. CC will allow you to make copies at
the data set level.
12.2 ESS Copy Services: considerations
Before you can use the PPRC or FlashCopy functions, you must have the appropriate feature
codes installed and enabled on the ESS.
Here are some things to keep in mind as you plan for ESS Copy Services on your ESS:

Only one FlashCopy at a time can be active on a volume, however, you can perform a
PPRC concurrently with FlashCopy on the same volume.

The primary and secondary volumes must reside within an ESS, for both PPRC and
FlashCopy. For PPRC, this would involve copying from one ESS to another ESS; for
FlashCopy, the copies must be within the same LSS of the ESS.

If you intend to manage PPRC using the ESS Copy Services Web interface, then
Ethernet, TCP/IP connectivity is needed between the two participating ESS subsystems
(primary and secondary) and the Web browser initiating and managing the PPRC
activities.

When using ESS Copy Services, you are required to nominate via the CONFIGURATION
panel, one ESS as the Copy Services primary server which is the central place for
collection of all information. A second ESS can be nominated as the backup server. You
will need to carefully plan which ESS you would nominate in which role to ensure
maximum availability.

There is a maximum of 2000 volumes allowed for a ESS Copy Services server. This
number includes all the primary and secondary PPRC volumes plus all the source and
target FlashCopy volumes. (Note: for z/OS when using TSO to control PPRC and
FlashCopy, then there is no 2000 volumes limitation).

The source and target volumes must have the same track capacities and the same
number of tracks per cylinder. The target volume must have the same volume capacity or
larger than the source volume.

For an XRC implementation the SDM (System Data Mover) placement must be decided
with all considerations taken into account. Locating the SDM at the secondary site has
good performance implications.
You will also need to check that the CRIT=YES - Light or Heavy options have been set
correctly in the ESS VPD by your IBM SSR.
You will need to have the IBM SSR check that all the ESS cluster hostnames were added to
the hostname list during installation. Otherwise the cluster hostname will not be found when
defining the ESS Copy Services primary.

200

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
12.3 PPRC
This section describes the configuration rules and invocation processes for PPRC for S/390
12.3.1 Rules for configuring PPRC links
The following are the PPRC configuration rules for establishing ESCON connections between
two ESS subsystems.

Up to eight ESCON links are supported between two ESS subsystems.

A primary ESS can be connected via ESCON links to up to four secondary ESS
subsystems.

A secondary ESS subsystem can be connected to any number of primary ESS
subsystems, limited by the number of ESCON links available.

PPRC links are unidirectional, because the ESCON port at the primary ESS is
reconfigured to act like an ESCON channel in a host S/390 processor. The primary
ESCON port is dedicated to PPRC, and cannot connect to an S/390 host while PPRC
paths are established.

If an ESCON port on the secondary ESS is connected to an ESCON director, it can
connect to either a PPRC primary ESS subsystem or a zSeries host. It does not need to
be dedicated to PPRC (Note: if the direction of the link is swapped, then the port will
become unavailable to the host).

The ESCON protocol has been streamlined with less handshaking and larger frames
transmitted between ESS. This has enabled the PPRC ESCON links to be extended to up
to 103 km.
An ESCON PPRC link can be used only to transmit data from the primary storage control to
the secondary. If you want primary and secondary volumes on each of two ESS, you need
ESCON PPRC links in each direction. The number of links needed in each direction depends
on the total write activity to all the primary devices in each ESS.
12.3.2 How to invoke PPRC for S/390
In this section we describe how to invoke PPRC for S/390.
ESS Copy Services Web interface
To set up the paths and pairs, you need to know the last five digits of the serial number and
logical subsystem number of the primary and secondary resources. These details can be
found using the ESS Copy Services Web interface or the OS/390 DEVSERV QDASD
command.
To invoke PPRC you need to perform the following tasks from the ESS Copy Services main
menu:
1.Using the PATHS panel, establish a logical path between the primary ESS LCU and host
adapter, and the target LCU and its host adapter.
2.Use the VOLUMES panel to find and select the source and then target volume PPRC
pairs and Establish, Suspend or Terminate the data transfer. You can optionally establish
both paths and volume pairs from this panel.
3.Alternatively, you can select the CONTROL UNIT panel to initiate or remove PPRC
relationships between all volumes on an LCU.
4.You can save previously defined PPRC path and pair definitions as tasks. Using the
TASKS panel, you can select and run the pre-saved set of tasks.

Chapter 12. ESS Copy Services for S/390
201
During steps 1, 2 and 3, once you have selected the resources that you are working with, a
right mouse click on the target resource starts a wizard to guide you through the selection of
the appropriate PPRC functions.
To assist you to remove unrequired resource information from the panels, a FILTER button is
provided to enable display of selected resources. For example, you could show only:

S/390 or open systems volumes

Source or target volumes

Physical or logical ESS
Each panel also has an INFORMATION button which will display PPRC status and other
general information about the selected LCUs, volumes or paths.
Detailed information regarding each panel mentioned above can be found in
IBM
TotalStorage Enterprise Storge Server Web Interface User’s Guide,
SC26-7346, the Web
Interface Users Guide, SC26-7346, and
Implementing ESS Copy Services on S/390,
SG24-5680.
In z/OS and OS/390 environments, PPRC can also be invoked using TSO commands or the
ICKDSF utility. Also, in these environments, the ANTRQST macro supports PPRC requests.
TSO commands
Specific TSO/E commands can be used in the z/OS and OS/390 environments to control
PPRC. These commands are extremely powerful, so it is important that they are used
correctly and directed to the correct devices. It is recommended that you place the TSO
commands in a RACF-protected library to restrict PPRC TSO commands to authorized
storage administrators only.
ICKDSF
ICKDSF supports PPRC operations in the z/OS, z/VM, VSE/ESA and stand-alone
environments. The PPRC functions are supported through the PPRCOPY command.
ICKDSF also provides a subset of the z/OS PPRC commands for stand-alone use when no
operating system is available.
12.4 FlashCopy
This section describes the configuration rules and invocation processes for FlashCopy. How
to invoke FlashCopy. These are the ways that FlashCopy can be invoked in the z/OS or
OS/390 environment:

DFSMSdss can be used to invoke FlashCopy in batch via the
ADDRSSU
program, with the
Copy Full
utility function.

TSO/E commands that are unique to FlashCopy. These commands are:
FCESTABL
,
FCQUERY
, and
FCWITHDR
.

DFSMSdfp Advanced Services ANTRQST macro which calls the System Data Mover API.

The ESS Copy Services Web interface presents you the panels needed to invoke
FlashCopy.
To invoke FlashCopy with ESS Copy Services panels, use the VOLUMES panel to find and
select the source and target volumes. Once presented with the wizard window, select
Establish
or
Withdraw
the FlashCopy pairs.

202

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
As with PPRC, you can save your FlashCopy definitions as tasks and run them at any time by
using the TASKS panel. You can also FILTER your displays and use the INFORMATION
button.
Once a relationship has been established between a FlashCopy source and target volume, a
background task commences that copies the entire source volume to the target. If initiating
the FlashCopy using the ESS Copy Services Web interface, you can suppress this copy task
by specifying a NOCOPY option.
If NOCOPY has been specified, any data about to be updated on the source volume is first
copied to the target volume. Hence, the target volume only contains pre-updated data, not a
complete volume copy. The T0 copy of the source is still available for use as long as the
source target relationship exists.
This relationship can be terminated (WITHDRAW) using the ESS Copy Services Web
interface. If NOCOPY was not specified, the relationship ends automatically once the
background source copy has been completed.
If you have the appropriate DFSMSdss PTF, you may also use the FCWITHDRAW for full
volume DUMP operations. If the source volume of the full volume dump operation is the
source of a FlashCopy relationship, specifying FCWITHDRAW will cause DFSMSdss to
withdraw the FlashCopy relationship when the dump has completed successfully.

© Copyright IBM Corp. 2002
203
Part 3
Implementation in
the open systems
environment
In this part we discuss the processes for implementing the IBM TotalStorage Enterprise
Storage Server in open systems environments. We first cover software requirements for the
different open platforms. Then we examine in length how to design and implement an ESS
logical configuration that meets your needs, with specific implementation guidelines for the
different servers. Finally, we discuss monitoring and managing, the migration of data to the
ESS, and considerations for implementing copy services.
The open systems environment is very broad and expands across many different types of
platforms. The guidelines presented in this part of the book must be complemented with the
detailed technical bibliography of the specific platforms you will be implementing.
Part 3

204

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment

© Copyright IBM Corp. 2002
205
Chapter 13.
Open systems support
This chapter guides you through the process of verifying that the hardware and software
requirements of the host server are adequate to support the attachment of the IBM
TotalStorage Enterprise Storage Server (ESS).
The guidelines discussed in this chapter are for configurations where the host system is an
open systems server that is to be attached to the ESS using either SCSI connections or Fibre
Channel connections.
13

206

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
13.1 General considerations
Figure 13-1 shows the multiplatform attachment characteristics of the IBM TotalStorage
Enterprise Storage Server (ESS).
Figure 13-1 ESS - open systems support
When starting to plan for the host systems that are going to be attached to the ESS, you
should commence at the following Web site:
http://www.storage.ibm.com/hardsoft/products/ess/supserver_summary_open.html
Here you will find the most current list of open servers that can be attached to the ESS. In
Table 13-1 we present the information available at the time of writing our redbook.
IBM S/390
eServer zSeries 900
AS/400
Other
Intel-based
Windows
NT Servers
Sun
Enterprise
Storage Server
SerialDisk and
IBM 7133 Serial Disk
S/390: Count-Key-Data
Remote mnagement with
IBM TotalStorage ESS Specialist
IBM TotalStorage ESS Expert
HP 9000
IBM
RS/6000
and
pSeries
Network
Other platforms:
Fixed Block
DG
Linux
IBM PC
Server
Netfinity

Chapter 13. Open systems support
207
Table 13-1 Open systems support summary
13.2 Operating system requirements
During the preinstallation planning you should go through the following considerations in
respect to the operating system (O/S) running in the servers you plan to attach to the ESS:

Is the host O/S at the minimum software level required to support the ESS?

Has all the recommended software maintenance and fixes for the O/S been installed and
tested?
The information needed to check the preceding considerations about the O/S requirements,
is at the following Web site:
http://www.storage.ibm.com/hardsoft/products/ess/supserver.htm
Once there, you click, or download, the PDF at
ESS Supported Servers
(see Section 1.5.1,
“ESS Web site” on page 9).
This document is updated with the latest support data. You will find in this document, detailed
information on the O/S requirements for each open systems platform connectable to the ESS,
for both SCSI and Fibre Channel attachment.
Platforms support O/S Levels
support
SDD
support
CLI
support
Load
balancing
SCSI
channel
support
Fibre
channel
support
pSeries AIX 4.2.1 - 5L
version 1
Yes Yes Yes Yes Yes
Compaq Open VMS No No No Yes No
Tru 64 4.x & 5.x Yes Yes No Yes Yes
Hewlett Packard HP/UX 10.20 -
11.00
Yes Yes Yes Yes Yes
iSeries AS/400 V4 release
4.0
No No No Yes Yes
Intel PC based servers Windows NT Yes Yes Yes Yes Yes
Windows 2000 Yes Yes No Yes Yes
Linux Redhat 7.1 No Yes No Yes Yes
SuSE 7.2 No Yes No Yes Yes
Novell NetWare 4.2 -5.1 Yes Yes No Yes Yes
Data General DG/UX 4.2 No No No Yes No
Sun Solaris Yes Yes Yes Yes Yes
xSeries/NUMA-Q DYNIX/ptx No Yes No No No
SGI RPQ RPQ RPQ No No RPQ

208

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
13.3 I/O adapters: considerations
During the preinstallation planning you will also go through the following considerations in
respect to the input/output (I/O) adapters your servers will use to attach to the ESS:

What type of connectivity will be used? SCSI and/or Fibre Channel?

What SCSI or Fibre Channel I/O adapters does your server support, to connect to the
ESS?
The information needed to check the preceding considerations about the host system I/O
adapter requirements, is at the following Web site:
http://www.storage.ibm.com/hardsoft/products/ess/supserver.htm
Once there, you click, or download, the PDF at
ESS Supported Servers
(see Section 1.5.1,
“ESS Web site” on page 9).
This document is updated with the latest support data. You will find in this document detailed
information on the I/O adapters that are supported for each host system server, to attach to
the ESS. The information is provided for both SCSI and Fibre Channel attachment.
13.4 Subsystem Device Driver (SDD)
The Subsystem Device Driver (SDD) program, that runs in the host system, provides path
failover/failback for increased availability, as well as path load balancing for improved I/O
performance.
If you plan to use the SDD program you will have to check how it is supported in your planned
configuration. This information is at the URL:
http://www.storage.ibm.com/hardsoft/products/ess/supserver.htm
Once there, you click, or download, the PDF at
ESS Supported Servers
(see Section 1.5.1,
“ESS Web site” on page 9).
This PDF contains a section that documents the SDD support in a matrix format, both for
clustered and for non-clustered configurations.
Also, from this same URL, you can click over the entry
SDD (Subsystem Device Driver)
.
This will take you to:
http://ssddom01.storage.ibm.com/techsup/swtechsup.nsf/support/sddupdates
This last Web page has the latest support information and downloads for SDD installation.
At this Web page you can access the
IBM Subsystem Device Driver User’s Guide
as PDF or
HTML.

Chapter 13. Open systems support
209
13.5 Command Line Interface (CLI)
The ESS Copy Services Web interface is available for all supported platforms. The ESS Copy
Services Web interface function is a Java/CORBA based application that runs on the host
system and requires a TCP/IP connection to each ESS under it. ESS Copy Services also
includes a Command Line Interface feature (CLI) for communicating with the ESS Copy
Services server from the host’s command line. The CLI code is supplied with the ESS
microcode and is specific to the code level.
If you are planning to use the CLI facility for the ESS Copy Services functions of the ESS,
then you should check the O/S environments that do support the CLI interface. This
information is at:
http://www.storage.ibm.com/hardsoft/products/ess/supserver.htm
Once there, you click, or download, the PDF at
ESS Supported Servers
(see Section 1.5.1,
“ESS Web site” on page 9).
This PDF contains a section on ESS Copy Services that documents the host O/S
requirements for supporting the Command Line Interface.
13.6 Additional considerations
When reviewing the software requirements, you may also have to consider the following:

Have the Independent Software Vendors (ISV) been contacted regarding their support of
the ESS?

Have any required ISV fixes and prerequisites been installed and tested?

Are there any considerations for volume manager software (Veritas VxVm, HP SAM, or
Sun Solaris DiskSuite)? These are software products that provide disk administration tools
to allow you to configure and manage the ESS defined devices, once presented to the
host for some open systems platforms.

210

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment

© Copyright IBM Corp. 2002
211
Chapter 14.
Open systems host setup tasks
The IBM TotalStorage Enterprise Storage Server is designed to handle open systems hosts
such as the IBM pSeries and RS/6000 with AIX, IBM iSeries and AS/400 with OS/400,
xSeries Servers with Windows NT, Windows 2000, Linux, Novell Netware, NUMA-Q, and
many other UNIX based systems. All of these systems use SCSI and/or Fibre Channel
generic devices, or specific disk emulations, to address the ESS disk storage.
This chapter discusses the setup tasks you need to perform and the parameters you need to
know to successfully connect the ESS to your open systems host, as well as to access the
disk devices defined during the logical configuration of the ESS. We discuss how to get these
parameters and the commands you will be using for the various open platforms.
14

212

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
14.1 General
The parameters discussed in this chapter are necessary for you to enter in as values,
required to complete the ESS open system host definitions and other ESS internal settings.
We recommend that before you commence to set up the logical configuration of the ESS, that
you gather this information together, so that you have it available to ensure smooth logical
configuration of the ESS to your open system platform(s).
You should obtain the following general information for each host platform you will be
connecting.

The number of host SCSI or Fibre Channel I/O adapters for each host.

The WWPN of each Fibre Channel I/O adapter card used to connect the host to the ESS.
This is necessary to complete the ESS logical setup planning.

How to get the host to recognize the disks, assigned to each SCSI or Fibre Channel I/O
adapter.

How to list the disks on each host and identify the ESS LUN serial number on each host
disk.
You will want to know how to get the host to recognize the ESS LUNs, to ensure that your
definitions match the ESS internal addressing for the target ID and LUNs. These values are
assigned by you using the ESS Specialist, during the logical configuration of the ESS, and
can be viewed in the ESS Specialist
Storage Allocation
panel (Tabular view).
The information presented in this chapter can be complemented with the following:

To work with the IBM TotalStorage Enterprise Storage Server Specialist (ESS Specialist),
refer to the IBM publication
IBM TotalStorage Enterprise Storge Server Web Interface
User’s Guide,
SC26-7346.

As an additional source of information when implementing FCP on AIX, Sun, HP and
Windows NT, refer to the IBM redbook
Implementing Fibre Channel Attachment on the
ESS,
SG24-6113

For the most current and up to date information on the number of supported LUNs per
target ID, for the different host servers, please refer to Chapter 1 of the
Host Systems
Attachment Guide,
SC26-7296
.
See Section 1.5.3, “ESS documentation Web site” on page 10 and Section 1.5.4, “Redbooks”
on page 12 for information on how to get ESS related documentation.
As an initial recommendation you should consider not adding more than 120 GB per SCSI I/O
adapter, and not more than 300GB per Fibre Channel I/O adapter. This recommendation is
because of the SCSI and Fibre Channel attachment characteristics, and not because any
limitation on the ESS.
14.2 AIX
The RS/6000 with AIX supports up to 32 LUNs per target ID for the Ultra wide SCSI adapters,
and a maximum of eight LUNs with SCSI-2 fast wide adapters. AIX 4.3.3 and 5.1 support up
to 64 LUNs per target ID. The number of target IDs per adapter is not a consideration when
you are attaching the ESS to the host with Fibre Channel adapters.

Chapter 14. Open systems host setup tasks
213
The AIX operating system contains entries in its Object Distribution Manager (ODM),
database to identify the ESS. However, the AIX operating system accesses the ESS through
its generic SCSI or Fibre Channel Data Distribution Manager (DDM) and, therefore, sees the
ESS LUNs as an hdisk.
Once the ESS has been connected to your processor, run the
2105inst
script, provided on a
CD-ROM with the ESS. On successful completion of the script, you can run the configuration
manager to detect the devices and add the hdisks with type IBM 2105B01.
14.2.1 I/O adapters in the host server
To obtain the number the SCSI or Fibre Channel of I/O adapters in your server, enter the
following commands:
lsdev -Cc adapter | grep scsi
for SCSI
lsdev -Cc adapter | grep fcs
for Fibre Channel
The example output for SCSI looks like the following:
scsi0 Available 10-60 Wide/Fast-20 SCSI I/O Controller
scsi1 Available 10-68 Wide/Fast-20 SCSI I/O Controller
The example output for Fibre Channel looks like the following:
fcs0 Available 14-08 FC Adapter
fcs1 Available 2A-08 FC Adapter
14.2.2 WWPN of the Fibre Channel I/O adapters
To get the WWPN of the Fibre Channel I/O adapters, enter the following command:
lscfg -vl fcs0 | grep Net
The output looks like this
Network Address.............10000000C9280712
14.2.3 Host recognition of newly assigned disks
To get the host server to recognize the newly assigned disks run configuration manager or
reboot the host:

To bring in the disks connected to this Fibre Channel type:
cfgmgr -l fcs0

To bring in the disks assigned to all adapters, type:
cfgmgr
14.2.4 ESS LUNs assigned as host hdisks
To list how the ESS LUNs have been assigned as hdisks to the host server, enter the
following commands:

To list all the hdisks, type:
lspv
The output looks like this:
hdisk8 000b7fad1ffac5ff none

To list all the hdisks assigned to Fibre Channel adapter fcs0, type:
lsdev -Cc disk | grep 14-08

214

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
The output looks like this:
hdisk8 Available 14-08-01 IBM FC 2105F20

To show the ESS LUNs serial number corresponding to the hdisk, type:
lscfg -vl hdisk8 | grep Serial
The output looks like this:
Serial Number...............40D15630
Where 40D is the LUN ID and 15630 is the ESS serial number.
If you have implemented and configured IBM Subsystem Device Driver (SDD) then enter the
following command to obtain the LUN to disk serial number.
lsvpcfg vpath
# where # is the vpath number.
If two adapters are configured on the host then the output looks like this:
vpath0 (Avail pv data03MQ1vg) 40D15630 = hdisk8 (Avail ) hdisk33 (Avail )
The host sees hdisk8 and hdisk33 as two hdisks, but they are the same LUN connected to
two paths. The definition of vpath is used to manage the ESS LUN on the host. This way the
host sees the vpath as the physical device and associates the two hdisk definitions to it.
If you connect the ESS to your processor and reboot your AIX system before you have run the
2105inst
script, the ESS devices will be detected and listed as hdisk type of Other devices.
You will need to delete these disks and then run
2105inst
.
Type in the following command to delete the device definition:
rmdev -dl hdisk#
or
rmdev -dl vpath#
where # is the number of the hdisk or vpath.
Once the hdisks are available, you can assign the hdisks to a volume group and put file
systems on the ESS LUNs.
14.2.5 FlashCopy
AIX marks disks with a Physical Volume ID (PVID). FlashCopy creates an identical copy of a
volume, including the PVID. You will need to change the PVID of the target or copy volume if
you want to use this volume from the same host.
This can be achieved by first clearing the PVID and then assigning a new unique one. For
example:
chdev -l disk -a pv=clear
chdev -l disk -a pv=yes
For detailed procedures on implementing FlashCopy on an AIX hosts, please refer to the IBM
redbook
Implementing ESS Copy Services on UNIX and Windows NT/2000,
SG24-5757
.
14.3 Compaq (DEC)
The Compaq Open VMS and Tru64 4.x supports up to eight LUNs per target ID, 0 to 7, for
SCSI adapters, and Compaq Tru64 5.x supports sixteen LUNs per target ID, 0-15, for SCSI
adapters. The number of target IDs per adapter is not a consideration when you are attaching
the ESS to the host with Fibre Channel.

Chapter 14. Open systems host setup tasks
215
14.3.1 I/O adapters in the host server
Use the following procedures to install and configure the ESS disk drives on a Compaq Tru64
4.x and 5.x systems:
1.Push the halt button to turn on the Alpha server
At the

>>>

prompt type in the following command:
show device
The system responds with a list of controllers and disks that are connected to the host. In
the description field on the right of the screen you can check for disk devices that begin
with the letters
dk
.
2.Type
boot
to restart the system.
3.Once the system has restarted, open two terminal windows and type in each:
uerf -R 300 | more

The output should show devices that start with
rz
in each window. For example:
rz28, rz29, rzb28,
and
rzb29
Compare the lists to determine which ESS devices you want to add to the host system.
14.3.2 WWPN of the Fibre Channel I/O adapters
To determine the WWPN of the Fibre Channel I/O adapters in your host server, type in the
following command:
PO>>>wwidmgr -show
Here is an example output with
10000000c922d469
as the WWPN:
[0] pga0.0.0.7.1 1000-0000-c922-d469 FABRIC FABRIC
Or type in the following command and look for the
wwn
number.
view /var/adm/messages
14.3.3 Host recognition of newly assigned disks
You must reboot.
If you install the devices after the initial operating system installation, you must make special
files that create the character devices needed for the file systems.
For Open VMS
1.Type
cd /dev
2.Type
./MAKEDEV rzxx
, where
xx
is the number portion of the device name. Repeat this for
each new drive that you installed.
Operating system device recognition for Tru64 4.x
Use the following procedures to install and configure the IBM ESS disk drive modules on a
Compaq Tru64 UNIX Version 4.0x host system.

216

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Console device check
The following procedures tell you how to perform a console device check.
1.Push the Halt button to turn on the AlphaServer.
2.Type
show device
at the
>>>
prompt to list the devices available to the AlphaServer.
The system responds with a list of controllers and disks that are connected to the system.
In the description field on the right of the screen, you should see a list of all devices
assigned by the ESS. Disk devices begin with the letters
dk
. If you do not see a list of
devices, verify the SCSI connections, connectors, and terminators on the bus. If you still
do not see a list of devices, check the ESS to ensure that the ESS is operating correctly.
3.When the list of the devices is displayed on your screen, type
Boo
t to restart the system.
The system performs self-test diagnostics and responds with the console prompt >>>.
Operating system device recognition
After the system has restarted, perform the following steps to verify that UNIX recognizes the
disks:
1.Open two new terminal windows.
2.Type
uerf -R 300|more
on the command line in each of the windows.
A list of device names that begin with the letters
rz
is displayed in each window.
For example, device names should look like the following:
rz28,rz29,rzb28,and rzb29
3.Compare the lists to determine which ESS devices you want to add to the system.
Device special files
If you install a serial storage architecture (SSA) device after the initial operating system
installation, you must make the device special files that create the character devices needed
for file systems.Perform the following steps:
1.Type:
#cd /dev
2.Type:
#./MAKEDEV rz xx
where xx is the number portion of the device name.
3.For each new drive that you installed in the SSA device type
#./MAKEDEV rz xx
where xx
is the number portion of the device name.
Initializing disk device
After the list of devices has been determined, you must label the disk volume sizes. Perform
the following steps to label the disks:
1.Write the new label by typing:
#disklabel –rw rz28 ESS
2.Verify the label by typing:
#disklabel rz28
The
#disklabel rz28
command shows the new partition layouts on the Compaq Tru64
and automatically detects the LUNs that are provided by the SSA device.
Configuring AdvFS
Before you create an AdvFS file system, you must design a structure by assigning a file
domain and file sets. Perform the following steps to create an AdvFS filesystem with file sets:
1.Type:
#cd/
2.Type:
#mkfdmn –rw /dev/rzXc vol1_dom
3.Type:
#mkfset vol1_dom vol1
4.Type:
#mkdir /vol1 #mount vol1_dom#vol1 /vol1
5.To display all mounted devices, type:
df -k

Chapter 14. Open systems host setup tasks
217
Configuring devices to mount automatically
To enable an AdvFS file system to start automatically, add an entry to the /etc/fstab file that
the mount command will issue during startup. Figure 14-1 on page 217 shows an example of
a modified /etc/fstab file.
In Figure 14-1 on page 217, the lines that are shown in bold type are the lines that were
entered since the initial operating system installation.
Figure 14-1 Example of a modified /etc/fstab file
Installing and configuring the Compaq Tru64 UNIX Version 5.0x
Use the following procedures to install and configure the ESS disk drives on a Compaq Tru64
UNIX Version 5.0x.
Console device check
These are the steps for doing a console device check:
1.Push the halt button to turn on the AlphaServer.
The system performs self-test diagnostics and responds with the console prompt >>>.
2.Type
show device
at the
>>>
prompt to list the devices available to the AlphaServer.
The system responds with a list of controllers and disks that are connected to the system.
In the description field on the right of the screen, you should see a list of all devices
assigned by the ESS. Disk devices begin with the letters
dk
. If you do not see a list of
devices, verify the SCSI connections, connectors, and terminators on the bus. If you still
do not see a list of devices, check the ESS to ensure that the ESS is operating correctly.
3.When the list of devices is displayed, type
Boot
to restart the system.
Operating system device recognition
To view the disk information for installed devices on the Compaq AlphaServer, type:
#/sbin/hwmgr -view devices
Table 14-1 shows an example of what you see when you type the
#/sbin/hwmgr-view
devices
command. The /dev/disk/dsk* device special files are the actual devices connected
to the Compaq AlphaServer. The default internal disk,
/dev/disk/dsk0a
is the disk the
Compaq AlphaServer uses for starting the operating systems.

218

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Table 14-1 Example of disk information for installed devices for a Compaq AlphaServer
If you install an SSA device after the initial operating system installation, you must make the
device special files that create the character devices needed for filesystems. Perform the
following steps:
1.Type:
#cd /dev
2.Type:
#./MAKEDEV /dev/rdisk/dsk1a
3.Type: #
./MAKEDEV /dev/rdisk/dsk1a
for each new drive that you installed in the SSA
device.
Initializing disk devices
After the list of devices have been determined, you must label the disk volume sizes. Perform
the following steps to label the disk volume sizes:
1.Write the new label by typing:
#disklabel –rw dsk1a ess
2.Verify label by typing:
#disklabel dsk1a
3.The
#disklabel dskla
command shows the new partition layouts on the CompaqTru64
and automatically detects the LUNs that are provided by the SSA device.
Configuring AdvFS
Before you create an AdvFS file system you must design a structure by assigning the file
sets.Perform the following steps to create an AdvFS file system with file sets:
1.Type:
#cd/
2.Type:
#mkfdmn –rw /dev/disk/dsk1a vol1_dom
3.Type:
#mkfset vol1_dom vol1
4.Type:
#mkdir /vol1
5.Type:
#mount vol1_dom#vol1 /vol1
6.To display all mounted devices, type:
df -k
Configuring devices to mount automatically
To enable an AdvFS file system to start automatically, add an entry to the /etc/fstab file that
the mount command will issue during startup. See Figure 14-2 for an example of a modified
/etc/fstab file.
In Figure 14-2, the lines that are shown in bold type are the lines that were entered since the
initial operating system installation.
Hardware ID Device Special File Manufacturing Model Location
3/dev/kevm N/A N/A N/A
28/dev/disk/floppy0c N/A 3.5 inch floppy fdi0-unit-0
30/dev/disk/dsk0c DEC RZ1DF-CB DEC bus-0-targ-0-lun-0
31/dev/disk/cdrom0c DEC RRD47 DEC bus-0-targ-4-lun-0

Chapter 14. Open systems host setup tasks
219
Figure 14-2 Example of how to label disks with volume sizes
14.3.4 ESS LUNs assigned as host disks
You can also identify the ESS LUN serial number on Fibre Channel attachments with the
following procedures:
1.Type
set mode diag
to put the console into diagnostic mode.
2.Type
wwidmgr -show wwid
Figure 14-3 shows an example of the output.
Figure 14-3 Example listing of LUN assignment for Compaq servers
In the example in Figure 14-3,
282
is the LUN ID and
15660
is the serial number of the ESS.
1.To verify attachment of the ESS LUNs on the host system, type in the following command:
hwmgr -view dev -cat disk
2.Here is an example of the output:
63: /dev/disk/dsk5c IBM 2105F20 bus-0-targ-253-lun-0
14.4 Hewlett Packard 9000
HP 9000 systems support up to eight LUNs, 0 to 7 per target ID, 0 to 7, for SCSI adapters.
The number of target IDs per adapter is not a consideration when you are attaching the ESS
to the host with Fibre Channel I/O adapters.
Once the ESS has been connected to your processor, run the
2105inst
script, provided on a
CD-ROM with the ESS. On successful completion of the script, you can run the configuration
manager to detect the devices and add the hdisks with type IBM 2105B01.
[0] UDID: -1 WWID:01000010:6000-1fel-492-4d20-0000-0000-28b1-5660 (ev:none)
[1] UDID: -1 WWID:01000010:6000-1fel-492-4d20-0000-0000-2881-5660 (ev:none)
[2] UDID: -1 WWID:01000010:6000-1fel-492-4d20-0000-0000-2821-5660 (ev:none)
Note:
Ensure that you have at least 1 MB minimum of hard disk space available to install
the 2105inst file.

220

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
14.4.1 I/O adapters in the host server
Once you have identified the host I/O adapters physically connected to the ESS with the
system administrator, you then can type in the following commands for SCSI adapters:
ioscan -kfn | grep SCSI
The output should look something like the example in Figure 14-4.
Figure 14-4 SCSI adapter identification - HP-UX
For Fibre Channel adapters, type in the following command:
ioscan -kfnC fc

Figure 14-5 shows an example output.
Figure 14-5 Fibre Channel adapter identification - HP-UX
14.4.2 WWPN of the Fibre Channel I/O adapters
The tool used to discover the WWPN is:
tdutil /dev/tdx

Here,
x
is the instance number.
For example:
/opt/fcms/bin/tdutil /dev/td0
This should return the WWPN along with all the other information of instance 0.
This is an example of output with
50060b000007a01a
as the WWPN.
N_Port Node World Wide Name = 0x50060b000007a01b
N_Port Port World Wide Name = 0x50060b000007a01a
14.4.3 Host recognition of the newly assigned disks
To see how the host server is recognizing the newly assigned disks, type in the following
command:
ioscan -fnC disk | more

When you are initially bringing new disks/LUNs in there won't be any special device files (ex.
/dev/dsk/c4t4d1, etc.) you must then run the command:
insf -e
ext_bus 0 0/0/1/0 c720 CLAIMED INTERFACE SCSI C896 Fast Wide Single-Ended
ext_bus 1 0/0/1/1 c720 CLAIMED INTERFACE SCSI C896 Ultra Wide Single-Ended
Class I H/W Path Driver S/W State H/W Type Description
=================================================================
fc 0 0/1/0/0 td CLAIMED INTERFACE HP Tachyon TL/TS Fibre Channel Mass Storage Adapter /dev/td0
fc 1 0/1/1/0 td CLAIMED INTERFACE HP Tachyon TL/TS Fibre Channel Mass Storage Adapter /dev/td1

Chapter 14. Open systems host setup tasks
221
Figure 14-6 shows a sample output.
Figure 14-6 Host view of newly assigned disks - HP-UX
14.4.4 ESS LUNs assigned as host disks
You can identify the ESS LUN serial number to disk number with the following command-line
interface command:
rslist2105s
Figure 14-7 shows a sample output.
Figure 14-7 LUN serial number to disk number - HP-UX
Note that 021 is the LUN ID and 17005 is the ESS serial number.
14.5 AS/400 and iSeries
The iSeries and AS/400 support only six target IDs and eight LUNs ranging from 0 to 7. The
AS/400 target address is always six. Fibre Channel attachment in the iSeries supports 32
target IDs, 1-32.
The information presented on this section can be complemented with the IBM redbook
iSeries in Storage Area Network A Guide to Implementing FC Disk and Tape with iSeries,
SG24-6220.
14.5.1 I/O adapters in the host server
Follow this procedure to recognize the I/O adapters in the host server:
1.Enter the command:
WRKHDWRSC *STG
Figure 14-8 shows the sample output, with
2766 Type
resource for the iSeries FC IOA and
6501 Type
resource for SCSI adapters.
disk 0 0/0/1/1.2.0 sdisk CLAIMED DEVICE IBM DMVS18D /dev/dsk/c1t2d0 /dev/rdsk/c1t2d0
disk 1 0/0/2/0.2.0 sdisk CLAIMED DEVICE IBM DMVS18D /dev/dsk/c2t2d0 /dev/rdsk/c2t2d0
d
isk name 2105 serial number
--------- ------------------
c4t4d1 02117005
c5t1d7 10F17005
c6t1d5 20D17005
c8t4d4 02417005

222

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Figure 14-8 Identifying the I/O adapters - iSeries and AS/400
2.You may need resource detail information for later steps.
For this you choose option 7 to display resource detail on the 2766 or 6501. Figure 14-10
shows the output information you will get.
Figure 14-9 Resource detailed information for I/O adapters - iSeries and AS/400
3.Take note of the Frame ID and Card position, and System bus, card and adapter.
Work with Storage Resources
System: RCHAST00
Type options, press Enter.
7=Display resource detail 9=Work with resource
Opt Resource Type-model Status Text
DC01
DC03
DC07
DC04
CMB04
DC05
SI03
CMB03
CMB01
DC32
DC06
2778-001
2763-001
2782-001
2749-001
2842-001
2768-001
6501-001
284B-001
284E-001
9337-5A0
2766-001
Operational
Not detected
Not detected
Operational
Operational
Operational
Operational
Operational
Operational
Operational
Operational
Storage Controller
Storage Controller
Storage Controller
Storage Controller
Combined function IOP
Storage Controller
Storage Controller
Combined function IOP
Combined function IOP
Disk Storage Controller
Storage Controller<<<<<< FC IOA <<<<<<<<
Display Resource Detail
System: RCHAST27
Resource name . . . . . . . : DC06
Text . . . . . . . . . . . . : Storage Controller
Type-model . . . . . . . . . : 2766-001
Serial number . . . . . . . : 10-22041
Part number . . . . . . . . : 0000003N2454
Physical location:
Frame ID 2
Card position C08
Logical address:
PCI bus:
System bus 26
System board 0
System card 16
Storage:
I/O adapter 6

Chapter 14. Open systems host setup tasks
223
14.5.2 WWPN of the Fibre Channel I/O adapters
These are the steps you need to follow:
1.On the screen for the AS/400 Main Menu panel, type
strsst
2.Go to the Sign On panel and enter a valid Service Tools user ID and password.
3.On the
System Service Tools
(SST) panel, type
1
to Start a Service Tool.
4.On the
Start a Service Tool
panel, type
7
to select Hardware Service Manager.
5.On the
Hardware Services
panel, type
1
to select Packaging Hardware Resources.
6.On the
Packaging Hardware Resources
panel, type
9
to select the system expansion unit
or System Unit, for the tower that your 2766 is installed in - see Frame ID above.
7.On the
Packaging Hardware Resources
panel, type
8
to select your 2766 Multiple
Function IOA.
8.On the Logical Resources Associated with the
Packaging Resource
panel, type
5
to
select Multiple Function IOA.
9.On the
Auxiliary Storage Hardware Resource Detail
panel, locate the field name for Port
name (see Figure 14-10). The WWID is listed as
Port worldwide name . . . . . . . . : 10000000C922D228
Figure 14-10 Identifying the WWPN - AS/400 and iSeries
14.5.3 Host recognition of the newly assigned disks
The iSeries host should recognize the newly assigned disks automatically. They will need to
be configured into an Auxiliary Storage Pool (ASP) before use.
14.5.4 ESS LUNs assigned as host disks
Type in the following command and follow the instructions:
WRKHDWRSC
, option 9 for your 2766 will show the 2105 type resources.
Figure 14-11 shows LUNs that are currently assigned along with LUNs that were assigned,
but have been since re-assigned elsewhere, hence are not detected.
Auxiliary Storage Hardware Resource Detail
0000003N2454
Storage IOA
I/O bus
Port worldwide name
6
Type-Model
System card
Operational
10-22041
I/O adapter
2766-001
Resource name
Controller
Serial number
DC06
System bus
16
System board
26
Status
Part number
0
PCI bus
Device
Storage

. . . . . . . . . . . . . . . . :
. . . . . . . . . . . . . . . . . . . . :
. . . . . . . . . . . . . . . . . . . . . :
. . . . . . . . . . . . . . . . . . . . . :
. . . . . . . . . . . . . . . . . . . . . :
. . . . . . . . . . . . . . . . . . . :
10000000C922D228
. . . . . . . . . . . . . . . . . . . . . :
. . . . . . . . . . . . . . . . . . . . . . :
. . . . . . . . . . . . . . . . . . . . . :
. . . . . . . . . . . . . . . . . . . . :
. . . . . . . . . . . . . . . . . . . . . . . . :

. . . . . . . . . . . . . . . . . . . . . . . :
. . . . . . . . . . . . . . . . . . . . . . :
. . . . . . . . . . . . . . . . . . . . . . . :
. . . . . . . . . . . . . . . . . . . . . . . :
. . . . . . . . . . . . . . . . . . . . . . . :
Description

224

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Figure 14-11 Listing of the assigned LUNs - iSeries
14.6 Intel based servers
The Intel based servers with Windows NT 4.0 Windows 2000 support up to eight LUNs per
target.
Before attaching a host with Windows NT and 2000, set the basic input/output system (BIOS)
for the SCSI adapters to “disabled”. This will ensure optimum performance is achieved. See
Chapter 2 in the
IBM TotalStorage Enterprise Storage Server Host Systems Attachment
Guide 2105 Models E10, E20, F10 and F20,
SC26-7296,

for more details
.
14.6.1 I/O adapters in the host server
See the next step to determine how may I/O adapters are on the system.
14.6.2 WWPN of the Fibre Channel I/O adapter
To find the WWPN for a Windows NT and Windows 2000 host systems follow these
procedures depending on the I/O adapter.
5= Work with configuration desciption 7= Display resource detail
Disk Unit
Storage Controller
Disk Unit
Disk Unit
Disk Unit
Disk Unit
Disk Unit
Disk Unit
Disk Unit
Disk Unit
Disk Unit
Disk Unit
Disk Unit
Disk Unit
Disk Unit2105-A05
DD024
DD034
DD025
DD022
DD044
DD042
DD019
DD036
DD021
DD020
DD045
DD035
DD018
DD041
Not detetcted
DC06
Opt Resource Type-model TextStatus
2105-A81
2105-A05
2105-A81
2105-A05
2105-A81
2105-A05
2105-A82
2105-A01
2105-A82
2105-A01
2105-A82
2766-001
2105-A81
Operational
Operational
Not detetcted
Operational
Not detetcted
Operational
Not detetcted
Operational
Not detetcted
Operational
Not detetcted
Operational
Not detetcted
2105-A01 Operational
System: RCHAST27
Work with Storage Controller Resources
Type options, press Enter.

Chapter 14. Open systems host setup tasks
225
Qlogic adapter
Perform the following steps:
1.Restart the server.
2.Press
Alt+Q
to get the FAST!Util menu.
If you have more than one Fibre Channel adapter installed, all the Fibre Channel adapters
are displayed. Scroll down to the adapter you want. Press Enter.
3.From the FAST!Util menu, scroll down and choose
Select Host Adapter
.
4.Scroll up and highlight
Configuration Settings
. Press Enter.
5.From the
Configuration Settings
menu, click
Host Adapter Settings
.
6.Write down the host adapter name, for example:
200000E0800C2D5
.
Note
: The nice feature of QLview is that you do not have to reboot the server. If you don't
install QLview, the reboot option is the only way to get the WWPN. The QLview utility can
be installed on a workstation as well as a server (it must be installed on the Hosts you
intend to query).
7.Click the desktop icon (that is a shortcut to the QLview utility).
8.Update the Machine box - using the Host Name of the Windows/NT or Windows 2000
server.
9.Press Enter or click
Connect
.
A panel is displayed with diagram of Qlogic adapters installed in the host with WWPN
displayed.
Emulex LP8000 adapter
Perform the following steps:
1.Click
Start -> Programs -> Emulex Configuration Tool
2.From the
Emulex Configuration Tool
window in the
Available Adapters
window,
double-click the adapter entry for which you want to display the WWPN information.
14.6.3 Host recognition of newly assigned disks
Perform the following steps:
1.Back up disk configuration on Windows host.
2.Assign LUNs to host using ESS Specialist.
3.Reboot host.
4.Use SDD to verify volume numbers displayed are the same as assigned, and that path is
available to all volumes.
5.Use Disk Administrator (NT) or Disk Management (2000) to create logical volumes (LVs).
6.Verify that the
writing signature
prompt matches the new Disk Volume numbers.
14.6.4 ESS LUNs assigned as host disks
You can use the Subsystem Device Driver (SDD) that comes with the ESS, utility to verify
LUNs assigned to the host system. You can also use the DiskAdmin command to invoke
device discovery.

226

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Perform the following steps:
1.Click
Start
->
Programs
->
Subsystem Device Driver
->
Subsystem Device Driver
Management
...
2.At the command prompt, enter:
datapath query device
3.Figure 14-12 shows a sample output:
Figure 14-12 Listing of the assigned LUNs - Intel based servers
In the sample shown in Figure 14-12:
TYPE
is the ESS Model, in this case 2105F20
SERIAL
is the made up of the Volume Number (vvv) and ESS Serial Number
(sssss) in the format vvvsssss
For example DEV#1 shows the Volume 01A and ESS Serial number 19445
Device Name
is the Windows/NT disk volume displayed in Disk Administrator or the
Windows 2000 disk volume displayed in Disk Management
Path#
initially only path 0 will be displayed after reboot when assigning new
LUNs
Path 1
will be displayed for the “dual” path (if multiple paths are available)
Path 1
displayed after Disk Administrator or Disk Management is used to
define Logical Volumes
14.7 Linux
Intel based servers with Linux support up to 128 LUNs per device. The standard Linux kernel
(up to 2.4.x) uses a major and a minor number address mechanism. Each disk device, in fact,
each partion on a disk device, is represented by a special device file; for example,
/dev/sdal
.
Total Devices : 5
DEV#: 0 DEVICE NAME: Disk2 Part0 TYPE: 2105F20 SERIAL: 01219445
=====================================================================
Path# Adapter/Hard Disk State Mode Select Errors
0 Scsi Port5 Bus0/Disk2 Part0 OPEN NORMAL 1057 0
1 Scsi Port5 Bus0/Disk7 Part0 OPEN NORMAL 1137 0
DEV#: 1 DEVICE NAME: Disk3 Part0 TYPE: 2105F20 SERIAL: 01A19445
=====================================================================
Path# Adapter/Hard Disk State Mode Select Errors
0 Scsi Port5 Bus0/Disk3 Part0 OPEN NORMAL 94 0
1 Scsi Port5 Bus0/Disk8 Part0 OPEN NORMAL 117 0
Note:
The disk number is different on the second path. These volumes will be shown as
follows:
OFFLINE volumes on Windows/NT
They do not display in Disk Management on Windows 2000 (“place holder” is maintained).

Chapter 14. Open systems host setup tasks
227
By default there are a maximum of 16 partitions per disk. Both the major and minor numbers
are 8 bit. There are 8 major numbers reserved for SCSI devices. Note that Fibre Channel
attached devices are handled as SCSI devices as well. The major numbers are 8, 65, 66, 67,
68, 69, 70 and 71 (defined from
SCSI_DISK0_MAJOR to SCSI_DISK7_MAJOR
in
/usr/src/linux/include/linux/major.h
). According to Linus Torvalds there will be no
additional major numbers allocated for disk devices in the future!
Furthermore, there are 256 minor numbers available for each of the 8 major numbers.
As there are a maximum of 16 partions per device, the following formula will provide the
maximum number of devices under Linux:
Number of devices = (number of major#’s) x (number of minor#’s) / (number of partions)
Number of devices = 8 x 256 / 16 = 128
You will need to do one of the following tasks to add the Fibre Channel I/O adapters to your
Linux host system. You have two options for loading the driver:

Compile it into the kernel.

Load it as a module.
14.7.1 Kernel configuration
Use the following procedure to configure the kernel:
1.Edit the top-level
Makefile
and specify the new kernel in
/usr/src/linux
.
Change the “
EXTRAVERSION=
” line according to your new kernel.
Note
: this step is to ensure that you preserve the existing kernel. It might not be necessary
when installing a new kernel in parallel.
2.Determine the appropriate processor generation and set the default kernel configuration:

cd /usr/scr/linux

cp -p configs/kernel-*-i686.config arch/i386/defconfig

make mrproper

make oldconfig
3.Configure the Linux kernel, change parameters as needed using one of the following tools:

make config

make menuconfig

make xconfig
4.Build the new kernel (it will be a compressed image)

make dep bzImage
5.Build the modules:

make modules
Note:
You can either use one of the Red Hat template kernel configuration files and
modify it or just start the kernel configuration tool, walking through all parameters.
After saving the configuration, you will find a
.config
file in
/usr/src/linux
. Make a
backup copy of the
.config
file for future reference (
make mrproper
deletes this file).
Note:
You can run all the compile steps and install the modules with this command:
make dep bxImage modules modules_install

228

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
You can get an updated version of the procedure to configure the Linux kernel, from the Web:
http://osl.storage.sanjose.ibm.com/cgi-bin/docs.cgi?/linux/Distributions
14.7.2 Loading the I/O adapter card as a module
Refer to the Qlogic Readme on the Qlogic page for detailed information on loading this card
as a module, at the following Web site:
http://www.qlogic.com/support/product_resources.asp?id=113
Click
Linux
and take a look at the 2.4 kernel readme.
This is the procedure:
1.Install the new kernel, its
System.map
and its modules from within
/usr/scr/linux

cp -p arch/i386/boot/bzImage /boot/vmlinuz-<version>

where <version> is the full kernel version as specified in this step.

cp p System.map /boot/System.map-<version>

make modules_install
2.Create an initial ramdisk image:

mkinitrd /boot/initrd-<version>.img <version>
3.Edit /
etc/lilo.conf
. Append the following lines:
– i
mage=/boot/vmlinuz-<version>

lable=new_kernel

initrd=/boot/initrd-<version>.img

read-only

root=/dev/your-root-partition
where your-root-partition is whatever you designated
it to be.
Note: In addition you might want to change the global
default=linux
line to
default=new_kernel
4.Test the modified
/etc/lilo.conf
and make the changes take affect by typing in the
commands:

lilo -t

lilo -v -v
5.Reboot the system and verify that the new kernel is running by typing in the following
command:

uname -f
14.7.3 I/O adapters in the host server
Type in the following command to list all instances of the installed adapters (this means the
number of adapters by number).
ls -l /proc/scsi/qla2x00


Chapter 14. Open systems host setup tasks
229
14.7.4 WWPN of the Fibre Channel I/O adapter
On Linux systems the required WWN is displayed in the system log file
/var/log/messages

when loading the driver. Type in the command:
/var/log/mesages
The output will display the WWN along with other information, look for a line similar to the
following:
Nov 2 08:55:10 skywalker kernel: scsi-qlal-adapter-port=210000e08b02e534
This WWN will be used when configuring the storage using the ESS Specialist
Modify Host
Systems
panel (See 15.2.4, “Modify host systems” on page 243) and specifying the
appropriate
Linux (x86)
host type.
14.7.5 Host recognition of newly assigned disks
All LUNs are automatically recognized after reboot. If you add LUNs later, the easiest way is
to reload the module and scan for attached devices.
To reload the module, type in the following command:
rmmod qla2x00
To scan for attached devices, type in the following command:
insmod qla2x00
14.7.6 ESS LUNs assigned as host disks
On Red Hat distribution, there are special device file entries available for the 128 devices. On
SuSE distribution, there are only special device files available for the first 16 devices. All other
devices have to be created manually using the
mknod
command.
The range of devices goes from
/dev/sda
(LUN 0) to
/dev/sddx
(LUN 127). To list, type in the
following command:
ls -l /dev/sda
14.7.7 Partitioning the disks
Before a file system can be created the disk has to be partitioned using the
fdisk
utility. You
have to specify the special device file of the disk you want to partition when executing
fdisk
.
For example:
fdisk /dev/sdb
14.7.8 Creating and using a filesystem
After the disk has been partitioned, as described in the previous section, the next step is to
create a file system. The file system can be created using either the
mke2fs
or
mkfs

command. Once the file system is created, it can be mounted and is ready to be used.

230

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
14.8 Novell NetWare
In this section we discuss the setup tasks for a Novell NetWare server.
14.8.1 I/O adapters in the host server
At the console prompt use the command:
list storage adapters
Each adapter appears as
A0
,
A1
,
Ax
where
x
is the adapter number.See Figure 14-13 on
page 230.
Figure 14-13 Adapter device list
14.8.2 WWPN of the Fibre Channel I/O adapter
Both Qlogic and EMULEX adapters can be used or you can find the WWPN in the BIOS of
the FC-Adapter.
Qlogic adapter
For the Qlogic adapter perform the following steps:
1.Restart the server.
2.Press
Alt+Q
to get the FAST!Util menu.
If you have more than one Fibre Channel adapter installed, all the Fibre Channel adapters are
displayed. Scroll down to the adapter you want. Press Enter.
1.From the FAST!Util menu, scroll down and choose
Select Host Adapter.
2.Scroll up and highlight
Configuration Settings
. Press Enter.
3.From the
Configuration Settings
menu, click
Host Adapter Settings
.
4.Write down the host adapter name, for example:
200000E0800C2D5
.
Emulex adapter
For EMULEX adapters, perform the following steps:
1.Boot to MSDOS.
2.Run the Emulex utility
LP6DUTIL
.
3.It will display the WWN, which is the number you are looking for.

Chapter 14. Open systems host setup tasks
231
14.8.3 Host recognition of newly assigned disks
At the system console, either reboot the host or type the following commands:
scan all A1
to Scan all LUNs on this SCSI adapter
s
can all luns
to Scan all LUNs on all adapters
A list of all new devices is displayed. You can now partions your new devices and make
volume groups.
14.8.4 ESS LUNs assigned as host disks
At the system console, type the following command (also see Figure 14-14.):
list devices
Figure 14-14 list devices output
To get the LUN ID serial number, you can use the command-line interface command:
rsList2105s
14.9 Sun
Sun systems with Solaris 2.5.1 and 2.6 will support up to eight LUNs per target, 0 to 7. Sun
Ultra B supports up to 32 LUNs, 0 to 31. You should check each individual system for their
specific support.
14.9.1 I/O adapters in the host server
Please check with your system administrator and physically identify them. You can also type
in the command:
prtconf
Here is a sample output:
QLGC,isp, instance #4

232

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
14.9.2 WWPN of the Fibre Channel I/O adapters
After installation of the adapters and restart of the system, view the
/usr/adm/messages
or
/var/adm/messages
log file and search for the line that contains the following:

For JNI Bus adapters, search for
fcawx: Fibre Channel WWNN
, where
x
is the adapter
number (0, 1,...) You can find the WWNN on the same line following the WWNN.

For JCI PCI adapters, search for
fca-pcix: Fibre Channel WWNN
, where
x
is the adapter
number (0, 1,...) You can find the WWNN on the same line following the WWNN.

For the Qlogic QLA2200F adapter, search for
qla2200-hbax-adapter-port-name
, where
x

is the adapter number (0, 1, ..)

For the Emulex LP8000 adapter, search for
lpfcx: Fiber Channel WWNN
, where
x
is the
adapter number (0, 1, ..)
14.9.3 Host recognition of newly assigned disks
Boot the host with
reboot -- -r
The host should boot properly and mount all the volumes. VERITAS Volume Manager will
identify all of its volumes. If you can not reboot right away you can enter in the series of the
next three commands:
drvconfig
disks

devlinks
14.9.4 ESS LUNs assigned as host disks
Type in the following command:
vxdisk list
Figure 14-15 shows a sample output:
Figure 14-15 LUNs listing, sample output -Sun
You can list the ESS LUN serial number to the host with the following command-line interface
command:
rslist2105s
Note:
If you have multiple I/O adapters installed, you will see more than one WWPN.
DEVICE TYPE DISK GROUP STATUS
c0t2d0s2 sliced - - online
c0t3d0s2 sliced rootdisk rootdg online
c1t1d2s2 sliced barkydg08 barkydg online
c1t1d3s2 sliced barkydg07 barkydg online
c1t1d4s2 sliced barkydg06 barkydg online

Chapter 14. Open systems host setup tasks
233
Figure 14-16 show an example output:
Figure 14-16 Output for rslist2105s - LUN serial numbers - Sun
Note that 021 is the LUN ID and 17005 is the ESS serial number.
14.10 NUMA-Q
To complement the information presented in this section, you can refer to the following Web
site:
http://techdocs.beaverton.ibm.com
14.10.1 I/O adapters in the host server
Run
devctl -c fabricn
, in where n is the number of the FC fabric connected to the target
volume. DYNIX/ptx will then discover the device.
14.10.2 WWPN of Fiber Channel I/O adapters
Type in the following command:
infodev -a
In the following example, the Fibre Channel I/O adapter whose WWPN is sought is
ff0
. The
line that begins with
Port WWN
contains the desired value. Perform this step for each Fibre
Channel I/O adapter that is to be configured.
/etc/infodev -a ff0
Figure 14-17 shows a sample output:
Figure 14-17 Identifying the WWPN - NUMA-Q servers
disk name 2105 serial number
--------- ------------------
c4t4d1 02117005
c5t1d7 10F17005
c6t1d5 20D17005
c8t4d4 02417005
Vendor ID : Emulex
Product ID : Firefly
Revision Level : FireFly 4
Device Class : ctlr
Node WWN : 10:00:00:00:c9:20:01:12
Port WWN : 10:00:00:00:c9:20:01:12
Located on : quad0

234

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
14.10.3 Host recognition of newly assigned disks
For each ESS storage device, do the following:
1.In the
NUMA Center Configuration Tool
screen, select
ESS Storage Device 1
(or
2
,
3
,
etc.) in the left-hand pane.
2.Update the IP Address 1 and IP Address 2 fields to reflect the subnet and/ or new host ID.
3.Fill in the
Description
box if desired.
4.Save the changes. From the
File
menu, choose
Save
.
Configure your system by following the instructions of your host system publication using the
ConfigApp
system administration tool. For detailed information you can visit the following site:
http://techdocs.beaverton.ibm.com/docs/ncrnab00/ch_5.htm#SE178140
14.10.4 ESS LUNs assigned as host disks
Use ptx/SVM's
vxdctl enable
command to list all ptx/SVM disks known to the operating
system's naming database.
The command
rsList2105s.sh
is used to list devices recognized by DYNIX/ptx. It is limited
by the value of the parameter
ARRMAX
, which defines the maximum number of items that an
array can contain. In the DYNIX/ptx operating system, that value is 1024.
Thus, one DYNIX/ptx instance can keep track of a maximum of 1024 device files on the
system. Since each logical volume within an ESS unit is mapped to a DYNIX/ptx device file,
1024 is the maximum number of logical volumes that can be configured on an ESS unit for
access byDYNIX/ptx.

© Copyright IBM Corp. 2002
235
Chapter 15.
ESS configuration for open
systems fixed block storage
This chapter describes detailed procedures using the IBM TotalStorage Enterprise Storage
Server Specialist (ESS Specialist) to configure fixed block storage in the ESS for use in the
open systems environment. At the end of the chapter we present sample scenarios that walk
you through the steps in an orderly fashion.
15

236

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
15.1 Introduction
The IBM TotalStorage Enterprise Storage Server Specialist (ESS Specialist) may be used to
do the complete storage configuration process for an ESS, defining S/390 and fixed block
(FB) storage, or it may be used to modify an existing configuration. For example, the ESS
Specialist might be used to define hosts and add volumes for one of the array configurations
that can be ordered for a new ESS.
The ESS Specialist is described in Chapter 5.1, “ESS Specialist” on page 80. To compliment
the information presented in the chapter you can refer to the IBM publication
IBM
TotalStorage Enterprise Storge Server Web Interface User’s Guide,
SC26-7346 (see
Section 1.5.3, “ESS documentation Web site” on page 10).
On the ESS Specialist
Storage Allocation
panel, two buttons appear, see Figure 15-1:
Figure 15-1 ESS Specialist: Storage allocation graphical view
Open System Storage
This is used to configure RAID and non-RAID FB storage
S/390Storage
This displays configured S/390 storage in tabular form, and leads
to further panels that are used to configure CKD storage.
15.1.1 Logical subsystem (LSS)
The ESS uses the concept of the Logical Subsystem (LSS) to internally manage logical
volumes (LVs). A fixed block (FB) logical volume in the ESS is known to host systems as a
LUN (logical unit number) associated with a Target ID in a SCSI or Fibre Channel host
adapter port. Open host systems do not need to know about logical subsystems in the ESS.
However, the internal organization of the ESS does impose some rules on how logical
devices are configured. Knowing the architecture will help you visualize and plan the
configuration. Figure 15-2 shows the relationship between clusters, LSSs, disk groups and
LUNs for an ESS with only 8 LSSs configured.

Chapter 15. ESS configuration for open systems fixed block storage
237
Figure 15-2 Relationship between clusters, LSSs, disk groups and LUNs
In the ESS, logical volumes reside in RAID arrays (disk groups) or non-RAID disks, on SSA
loops. Each SSA loop has from two to six disk groups of eight drives (not to be confused with
the physical drive enclosures, or 8-packs in which the disks are installed). In Figure 15-2, we
show the information for an ESS of only eight LSSs. These eight logical subsystems for FB
storage in the ESS are numbered from 10 to 17. You can have up to 16 LSSs for FB storage,
the last eight being numbered from 18 to 1F. We are only showing sixteen disk groups, two
within each LSS. We hope that this example will help you visualize the architecture of the
ESS.
In the next example (Figure 15-3), DA 2 manages up to three disk groups on DA pair 2, loop
A, and up to three disk groups on loop B. Those disk groups, and the logical volumes on
them, are associated with LSS 12, in cluster 1. The disk groups on DA pair 2, loops A and B,
that are managed by DA 3, are in LSS 13, managed by cluster 2.
LSS Disk Group
LUN
Cluster

238

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Figure 15-3 Relationship between LSSs DAs and SSA loops
If the ESS has S/390 storage configured as well as FB storage, there is a second,
independent set of LSSs that manages the S/390 logical volumes. These are numbered 00 to
0F (see Figure 15-3). Each of them corresponds to one S/390 Logical Control Unit (LCU). A
rank (RAID array or JBOD) may be formatted for FB and managed by an LSS, or for S/390
and managed by an LCU. Unlike an FB LSS, an S/390 LCU must be defined to the ESS, and
also to any S/390 hosts that use logical volumes in the LCU.
Summary of LSS attributes
An LSS is uniquely associated with one storage cluster, one DA, and half of the disk groups
on each loop. The FB ranks in those disk groups, and the logical volumes on the FB ranks,
are managed by the LSS.

An LSS may have from zero up to six disk groups.

There can be up to 16 FB logical subsystems (LSSs) in an ESS (and from zero to sixteen
S/390 LCUs as well).

An LSS can have up to 256 logical volumes (LUNs) assigned.

An LSS may have as few as one LUN assigned. That is, a single LUN may contain virtually
all the FB capacity in an entire array, and this array may occupy the only disk group in the
LSS.
Open host systems do not need to be aware of the internal logical structure of the ESS, but
logical subsystems are visible to the ESS user in several ways:

When managing PPRC using the ESS Copy Services Web interface, you can specify a
group of volumes either by ESS subsystem, or by LSS.

When using FlashCopy, volumes can be copied only to other volumes in the same LSS.

Each target ID in an ESS SCSI port maps to one LSS. That is, all addressable logical unit
numbers (LUNs) under a target ID map to logical volumes in the same LSS. This means
that when volumes from more than one LSS are added to an SCSI port, a separate target
ID in the port is assigned for each LSS. Note that Target IDs do not apply to Fibre
Channel.

Chapter 15. ESS configuration for open systems fixed block storage
239
15.1.2 FB logical volumes and SCSI LUNs
A fixed block logical volume is often referred to as a LUN, because it maps to a single LUN in
any SCSI port to which it has been added.

Unlike S/390 volumes, fixed block volumes must be configured to an ESS SCSI port
before a host can address them.

A fixed block volume can be configured to two or more SCSI ports. This allows multiple
paths from a host to the volume, and it allows hosts on different SCSI buses to address the
volume.

An LSS can appear as only one target ID at any SCSI port. A volume can appear as only
one target/LUN per port.

The volume may appear as a different Target/LUN in each SCSI port to which it is added.
(Note that there is an option to make it the same when modifying volume assignments).
Targets and LUNS are assigned to a port in the order that volumes are added. The
target/LUN assigned for a volume depends on what has been previously assigned to that
port for the same, and other logical subsystems.

One target ID on the SCSI bus must be reserved for a host initiator. Up to 14 targets on a
given LSS may be used by other hosts or devices on the bus. An ESS SCSI port can have
up to 15 addressable target IDs. Each target ID maps to a single LSS. Any one SCSI port
cannot address volumes from more logical subsystems than the number of target IDs
available.

Because ESS supports the SCSI 3 protocol, each addressable target ID in a SCSI port
can support up to 64 LUNs. This allows any one ESS SCSI port to support up to 960
LUNs, of the maximum of 4096 logical volumes that can be configured in an ESS.

UNIX and NT hosts may not support 64 LUNs per target. The number of LUNs
addressable per target is limited to the number supported by the host operating system
(typically 1, 8, or 32).
15.1.3 FB logical volumes and Fibre Channel LUNs
The host sees the FB logical volume as a LUN, because it maps to a single LUN in any Fibre
Channel host to which it has been added.
Unlike LUNs assigned to SCSI ports, target IDs, and LUNs per target ID, is not valid with
Fibre Channel.
15.2 Fixed block storage configuration
In this section, we describe the procedures needed and the ESS Specialist panels used to
configure FB storage.
15.2.1 Task sequence
We recommend that you do the necessary steps for configuration of FB volumes in the order
indicated in Figure 15-4. On the ESS Specialist
Storage Allocation
panel there are five
buttons in the lower part of the panel. The functions you will use are:
Modify Host Systems
Define the open system hosts that will address logical
volumes in the ESS. You will define the SCSI or Fibre
Channel attached host systems by type and name. You will
also select the WWPN and assign Fibre Channel attached
host systems to specific ESS fibre ports.

240

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Configure Host Adapter Ports
Associate each SCSI or Fibre Channel port with the hosts it
attaches. For SCSI, define for each port the target IDs on the
SCSI bus. For FCP you will be defining the access mode
and also the topology.
Configure disk groups
Configure the fixed block RAID (or eventually non-RAID)
storage you require.
Add Volumes
Create logical volumes (LUNs) and associate them with
SCSI ports or Fibre Channel adapters.
Modify Volume Assignments
Associate logical volumes with additional SCSI ports or
Fibre Channel adapters.
Figure 15-4 Recommended configuration sequence for FB storage
15.2.2 Connecting to ESS Specialist
When you first log in to the ESS, click the
Storage Allocation
panel. Next, click
Open
Systems Storage
(see Figure 15-1 on page 236), and then you get to thee screen shown in
Figure 15-4. Next, click
Configure Disk Groups
.
31 2
4 5
Note:
When using the ESS Specialist, ensure that the terminal that the SSR uses is not
logged on via the serial port. If you see a message that one or more components are in
service mode, this is probably the cause.

Chapter 15. ESS configuration for open systems fixed block storage
241
15.2.3 Configure Disk Groups
The
Fixed Block Storage
panel appears as shown in Figure 15-5. In this panel is a table
showing all the available disk groups in the ESS. Any disk groups that are already configured
as S/390 are not shown. Use the scroll bar to view the entire table. You can select a disk
group on SSA loop A, or loop B on one of the four device adapter (DA) pairs in the ESS. Each
SSA loop can have up to six disk groups.
For the selected disk group, choose one of the following:
RAID Array
Creates a RAID 5 array (6+P+S, or 7+P) using eight disk drives on the
specified SSA loop
Non-RAID
Defines a group of disk drives as individual, non-RAID ranks
Undefined
Undefines an existing group
When you select
RAID
, a RAID 5 array is defined. The
Storage Type
and
Track Format
fields
in the table are RAID Array and Fixed Block. The Capacity field has the formatted array
capacity. In Figure 15-5, adapter pair 1, loop A, group 2 has a capacity of 105.24 GB. This is
the formatted capacity of a 6+P+S RAID array of 18.2 GB disk drives.
Figure 15-5 Configuring fixed block disk groups
If you select
non-RAID
, the row in the table expands to eight rows, one for each of the disk
drives in the group. See Figure 15-6. The
Storage Type
and
Track Format
fields in the table
are non-RAID and None. The
Capacity
field has the unformatted capacity of a single disk
drive.
Non-RAID disks can then be individually selected, and the track format changed to fixed
block. A non-RAID disk group may contain both FB and CKD disks in any combination. If a
disk group is defined as non-RAID using this panel and any disks in the group are not defined
as FB track format, these disks are available to be defined as CKD track format using the
S/390 Storage
panels. Similarly, if a non-RAID disk group is defined using the S/390 panels,
any disks not defined as 3390 appear in the
Fixed Block Storage
panel (see Figure 15-5) and
are available to be defined as FB.

242

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
In Figure 15-6, none of the non-RAID disks has been defined as fixed block.
Figure 15-6 Configuring fixed block storage
When the required groups are defined in the panel, each array or non-RAID disk you have
created has
Defined
in the
Modification
column of the table. See Figure 15-5 on page 241.
Click
Perform Configuration Update
. Message 1402 is presented. Click
Yes
to continue.
See Figure 15-7.
Figure 15-7 ESS Specialist: Message 1402
Deleting a disk group
You can undefine an existing array, or a group of non-RAID disks by selecting Undefined.
Warning 1802 is presented. See Figure 15-8 on page 243.

Chapter 15. ESS configuration for open systems fixed block storage
243
Figure 15-8 ESS Specialist: Warning 1802
Note that all logical volumes and data on the disk group will be discarded if you proceed. To
cancel the reconfiguration, click
No
and then click
Cancel Configuration Update
to return to
the
Storage Allocation
panel now.
If you want to proceed with the deletion, click
Yes
. The disk group has
Undefined
in the
Modification
column. Click
Perform Configuration Update
. When a single non-RAID disk is
undefined, the entire disk group is deleted. If any of the disks in a non-RAID group has been
defined as a S/390 device, you cannot delete the group from here without first going to the
S/390 Configure Disk Groups
panel and changing the track format of all disks in this group to
none. If all the disks in the group that are visible here have their track format changed to none,
the group can be deleted from the
Configure CKD Disk Groups
panel.
15.2.4 Modify host systems
Open hosts must be defined to the ESS in order to access ESS logical volumes.
To define a SCSI or Fibre Channel host to the ESS, on the
Storage Allocation - Graphical
View
panel, select
Modify Host Systems
. The panel appears as shown in Figure 15-9. In
this example you can see previously created hosts. Enter in the Nickname, Host Type, Host
Attachment and Hostname/IP Address fields.
Note:
Either the Hostname or the IP Address can be entered. If you know both, you can
enter both, but it is not required to enter both.

244

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Figure 15-9 ESS Specialist: Modify Host System panel for SCSI
To configure a Fibre Channel host, the Nickname, Host Type, Hostname/IP Address, WWPN
and Fibre Channel Ports, are the required fields that you need to define. See Figure 15-10.
Figure 15-10 ESS Specialist: Modify host systems panel for Fibre Channel

Chapter 15. ESS configuration for open systems fixed block storage
245
Host Nickname
We suggest that you define a meaningful nickname. For example, if you have two Fibre
Channel ports attached to your host (hostname kolob) and the first host port adapter is fcs0,
then give the first host port the nickname of kolob_fcs0 and the second port the nickname of
kolob_fcs1.
World-Wide Port-Name
Enter in the WWPN for that adapter.To obtain the WWPN refer to the instructions on how to
get the WWPN in Chapter 14, “Open systems host setup tasks” on page 211, for your host
type. In some cases if the host is physically connected to the ESS then the WWPN will appear
in the pull down box.
Fibre-Channel Ports
You will need to determine if you are connecting through a direct cabling scheme or through a
switch fabric. If you are connecting through a Switch Fabric, for example a 2109 switch, then
we suggest you choose
All installed ports
, and configure the zoning in the switch to make
the connection from the ESS to the host. If you are connecting directly without a switch, then
you may choose the bay, port and adapter. See the drop down box in Figure 15-11.
Figure 15-11 ESS Specialist: Fibre-Channel Ports
When you create FB logical volumes and associate them with specific SCSI or Fibre Channel
ports, the ESS Specialist uses the host type information to determine how many SCSI and/or
LUNs can be created for each target ID in the port. Fibre Channel supports 1 target per host,
16,000 LUNs per target and 128 initiators per FCP. The port uses the defined SCSI or Fibre
Channel parameters for the hosts when communicating with a host on the bus.
For SCSI connections the ESS conforms to the SCSI-3 standard, and supports a maximum of
32 SCSI port (two ports per adapter), a maximum of 15 targets per SCSI adapter, a maximum
or 64 LUNs per target, depending on the host system type and a maximum or 512 SCSI-FCP
host logins or SCSI-3 Initiators per ESS. Hosts using the SCSI-2 standard support 16 target
IDs per bus, but only eight LUNs per target ID. Refer to Chapter 13, “Open systems support”
on page 205 to determine the maximum number of supported LUNs per target ID for your
host type.
You need to plan the storage allocation process carefully to ensure that you can define your
required configuration, or modify it as needed.
Host Name
To add a SCSI host on the
Modify Host Systems
panel enter the name (128 characters max)
by which you will identify the host in the Host Name field. If the host is LAN connected, we
suggest that you use the server’s TCP/IP hostname (including domain).

246

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Host Type
Select the host type from the pull-down list. See Figure 15-12 on page 246.
Figure 15-12 ESS Specialist: Host Type drop down box
The ESS may support additional SCSI host systems to those in this list. For current
information on supported hosts, refer to:
http://www.storage.ibm.com/hardsoft/products/ess/supserver.htm
If your host does not appear in the list, select a host type similar to yours. You can edit the
SCSI parameters for the host later on the
Configure SCSI Ports
panel.
Click the
Add
button. The list of hostname appears in the
Hosts Type
pull down box. You can
add up to 64 hosts if required.
Removing a host
You can remove any of the previously defined hosts that are no longer required by selecting
the
Host Systems List
of the
Modify Host Systems
panel and clicking on the
Remove
button.
See Figure 15-12.To activate the changes, click the
Perform Configuration Update
button.
Notes:


The action of removing a system from this list will not affect machine operation.
Therefore, you can use this process (remove and add) to modify a host name or type
entry.

You should remove a host system from the list only after you disconnect it physically
from the storage server.

Chapter 15. ESS configuration for open systems fixed block storage
247
15.2.5 Configure Host Adapter Ports
To define the hosts that will attach to a port, on the
Open System Storage
panel, click
Configure Host Adapter Ports
. See Figure 15-4 on page 240.
SCSI adapter ports
The panel appears as shown in Figure 15-13 for SCSI adapters.In this example, the
hostname
tucke
r has been added to the
Map of SCSI IDs on Bus
. Please ignore it for this
explanation.
Figure 15-13 ESS Specialist: Configure SCSI host adapter ports
The
Configure Host Adapter Ports
panel identifies the attachment of one or more host
systems to an ESS SCSI host port. A host system may attach to multiple ports via a separate
SCSI adapter and bus cable for each port. A host system may also attach to two ESS ports
using a single SCSI adapter and SCSI cable. In this case, because the ESS ports provide
electrical termination of the bus, they must be at the two ends of the cable. The two ESS ports
may be in the same, or different ESS subsystems.
The panel is used also to reserve target IDs for initiators and devices on the bus external to
the ESS port. The remaining IDs are available for allocation to the ESS port when you add
logical volumes.
Selecting a SCSI port
The first row of the panel (see Figure 15-13) contains icons for the installed host adapters.
The ESCON,FICON and Fibre Channel adapter icons have a different appearance than the
SCSI adapter icons. In Figure 15-13, the SCSI adapters are adapter that appear to be void of
any marking in them. Each adapter contains two selectable ports. You may select only one
port at a time for configuration.
Select a port either by clicking on the port icon in the first row or by scrolling the port selector
box to show the desired port number. The selected port is highlighted in yellow and the port
number is identified in the
Host Adapter Port
field.
SCSI adapter port

248

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
If a host system is attached to two ports via a common SCSI bus cable, use the
Second Bus
Connection (optional)
field, to identify the ESS port attached to the other end of the cable.
The panel will configure both ports according to a common
SCSI Bus Configuration
and
ensure that the
Map of SCSI IDs on BUS
for the ports are consistent with each other.
From the
SCSI Bus Configuration
pull-down, select the type for the host you are attaching to
the port.
A list of previously defined hosts (if any) that match the selected bus configuration appears in
the
Available Hosts
list.
Editing a SCSI bus configuration
If you want to examine or modify any of the SCSI bus parameters for your host system, click
the
Edit
button. The
SCSI Bus Configuration
window is displayed. Refer to Figure 15-14.
Figure 15-14 ESS Specialist: SCSI bus configuration window
In this example, the IBM_RS/6000 bus configuration is displayed. You can select any
available configuration from the
Bus Configuration Name
pull-down list on this panel. Any
parameter can be changed by selecting an alternative value from the associated pull-down
list. The altered configuration must be saved under a name you specify in the
Save As
field.
You cannot alter the original bus configuration. Click
Perform Configuration Update
to save
the new bus configuration.
Click
OK
to return to the
Configure Host Adapter Ports
panel. Your newly defined bus
configuration should now appear in the
SCSI Bus Configuration
list. If you select this
configuration, any host systems associated with the original bus configuration (in this
example, IBM_RS/6000), should appear under the new configuration, provided that you did
not alter the bus width parameter from 16 bits to 8 bits. The hosts also continue to be
available under the original bus configuration.
Adding hosts
After you have selected the correct
SCSI Bus Configuration
as described above,
Available
Hosts
lists contains the previously defined hosts that conform to the selected SCSI bus
configuration. Another entry, “Unrelated Host or Device” also appears. See Figure 15-13 on
page 247. This is normal and should appear here. Select another host from the
Available
Hosts
list.

Chapter 15. ESS configuration for open systems fixed block storage
249
In the
Map of SCSI IDs on Bus
, ensure that there are no host entries or
Unrelated Host or
Device
entries that are not required. Any such entries restrict the Target IDs available for
allocation to hosts you add, and to logical volumes you will add later. Remove unwanted host
entries first, then unwanted
Unrelated Host or Device
entries. If the last remaining entry is not
required, change the target ID to 8. You can remove it after you have added a host.
Ensure that any remaining host or
Unrelated Host to Device
entries are set to the correct
target IDs. Any entry
“Volumes Attached to SCSI ID”
cannot be changed to a different target
ID. See Figure 15-13 on page 247. These entries are previously assigned ESS logical
volumes. However, an entry can be deleted if not required. Any data on the volumes is
discarded.
Click
Add
. The selected host in the
Available Hosts
list appears in the
Map of SCSI IDs on
Bus
list. The first added host defaults to SCSI ID 7. This is the most commonly used target ID
for an initiator.
The SCSI ID is the SCSI ID of the adapter residing in the host. The adapter in the host comes
with a default address which is usually set to a specific value. Typically for single host SCSI
adapters, the SCSI ID is 7, but we recommend consulting the documentation for your system
to determine the ID and for instructions on how to change it.
Using the spin buttons, the target ID can be changed to any value 0-15 to match the target ID
of the host.
You can attach up to four hosts to any one ESS port, but they must be of the same host type.
If you select a different
SCSI Bus Configuration
to the one currently in use for the port,
warning 1610 is presented. See Figure 15-15.
Figure 15-15 ESS Specialist: Warning 1610
Second, and subsequent added hosts default to target IDs 6,5,4 unless existing entries in the
table have these IDs. Although the ESS supports up to four hosts per SCSI port, IBM does
not recommend multiple hosts on a bus because it results in lower performance due to
increased bus arbitration. Where multiple hosts are configured, we recommend that the target
IDs be set to 7,6,5,4. This provides the most efficient arbitration.
The host systems must support attachment of the number of hosts you want to add to the
bus. All SCSI IDs in the
Map of SCSI IDs on Bus
list must be unique. This is not checked by
ESS Specialist until you click the
Perform Configuration Update
button. The message
Error 1519: Duplicate SCSI ID Assigned
is presented if any ID appears more than once in
the list.
Figure 15-16 shows an example
Map of SCSI IDs
. The first host,
TestRS1
has been added at
target ID 7. Then two volumes were added to the port using the
Add Volumes
panel. They had
target IDs 6 and 5 assigned automatically by ESS Specialist. The volumes are attached at
different IDs because they are not in the same LSS.

250

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Figure 15-16 ESS Specialist: SCSI ID map
Subsequently, an unrelated host or device has been added at ID 4, and host
TestRS2
added
at target ID 3. The two volumes are available to both the
TestRS1
and
TestRS2
hosts.
You can control the target IDs at which logical volumes are later added on the
Add Fixed
Block Volumes
panel. When all the necessary hosts, and
Unrelated Host or Device
entries
have been added to the port, you can reserve any of the remaining unassigned IDs by adding
an
Unrelated Host or Device
entry set to that ID. This can be removed later if you want to
assign the ID to another host, or to additional logical volumes.
Note that when adding logical volumes to a port, ESS Specialist assigns new target IDs to the
port as required, in the order 7 to 0, then 15 to 8, skipping any IDs already assigned to:

Hosts

Unrelated hosts or devices

Existing logical volumes
If two ESS ports are on the bus, each port must have
Unrelated Host or Device
entries for any
target IDs assigned to logical volumes in the other port. This prevents possible duplicate IDs
on the bus if further volumes are added to either port from a new LSS. This is managed
automatically by ESS Specialist if the
Second Bus Connection
field is used to specify the
second ESS port on the bus.
The number of devices you can add to the port or, more precisely, the number of different
Logical Subsystems (LSS) to which added volumes can belong is limited to the number of
unreserved target IDs available for the port.
When the required hosts have been added, and the
Map of SCSI IDs on Bus
list is correct,
click the
Perform Configuration Update
button.See Figure 15-13 on page 247.
Once you have defined the SCSI host, you can connect it to the ESS. The physical link is
made by connecting a suitable cable between the ESS port and the host SCSI port. If other
SCSI hosts or devices are cabled onto the bus, the ESS port must be at one end of the cable.
The physical connection should be made after defining the connection through the ESS
Specialist
Configure Host Adapter Ports
panel.
Each SCSI adapter has two ports: A and B. In the host adapter icons on the panel, port A is
on the left, port B on the right. On the adapter, port A is the upper SCSI connector. SCSI
adapters in the ESS can be located anywhere in the four host adapter bays.
Fibre Channel adapter ports
To define the hosts that will attach to a port, on the
Open System View
panel, click
Configure
Host Adapter Ports
.

Chapter 15. ESS configuration for open systems fixed block storage
251
Selecting a Host Adapter port
The first row of the
Configure Host Adapter Ports
panel contains icons for the installed host
adapters. The ESCON, FICON and SCSI adapter icons have a different appearance to the
Fibre Channel adapter icons. In Figure 15-17, the Fibre adapter is the third one, in bay1.
Figure 15-17 ESS Specialist: Configure Fibre port panel
Select a port either by clicking on the port icon in the first row or by scrolling the port selector
box to show the desired port number. The selected port is highlighted in yellow and the port
number is identified in the
Host Adapter Port
, drop down box.
Figure 15-17 shows the following fields:

Fibre Channel Access mode. This field is greyed out and not selectable if you chose
Access Restricted
when planning with the configuration worksheets. This function
enables a from of zoning and LUN masking by the use of the Nicknames assigned to each
host adapter port.

World Wide Port Name. This field is greyed out and is the WWPN of the ESS I/O host
adapter port, not to be confused with the open system host adapters WWPN. You can not
change the WWPN of the ESS Host (I/O) Adapter Port.

Fibre Channel Topology. You may choose
Point to Point (Switched Fabric)
or Arbitrated
Loop.

Fibre Channel Protocol. You can only choose
FCP (Open Systems)
.
15.2.6 Adding fixed block volumes
In order for hosts to access data located in the ESS, the ESS storage must be partitioned into
LUNs. Each FB logical volume is seen by the host as a SCSI target/LUN, even though the
LUN is physically allocated in SSA disk storage in the ESS. A LUN cannot span ranks, but
many LUNs in different ranks can be accessed through a single host port.

252

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Although LUNs must initially be defined to a particular host and port, a volume can
subsequently be added to one or more additional ports. The volume is accessible by any host
configured to any port to which the volume has been added.
LUNs are selectable from a range of sizes, from 0.1 GB up to 440 GB. For the iSeries and
AS/400, only 4.19, 8.59, 17.55, 13.5, 20.56, 35.16, 36 and 70.56 GB are available.
To define LUNs, and associate them with a SCSI or Fibre Channel port, on the
Open System
Storage
panel, select
Add Volumes
. The
Add Volumes (1 of 2)
panel is displayed. Refer to
Figure 15-18.
Figure 15-18 ESS Specialist: Add Volumes
Select host
On this panel, the first row contains icons for the defined host systems. The second row
contains icons for the installed host adapters. When you select a host by clicking its icon, the
background of the host icon changes to black to indicate selection, and the SCSI ports
attached to that system are highlighted in yellow.
Any disk groups that currently have volumes added to any of the highlighted ports are now
shown in color. The color code indicates current allocation of the space in the groups. See the
legend in the top right corner of the panel.
Select port
Select the desired port by clicking on the left side of the host adapter icon for port A, or the
right side for port B. The background of the selected port changes to black to indicate
selection.

Chapter 15. ESS configuration for open systems fixed block storage
253
Select disk groups
All disk groups available for allocation of fixed block volumes are now shown in color. Any
groups shown in outline only are assigned as S/390 storage. Select one or more groups to
have volumes allocated by clicking on each group. The background of the groups changes to
black to indicate selection. You can allocate volumes to any combination of FB disk groups by
selecting them here. You may want to include only RAID groups, or only non-RAID groups in
any one selection.
If you select a host or a port and then want to change your selection before clicking the
Next>>>
button, you can click the
Clear View
button and start again. You can deselect a disk
group simply by clicking on it.
If you select a non-RAID group, the panel shown in Figure 15-19 is displayed. (Note that
AS/400 does not support non-RAID LUNs.)
Figure 15-19 ESS Specialist: Non-RAID disk group
The disks colored green are to indicate that they are available for allocation of FB volumes,
we have selected disks 1 and 2. Their background is changed to black to indicate selection.
Once the desired disks in the array are selected, click
Done
.
Disk groups on different DA pairs are in different logical subsystems. Two, or three groups on
the same loop may be in one LSS, or two different LSSs. Where no expansion frame is
installed, there are only two disk groups per loop, and they belong to LSSs in alternate
clusters. On the
Storage Allocation -- Graphical View
panel you can click each group and view
its status in the Information box. On an SSA loop, even numbered groups belong to one
Cluster 1; odd numbered groups belong to Cluster 2.
Note:
You must select a host, one port, and at least one disk group before proceeding to
the
Add Fixed Block Volumes (2 of 2)
panel.

254

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
This is of interest because each LSS that has one or more disk groups represented in your
selection needs a new target ID in the associated SCSI port if you add volumes from those
groups, and that LSS has no volumes already allocated to the port. If there are no available
IDs for the port, you may still be able to add new volumes, provided you do not select disk
groups in any LSS that does not already have logical volumes allocated to the port.
When your selection of host, port adapter and disk groups is complete, click
Next >>
to
proceed to the
Add Volumes (2 of 2)
panel. See Figure 15-20. If you want to cancel instead,
then click
Cancel Adding Volumes
to return to the
Open System Storage
panel.
Figure 15-20 ESS Specialist: Add Volumes (2 of 2) panel
This panel is used to allocate volumes in the selected disk groups, and add them to the
selected SCSI or Fibre Channel port. We suggest that you add or carve out your LUNs from
largest to smallest. This will help you manage other functions like FlashCopy later.
Under
Available Free Space
are three items:
Storage Type
RAID 5 or Non-RAID
Available Capacity
The total unallocated space in fixed block ranks in the selected disk
groups. Selected non-RAID disk groups may have individual ranks
(disks) that are either defined as S/390 or undefined.
Maximum volume size
The largest contiguous unallocated space in any FB rank in the
selected disk groups.
Add volumes
To add volumes, follow these steps:

Click the
Storage Type
field at the top of the
Add Volumes (2 of 2)
panel.

In the
Select a Volume Size
field, use the scroll button to select the required volume
capacity you want, in gigabytes (GB).

In the
Number of Volumes
field, enter the number of volumes of this size you want to
create.

Chapter 15. ESS configuration for open systems fixed block storage
255

From
Volume Placement
, select to do either of the following:
– Place volumes sequentially, starting in first selected storage area.
– Spread Volumes across all selected storage areas.

Click the
Add >>
button.
Because the iSeries volumes and open systems volumes are incompatible, you cannot add
both types of volumes to a single host port on the
Modify Volume Assignments
panel.
You may add additional volume size and quantity selections until the required configuration for
the selected port has been defined. To remove a volume from the list, click the volume to
select it. It highlights. Click the
<< Remove
button. The Remaining Free Space area in the
Volumes to be added table is recalculated each time an entry in the list is added or removed.
Clicking on
<< Back
at any time returns you to the Add Volumes (1 of 2) panel.
When the
New Volumes
to be added list is correct, click
Perform Configuration Update
. See
Figure 15-20 on page 254. Clicking on
Cancel Configuration Update
returns you to the
Open System Storage panel.
On this panel, when you click the host for which you just added volumes, the color coding in
the disk groups shows the modified space allocation. If you have no further volumes to add,
click
Cancel Adding Volumes
to return to the
Open System Storage
panel.
Notes:


The selector fields do not allow you to select a volume size that exceeds the largest
possible volume size, and they do not allow you to select a combined volume size and
quantity that cannot be allocated in the total free space, selected storage.

If there is more than one rank in your selection of disk groups, you can control how the
volumes are distributed in the selected storage, using the
Volume Placement
options.
You can fill the selected storage areas sequentially with allocated volumes, in the order
the ranks (disk arrays or non-RAID disks) were selected, and in the order the volumes
are allocated. Alternatively, you can evenly distribute allocated volumes one per rank in
rotation around the selected ranks.

If both RAID disk groups and non-RAID disk groups are selected, the volumes are
allocated among both RAID and non-RAID ranks as described. This is probably not
desirable. To avoid this, include only RAID groups, or only non-RAID groups in any one
selection of storage.

If the selected host is an iSeries, the logical volumes created are iSeries volumes.
These use a 520 byte-per-block logical track format. For other open hosts, the added
volumes have the conventional 512 byte per block logical track format. For clarity, we
repeat here that the physical SSA disks in the ESS are formatted using a 524
byte-per-block format. This accommodates both AS/400 and open systems logical
volumes on the same physical disk or array.

256

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
15.2.7 Modify Volume Assignments
The
Modify Volume Assignments
panel is used to associate logical volumes with additional
ports, and to remove volumes from ports as needed. The
Modify Volumes Assignments
panel
is shown in Figure 15-21.
Figure 15-21 ESS Specialist: modify volume assignments panel
The
Modify Volume Assignments
panel is a table of eight columns. A minimum of one row
exists for each defined volume. If a volume is assigned to a port then the row identifies the
assigned host port and host system. If a volume is assigned to two or more Fibre Channel
hosts, the table contains an additional Host Nickname entry on the same row, for each extra
Fibre Channel host assignment. Refer back to Figure 15-21 to see four entries on the same
row. If a volume is assigned to two or more SCSI ports, the table contains an additional row
for each extra SCSI port assignment. See Figure 15-22.
Note:
The
Add Volumes
panel allows you only to create new volumes; it does not allow you
to remove volumes created in a previous operation. Entries in the
New Volumes
to be
added list, but not yet created can be removed. Once the volumes are physically allocated,
they cannot be deleted. You can, however, unassign volumes through the
Modify Volume
Assignments
panel. This removes all association between the volumes and any port, but
the volumes, and the data on them continue to exist. Undefining a disk group on the
Configure Disk Groups
panel deletes all volumes residing in the group. Any data on the
volumes is discarded.

Chapter 15. ESS configuration for open systems fixed block storage
257
Figure 15-22 ESS Specialist: Dual row SCSI port assignments
Volume table
These are the Volume table columns:
Volume
Displays the volume serial number if the volume is assigned to a host
port; otherwise the volume is unassigned. The volume serial is an
ESS-generated eight digit hexadecimal number created by
concatenating the logical volume ID and the ESS machine serial
number. The logical volume ID is a three digit number assigned in
sequence to logical volumes as they are created. The volume serial
number is unique among ESS subsystems. For example, if the serial
number of the ESS is FCA35, then the very first LUN created in LSS
10 will be 000-FCA35. The first three digits (000) are the LUN hex ID.
Location
Displays the location of the logical volume expressed as DA
pair/cluster/loop/rank/volume number. Rank is either array or disk
group, as applicable.
LSS
Displays the Logical Subsystem that the volume is in.
Volume Type
Displays the volume type as either AS/400 or Open Systems.
Size
Volume size in gigabytes.
Storage Type
RAID array or non-RAID.
Host Port
Displays the assigned SCSI port expressed as Bay/Adapter/Port. If a
volume is assigned to multiple ports, there is a row for each unique
volume and port number combination. For FC only one row is
displayed for Fibre Channel.
Hosts Nicknames
Displays a list of host system host nicknames that are associated with
the volume and port number combination.
Use the scroll bar to move the table past the viewing window as required.

258

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Sort options
Each column in the table has a pull-down combo-box, containing the entries no sort, first,
second, and third. Select these options in sequence for the columns of your choice to
establish a sort hierarchy. When a sort choice is selected for one column, it is no longer
shown for the other columns. Click the
Perform Sort
button to rearrange the table according
to the sort options you selected.
For example, if you select first for the hosts column, and second for the size column, the table
shows the entries sorted by host. The entries for each host are further sorted by volume size.
Volume selection
Click a single row in the table to select it. It highlights in grey. To select additional rows, use
Ctrl-left-click on each row. For a range of rows, select a row at one end of the range, then use
Shift-left-click on the row at the other end. To deselect highlighted rows, use Ctrl-left-click on
them. To deselect all but one highlighted row, simply click on it.
Select action
When configuring SCSI, you assign volumes to the port. When configuring Fibre Channel,
you assign LUNs to a host not a port. Perform one of the following actions for a selected
volume(s) by clicking one of the action buttons at the bottom of the panel:
Assign selected volume(s) to target hosts:
Adds selected volumes to the port/host you select from the pull-down in the associated
port selector box.
Use same ID/LUN in source and target:
Uses the same LUN ID when selecting the same LUN to assign to an additional port/host.
Unassign selected volume(s) from target hosts:
Removes selected volume(s) from a port/host.
Click the
Perform Configuration Update
button. You are returned to the same panel. To
return to the Storage Allocation panel, click
Cancel Configuration Update
, or on
Storage
Allocation
.
You cannot assign a volume to a port if:

The volume type (AS/400 or open systems) is not compatible with hosts assigned to that
port.

Addition of the volume requires that a new SCSI ID be allocated in the port and one is not
available (all 16 IDs on the bus are reserved or allocated). Addition of a volume does not
require a new SCSI ID to be allocated if the volume is in the same logical subsystem as
any existing device on the port.

The volume is in an LSS whose target ID in the port has the maximum number of LUNs
assigned for the port’s host type.
In each case, the port is not available in the selector box.
Note:
To add a volume, you must select a port in the selector box. See Figure 15-17 on
page 251.

Chapter 15. ESS configuration for open systems fixed block storage
259

Using the
Assign
and
Unassign
functions, you can:
– Assign a volume to a second or subsequent host port/host. This allows hosts on
different buses to address the volume, or a single host to address the volume over two
or more paths.
– Assign a currently unassigned volume to a host port.
– Remove a volume from a port/host.
– Move a volume from one port to another.
– Leave the volume unassigned by removing its last port/host assignment.

Leaving a volume unassigned does not affect the volume definition. The data remains on
the volume and will be accessible when the volume is next assigned to a port/host.

A volume, once created, cannot be individually deleted from the disk group. To remove
logical volumes from a rank (array or non-RAID disk), you have to undefine the entire disk
group. Once you undefine the group, all logical volumes and data on that group are lost.
Refer to Figure 15-6 on page 242. You can then redefine the group and add new volumes.

If a volume is accessed by more than one open system host, the hosts should be
architecturally compatible. The format used to store data on disk (that is, the file system)
should be the same. Data corruption may occur if hosts are not compatible. Compatibility
of hosts is partially enforced by allowing only one Host Type to be associated with a SCSI
port. However, this does not prevent a volume from being accessed by incompatible host
systems through different ports.
15.3 Sample procedures
In this section we present sample configuration procedures. We recommend that you do them
in the order listed, in order to obtain the best results.
15.3.1 Formatting RAID-5 disk groups
Perform these steps:
1.Log into the ESS Specialist application.
2.Click
Storage Allocation
.
3.Click
Open System Storage
.
4.Click
Configure Disk Group
.
5.Select disk group to configure (area will darken).
6.Select
Storage Type
,
RAID 5
.
7.Select
Track Format
,
Fixed Block
.
8.Click
Perform Configuration Update
to commit or click
Cancel Configuration Update
to
cancel.
15.3.2 Creating a SCSI logical host
Perform these steps:
1.Log into the ESS Specialist application.
2.Click
Storage Allocation
.
3.Click
Open System Storage
.

260

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
4.Click
Modify Host Systems
.
5.Enter Appropriate value into Host Nickname Field.
6.Select
Host Type
, RS6000 is default.
7.Select
Host Attachment Type
, SCSI is default.
8.Enter a valid host name in Host Name/IP Address field.
9.Click
Add
.
10.Click
Perform Configuration Update
to commit or click
Cancel Configuration Update
to
cancel.
15.3.3 Creating a Fibre Channel logical host
Perform these steps:
1.Log into the ESS Specialist application.
2.Click
Storage Allocation
.
3.Click
Open System Storage
.
4.Click
Modify Host Systems
.
5.Enter appropriate value into Host Nickname Field.
6.Select
Host Type
, RS6000 is default.
7.Select
Fibre Channel Attached
Type.
8.Enter a Valid host name in Host Name/IP Address field.
9.Enter the
World-Wide Port-Name
.
10.Select
All installed ports
.
11.Click
Add
.
12.Click
Perform Configuration Update
to commit or click
Cancel Configuration Update
to
cancel.
15.3.4 Assigning SCSI Host Adapters to a logical host
Perform these steps:
1.We recommend that you reserve SCSI ID 7 by setting an unknown device to the ID.
2.Make sure all adapters that are assigned to a single host have the same SCSI ID,
(preferably SCSI ID 6).
3.Use SCSI ID 6 for the primary HACMP node and SCSI ID 7 for the failover HACMP node.
4.Log into the ESS Specialist application.
5.Click
Storage Allocation
button in the Navigation Frame on the left side of the browser
window. This will bring up the graphical storage allocation screen.
6.Click
Open System Storage
button at the bottom of this window. This will bring up the
Open System Storage
screen.
7.Click
Configure Host Adapter Ports
button.
8.Select the adapter (for example: Bay1 Adatper1 PortA), from the graphical bay
representation or, from the
Host Adapter Port

selector bar.
9.Select the host type from the
SCSI Bus Configuration
selector bar.
10.Select the host name from the
Available Hosts
selector field.

Chapter 15. ESS configuration for open systems fixed block storage
261
11.Click
Add
.
12.The Host Name now appears in the Map of SCSI IDs on Bus. Select desired SCSI ID from
Selector Bar.
13.Click
Perform Configuration Update
to commit or click
Cancel Configuration Update
to
cancel.
15.3.5 Assigning Switched Fibre Channel host adapters to a logical host
Perform these steps:
1.Log into the ESS Specialist application.
2.Click
Storage Allocation
button in the Navigation Frame on the left side of the browser
window. This will bring up the graphical storage allocation screen.
3.Click
Open System Storage
button at the bottom of this window. This will bring up the
Open System Storage
screen.
4.Click
Configure Host Adapter Ports
button.
5.Select an adapter (for example: Bay1 Adatper1 PortA), from the graphical bay
representation or, from the
Host Adapter Port

selector bar.
6.Verify that
Storage Server Attributes
is
Access_Restricted
. This attribute is previously set
by the IBM SSR upon installation in a switch environment.
7.Select from the Fibre Channel Topology

selector bar,
Fibre Channel Point to Point
.
8.Click
Perform Configuration Update
to commit or click
Cancel Configuration Update
to
cancel.
15.3.6 Assigning Direct Fibre Channel host adapters to a logical host
Perform these steps:
1.Log into the ESS Specialist application.
2.Click
Storage Allocation
button in the Navigation Frame on the left side of the browser
window. This will bring up the graphical storage allocation screen.
3.Click
Open System Storage
button at the bottom of this window. This will bring up the
Open System Storage screen.
4.Click
Configure Host Adapter Ports
button.
5.Select an adapter (for example: Bay1 Adatper1 PortA), from the graphical bay
representation or, from the
Host Adapter Port
selector bar.
6.Verify that Storage Server Attributes is
Access_Restricted
. This attribute is previously set
by the IBM SSR upon installation in a Switch environment.
7.Select from the Fibre Channel Topology

selector bar,
Fibre Channel Arbitrated Loop
.
8.Click
Perform Configuration Update
to commit or click
Cancel Configuration Update
to
cancel.
15.3.7 Carving LUNs from RAID5 Disk Groups
All carving should take place using a temporary DUMMY Host and DUMMY Host Adapter.
This procedure requires that the DUMMY Host has already been created by you, and that the
DUMMY Host Adapter has already been assigned.

262

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Procedure
Perform these steps:
1.Log into the ESS Specialist application.
2.Carve disk groups in the order they were installed. Carve the entire disk group before
moving to the next.
3.Click
Storage Allocation
.
4.Click
Open System Storage
.
5.Click
Add Volumes
.
6.From the
Add Volumes (1 of 2)
panel. Select the host system to be used as the DUMMY
Host.
7.Select the host adapter to be used as the DUMMY Host Adapter.
8.Select the next disk group to be carved.
9.Click
Next
.
10.Click
Add Volumes (2 of 2)
:
– First LUN loop pass: Select the largest and primary LUN size for this disk group from
the
Volume Attributes
Panel.
– Second LUN loop pass: Select the next largest LUN size for this disk group from the
Volume Attributes
panel.
11.Select the LUN count for this LUN Size into the
Number of Volumes
entry field.
12.Click
Place Volumes Sequentially
13.Click
Add
14.Repeat the previous steps until the
Available Space
field indicates you have selected all
the space available.
15.After you have selected ALL available space in this disk group, then do the following:
– Click
Perform Configuration Update
to commit or click
Cancel Configuration
Update
to cancel.
– Repeat all the previous steps until all un-allocated disk groups have been carved.
– Notify to the IBM SSR to have them verify that ALL disk groups have been successfully
carved.
Note:
If there are no un-used/spare host adapters then you can temporarily use a REAL
Host and one of the REAL Host Adapters already assigned to it. We recommend that you
select a REAL Host not running HA or SDD.
Note:
Always start with the largest LUN size required for this disk group and work down;
for example: 16, 8, 1.
Note:
Use caution when selecting sizes (observe decimal placement in selections!)
Note:
Your entries will transfer to the
New Volumes

field, the
Available Space
field will
update and, your LUN size selections will change in the
Volume Attributes
field.

Chapter 15. ESS configuration for open systems fixed block storage
263
16.Un-assign all new LUNs from the DUMMY host and host adapter.
15.3.8 Assigning LUNs to Fibre Channel host adapters
Perform these steps:
1.Log into the ESS Specialist application.
2.Click
Storage Allocation
.
3.Click
Open Systems Storage
.
4.Click
Modify Volume Assignments
5.Sort by volume. (This is the default selection.)
6.Select the first Target LUN in
Modify Volume Assignments
panel. The selection will
darken.
7.Select the second and subsequent target LUNs in
Modify Volume Assignments
panel
while holding down the Shift key; each selection will darken.
8.Click
Assign Selected Volume(s)
to the target hosts after you have selected all LUNs for
current operation.
9.Select the first target host/adapter combination in
Target Host
field, the selection will
darken.
10.Select the second and subsequent target host/adapter combinations while holding down
the Shift key; each selection will darken.
11.After you have selected ALL host/adapter combinations for this operation, then do this:
– Click
Perform Configuration Update
to commit.
or
– Click
Cancel Configuration Update
to cancel.
15.3.9 Assigning LUNs to SCSI Host Adapters
Perform these steps:
1.Select next LUN group. See important notes previously, for criteria of a LUN group.
Extract the following information from the LUN from the tabular view of the
Modify Volume
Assignments
panel.
– Target LUN(s) IDs
– Target Host Name(s)
– Target Host Adapter(s)
2.Log into the ESS Specialist application
3.Click
Storage Allocation
.
4.Click
Open Systems Storage
.
5.Click
Modify Volume Assignments
6.Sort by Volume.
7.Click Volume Spinner Bar; Select:
First
.
8.Click
Perform Sort

9.Select the first target LUN in
Modify Volume Assignments
panel. The selection will darken.
10.Select the second and subsequent target LUNs in
Modify Volume Assignments
panel
while holding down the Shift key, each selection will darken.

264

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
11.Click
Assign Selected Volume(s)
to the target hosts after you have selected all LUNs for
current operation.
12.Select the first target host/adapter combination in
Target Host
panel, the selection will
darken.
13.Select the second and subsequent target host/adapter combinations while holding down
the Shift key; each selection will darken.
14.After you have selected ALL host/adapter combinations for this operation then:
– Click
Perform Configuration Update
to commit or click
Cancel Configuration
Update
to cancel.
– For each LUN, verify that the SCSI ID is identical for all of its adapter assignments;
see the procedure in the next section.
15.3.10 Verifying a LUN and SCSI IDs
This is the procedure to verify that the LUNs have identical SCSI IDs for each adapter
assignment:
1.Log into the ESS Specialist application.
2.Click
Storage Allocation
.
3.Click
Tabular View
.
4.Sort by Volume and Host.
5.Click Volume Spinner Bar, select:
First
.
6.Click Host Spinner Bar, select
Second
.
7.Click
Perform Sort
.
8.Scroll through
Storage Allocation - Tabular View
panel to the desired LUN.
9.You should see a LUN entry for each host adapter, assigned to it.
10.Verify that the volume has the same SCSI ID for each adapter to which it is assigned.

© Copyright IBM Corp. 2002
265
Chapter 16.
Data migration in the open
systems environment
This chapter gives you the information necessary to plan the methods and tools your
installation will be using when doing the migration of the existing data into the IBM
TotalStorage Enterprise Storage Server.
The information presented in this chapter is basically oriented to the open systems users
(UNIX; iSeries; and Intel-based servers environments) that must move data from non-ESS
storage subsystems into ESS storage-subsystems.
To complement the recommendations and information discussed in this chapter, you may
also refer to Chapter 4 (Migrating data to the ESS) of the
IBM TotalStorage Enterprise
Storage Server User’s Guide 2105 Models E10, E20, F10 and F20,
SC26-7296.
16

266

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
16.1 Migrating in the open system environment
Due to the broad differences in open systems environments, we suggest that you use the
general methodology we are presenting here and then determine the specific courses of
action based on your host platform.
You can complement the recommendations and information discussed in this chapter
referring to chapter 4 (Migrating data to the ESS) of the
IBM TotalStorage Enterprise
Storage Server User’s Guide 2105 Models E10, E20, F10 and F20,
SC26-7296.
16.1.1 Method selection
Most methods of data migration affect everyday operation of a computer system. When data
is moved, it must be in known state, typically requiring updates or changes to cease while the
movement occurs. Depending on the amount of data you are moving and your methods, data
could be unavailable for an extended period of time, perhaps several hours.
Factors like creation of new logical volumes or file systems, modification of configuration files,
and data integrity checks, contribute to the unavailability of data that you are migrating. This
section describes some of the aspects which you must consider when selecting a method to
use for migrating data.
Select the method that best meets your criteria for:

The amount of data to be migrated

The amount of time available

The availability of spare disks or tape devices for temporary storage

The format of the data itself
You should select the method that is the best compromise between efficiency and the least
impact on the users of the system.
16.1.2 Replacing existing storage
In general when the ESS replaces existing storage, you must partition it so that its logical
disks are similar in configuration to the drives that it is replacing. New configurations should
be large enough to accommodate the existing data.
7133 serial disks
The ESS supports existing 7133-020 or 7133-D40 drawers. Whenever these units are
attached to an ESS, the drives must be reformatted for ESS usage. The ESS uses a different
format and internal drive information. Therefore, you need to backup all data on drives that
you are migrating before commencing movement of the drawers.
For configuration guidelines, when attaching your existing 7133 Serial Disk to the IBM
TotalStorage Enterprise Storage Server you can refer to the IBM redbook
IBM TotalStorage
Enterprise Storage Server,
SG24-5465.
Non-IBM disks
It is possible to free up redundant cabling to external disk storage units and use that port to
connect the replacement ESS in parallel. This way you can mirror LUNs or disks on the
existing unit to the replacement unit, by means of software techniques. Once the volumes are
mirrored, you can then break the mirror from the existing unit, and remove the old unit.

Chapter 16. Data migration in the open systems environment
267
Table 16-1 shows an example of the overall procedure on and AIX platform, migrating from a
Symmetrix unit to an ESS unit. You may be able to apply the concept and strategy to other
open system platforms.
Table 16-1 Example migration procedure
Process Commands Explanation
Phase 1, Identify hdisks,
gatekeepers and channels

Identify the gate keepers on the
Symmetrix, look for the disks
that are 2880 in size. make note
of the rhdisk#
/usr/lpp/Symmetrix/bin/inq
where # is the number of the
rhdisk
Identify the new adapter
connections to obtain the SCSI
adapter number
lsdev -Cc adapter | grep scsi
If you are using Fibre Channel
then replace
scsi
with
fcs
Identify the hdisks on each of
two channels, make note of
which hdisk# resides on which
adapter#
lsdev -Cc disk | grep 10-08
For example: the scsi #
previous step is equal to 10-08
Phase 2, Copy lv data from one
hdiskpower on the Symmetrix
to the hdisk on ESS
Identify which Volume groups to
mirror
lsvg -o
Shows all volumes varied on
Check the current information
lsvg -l vg_name
lsvg -p vg_name
This identifies all the logical
volumes in the volume group
and what disk they are on
Add disk(s) to volume group.
Spread the disks evenly across
the vgs for best performance
extendvg vg_name hdisk#
Where
vg_name
is the volume
group name and # is the hdisk
number on the ESS unit.
Verify that new hdisk has been
added to the volume group
lsvg -p vg_name
Where vg_name is the volume
group name
Mirror data onto new hdisk
mklvcopy lv_name 2 hdisk
Where
lv_name
is the logical
volume name and # is the new
hdisk number you have added
Sync the volume groups
syncvg -l lv_name
Where lv_name is the logical
volume name
Verify that logical volume has
copied on hdisk#
and is still on hdiskpower#
lspv -l hdisk#
lspv -l hdiskpower#
Where
hdisk#
is the new hdisk
and
hdiskpower#
is the old disk
you are migrating from
Verify that the sync isn't
showing
stale, it
should
show as
sync’d
lsvg -l vg_name
If the lv still shows stale then
you need to resync it before
proceeding
Phase 3, remove the disk from
the Symmetrix
Once sync’d, remove the
mirrored copy from old
hdiskpower#
rmlvcopy lv_name 1
hdiskpower#

268

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
16.2 Data migration for iSeries systems
The ESS can attach to the iSeries by either SCSI (V3R1 or later) or Fibre Channel (V5R1
only). The steps to add the external storage to the iSeries are the same regardless of the
adapter. With a SCSI adapter, the LUNs will report into the iSeries as device type 9337. With
a Fibre Channel adapter the LUNs report in as 2105 device types.
DASD, direct access storage devices (more commonly known as disk drives) are either
internal to the iSeries or attached externally. In the iSeries, disk drives are grouped together
into auxiliary storage pools (ASPs). The iSeries expects to see a separate device address for
each drive in the subsystem. The ESS meets this requirement by reporting unique addresses
for each virtual drive defined to the iSeries.
If you are using the ESS to migrate into it, data that is currently on the internal drives of the
iSeries then you must consider that the Load Source Unit (LSU) is a special disk in the
iSeries. This is the disk that is used to IPL the system (among other things), it is similar to a
boot drive. All other “user data” can be located on external DASD units, but the LSU must be
always an internal drive. This is because the system can not be IPL’d from an I/O adapter
(IOA) supporting external drives.
If you have an external disk subsystem to the iSeries, that you are replacing with a ESS, you
need to migrate the data to the ESS. You can use your existing host utilities for this data
migration.To compliment this information, you can refer to chapter 4 of the
Users Guide
found
on the Web at:
http://ssddom02.storage.ibm.com/disk/ess/documentation.html
You can select from several methods to migrate data to the ESS:

You can use logical ADD and REMOVE functions.

You can use save and restore methods with tape devices.

You can also use these methods if you remove an existing disk subsystem before you
install the ESS.
Make sure data is off old
hdiskpower
lspv -l hdiskpower#
Make sure data is on new hdisk
lspv -l hdisk#
Move the old hdiskpower out of
the volume group
reducevg vg_name
hdiskpower
Remove the old device
definition from the host
lrmdev -dl hdiskpower#
Where x is the number of disk
you want to delete
Remove the cabling and the old
unit from the host.
Process Commands Explanation
Note:
Whether the ESS LUNs report to the iSeries system as Unprotected or Protected
(Redundant Array of Independent Disks), any OS/400 reference to starting or stopping
device parity on these disk units does not apply. With the appropriate V4R5 PTFs or with
V5R1, ESS LUNs created for iSeries as Unprotected can be used to mirror either the LSU
or other internal or external unprotected units.

Chapter 16. Data migration in the open systems environment
269
Since the ESS always reports to the iSeries as a Redundant Array of Independent Disks
(RAID) (protected 9337 or 2105), any reference to starting or stopping device parity does not
apply.
For information on how to use these procedures, see the AS/400 Backup and Recovery
Guides for the release of the operating system you have on your host system. Also refer to the
IBM redbook
iSeries in Storage Area Network A Guide to Implementing FC Disk and Tape
with iSeries,
SG24-6220.
16.3 Data migration for UNIX systems
For UNIX systems, you can use a variety of methods for copying or moving data from one disk
drive to another. The host sees the ESS as one or more generic devices.
16.3.1 Migration methods
In this section we discuss various migration methods.
Volume management software
Volume management software provides specific tools for wholesale movement of data.
Management software provides simple robust methods that you can generally use during
production without disturbing users.
AIX, Solaris, and HP-UX all have volume management software that directly controls the
disks and storage devices attached to the system. It provides the system administrator with
tools for configuring and managing disk drives, file systems, paging, and exchanging spaces.
The software also provides the operating system interface to the data.
Direct copy
If the data you are migrating resides as individual files on UNIX file systems, and no volume
management software is available, use a utility. A utility which supports a direct copy feature,
such as
cpio
with the
-p
(pass) option provides the next easiest method of moving the data.
The
cpio
feature is available on all of the UNIX operating systems which support the ESS.
Backup and restore
In some cases, the only method available to transfer data is to back it up to a tape device and
restore it to the new disk. This method is obviously slower, because tape devices are
essentially slow devices. However, if disks are being removed before you install the ESS, the
only way to move the data is with an intermediate tape device.
Backup and restore procedures will generally have the most impact on the system usage.
They require that databases and file systems be in quiescent states to ensure a valid
snapshot of the data.
Dump and restore
These commands are similar in function to the Backup and Restore command; you find them
on almost all forms of UNIX. They too require an intermediate device.
Other commands
You can find other commands on UNIX systems for archiving the data. Again, these
commands require that you create an intermediate archive, usually on a tape drive or spare
disk drive.

270

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
You may not be able to use the volume management methods of migrating data in the
following cases:

For databases that use raw file systems

For logical volumes or methods other than a UNIX file system
This is most probable if the database uses volume serial numbers in its licensing code, or
validity checking.
If databases use licensing methods or validity checking, you may only be able to:

Export the database from its old locations

Import the database to its new location
It is up to the database software to provide some mechanism to move the data. This may take
the form of a standard database backup and restore if it does not have any specific tools for
movement of data.
16.3.2 The AIX logical volume manager (LVM)
The AIX LVM provides methods that you can use at any time without disrupting access to the
data by users. You may notice a small performance degradation, but this is better than having
to shut down database or require users to log off the system.
The AIX LVM provides useful tools and utilities for migration of data as part of the AIX base
operating system release. You can use these tools to move data to and from the ESS as you
would on any other disk drive connected to an AIX system.
These methods for data migration work below the file system level. They do not care what
sort of data resides on the logical volume, whether it is a UNIX file system or a raw database.
Data migration using AIX LVM mirroring
This is the procedure for using AIX LVM mirroring for data migration:
1.First determine which logical volumes need to be migrated.
A sample scenario could be one in which old storage might be reused at a later time for
new data, after the current data is migrated to ESS storage. For example, you may want all
of the data residing on the SSA disks to be moved to the ESS. After this migration is
complete, the old SSA disks might be used on a different system or installed in the ESS for
future use. In this case, in order to move the data while keeping it available to users, you
will have to mirror the logical volumes in which the data is stored.
The following two steps describe how to identify those logical volumes:
a.Make a list of the physical disks (hdisks) that need to be migrated.
An hdisk may represent a RAID 5 array and, therefore, be associated with several
pdisks (physical volumes). In this case you will still have to mirror all the logical
volumes contained in the array, even if some of the pdisks in the array will not be
replaced by ESS storage.
b.Make a list of the logical volumes that need to be migrated.
For each hdisk, run the command
lspv -l hdiskx
where
x
is the hdisk number. This
output will give you the name of all the logical volumes that reside on the hdisk. If there
is a filesystem mounted on the logical volume, this output will also give you the mount
point for that filesystem.
Based on the outputs of your lspv
-l hdiskx
commands, create a list of all the logical
volumes that will need to be mirrored.

Chapter 16. Data migration in the open systems environment
271
2.Mirror the logical volumes.
Mirroring does degrade system performance slightly. However, most users will prefer
performance degradation to downtime. On the other hand, downtime may be scheduled in
order to facilitate faster mirroring and prevent users from having to put up with degraded
performance. Mirroring of individual logical volumes should be done in accordance with
the migration schedule that was agreed upon by the user.
The following steps describe how to mirror logical volumes.
a.Mirroring individual logical volumes
Individual logical volumes can be mirrored with the command:
mklvcopy -e m -s y -k lvname # hdiskx hdisky
Where
lvname
= the name of the logical volume without the /dev/ prefix
#
= the total number of copies desired of the logical volume (in most cases, this will be
2)
hdiskx
= the source hdisk
hdisky
= the target hdisk
Note: This command will create a synced copy of the logical volume. The amount of
time for the system to produce the mirrored and synced copy will depend on the
amount of I/O operations performed on the logical volume as well as the amount of
data in the logical volume.
b.Mirroring all the logical volumes in a volume group.
An entire volume group can be mirrored and synced with the command:
mirrorvg vgname hdiska hdiskb hdiskc
Where
vgname
= volume group name
hdiska, hdiskb, hdiskc
= the physical disks to use for each copy of the volume group
Note: This command inherently uses strict allocation, which keeps each copy of a
logical volume on a different disk. This command will not mirror a dump device if the
dump device is not paging space. Use the
mklvcopy
command instead.
3.Split the logical volume copies.
Once you have made duplicate copies of each of the logical volumes that need to be
migrated, the next step is to separate the copies. This will allow you to test the new copies
of the logical volumes to ensure that they contain all the customers’ information before you
remove the old copies.
The following process should be used to split each logical volume:
Note:
There may be logical volumes that span more than one hdisk. In this case, the
logical volume will be indicated in the output of more than one lspv command that
you ran. You should take care not to confuse the number of PPs that a Logical
Volume uses on a single disk with the number of PPs that a logical volume uses on
the entire system. In other words, do not rely on the output of this command to
indicate whether or not the logical volume is mirrored.

272

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
a.Close the logical volume by either unmounting the filesystem that is mounted on the
logical volume or by killing the process that is using the logical volume to be split. You
can see if any processes are using a logical volume by using the
fuser
command. You
can verify that a logical volume is closed by using the
lsvg -l vgname
command.
b.Use the following command to remove a copy of the logical volume from the mirrored
copies. This command will separate a copy from the mirrored copy or copies and
rename that copy, thereby creating a separate, new logical volume on the specified
hdisk.
splitvcopy -y newlv -c # hdiskx
In this command:
-y
= indicates that you will supply the name of the logical volume that will be created
newlv
= the name of the new logical volume to be created
-c
= indicates that you will designate the maximum number of copies that will be left
in the logical volume that you are copying from
#
= the number of copies that will be left in the logical volume that you are copying
from
hdiskx
= the hdisk that you specify here will contain the renamed copy of the logical
volume
4.Remount the filesystem.
The following process should be used to remount each filesystem, thereby making the
data on the logical volume available to users and applications
a.Verify that you have the original logical volume (the same name that is in
etc
filesystems, not the name of the new logical volume that was created by the
splitlvcopy
command) on the new, ESS disk.
b.Remount the filesystem by using the
mount
command or using
smit
5.Test the newly migrated data.
Now that you have a new copy of the data on ESS disk, you can go ahead and test all data
according to the test plan designed by the customer.
6.Remove the old copies of data.
You may or may not want to remove old logical volumes depending on whether the old
storage will be removed from the system.
a.Remove each logical volume copy.
b.You will need to recall the names you gave to the new logical volumes created by the
splitlvcopy
command.
Note:
It is recommended that you use a naming convention that is consistent for
all the new logical volumes that will be created by the
splitlvcopy
command.
This way, it will be easy to recall the names of the new logical volumes if you
need to remove them after testing. For example, if splitting a logical volume
called lv00 or hd6, use the names xlv00 and xhd6, respectively.

Chapter 16. Data migration in the open systems environment
273
16.4 Data migration for Windows NT and 2000
You can migrate data to an ESS by using standard Windows NT and 2000 commands. No
special tools or methods are required for transferring data to ESS. Windows NT and 2000
recognizes the logical volumes in an ESS as normal physical SCSI or FCP disks. For details
on migration methods and commands, refer to chapter 4 of the
Users Guide
found on the
Web at:
http://ssddom02.storage.ibm.com/disk/ess/documentation.html
16.4.1 Migration methods
There are several different methods you can use for migration data in Windows NT
environments. You can use for example basic MS-DOS commands, such as
Copy, Backup,
Restore
or you can use drag-and-drop technique to achieve the same result. You can also
find many products on the market from different vendors who provide the storage
management software for migration and copying. Widely accepted as such a product is IBM
TotalStorage Tivoli Storage Manager (TSM).
16.5 Data migration across ESSs
The information in the previous sections is oriented to the data migration scenarios where
data from non-ESS storage subsystems is moved (or copied) to ESS storage subsystems.
If your data migration has to do with moving data from one ESS to another ESS you still can
use the techniques described so far. But for this ESS to ESS migration scenarios you also
have the additional alternative of using the ESS Copy Services functions. More precisely,
you can take advantage of the optimized ESTABLISH and RESYNCH operations over PPRC
pairs of logical volumes at long distances using the INRANGE Fibre Channel Directors. See
“INRANGE” on page 22.
16.6 IBM migration services
This is the easiest way to migrate data, because IBM will assist you throughout the complete
migration process. In several countries IBM offers a migration service. Check with your IBM
sales representative about migration services for your specific environment and needs.

274

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment

© Copyright IBM Corp. 2002
275
Chapter 17.
Managing and monitoring
the ESS
In this chapter we examine some of the tools that are available for managing and monitoring
the ESS in the open systems environment, on the host operating system level. Each
operating system has some form of storage management capability. In addition to this, you
can use the IBM TotalStorage family of products.
17

276

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
17.1 IBM StorWatch Enterprise Storage Server Expert
The IBM StorWatch Enterprise Storage Server Expert (ESS Expert) is a software tool
designed to help storage administrators manage and monitor IBM storage servers. It is an
optional product that the storage administrator has for managing and monitoring the
performance, the capacity, and the components of the storage environment.
This tool is described in Chapter 6, “IBM StorWatch Enterprise Storage Server Expert” on
page 87 of this redbook. To know how to use the ESS Expert you can consult the IBM
redbook
IBM StorWatch Enterprise Storage Server Expert Hands-on Usage Guide,
SG24-6102.
17.2 General considerations
When planning for an efficient administration of your storage environment, and thinking in
terms of management and monitoring tools and procedures you will need, you should ask
yourself the following questions:

What tools are available to help you in the administration of your system?

Do you have access to any Web site tools and documentation?

Do you have access to software technical support?
We suggest that after you read this chapter, you will be able to address these considerations.
The following sections in this chapter should help you start to formulate ideas to resolve these
proposed questions.
17.3 Managing and monitoring tools
Many open system platforms have software tools designed to help system and storage
administrators to manage and monitor disk storage devices. It is beyond the scope of this text
to go into detail about the various tools and commands available on each open system
platform. We will only examine a few.
17.3.1 AIX
System Management Interface Tool (SMIT)
Typing in the commands
smit
on a Graphical User Interface (GUI) or
smitty
on a non GUI,
present excellent tools you can use to control adapters, disks, and arrays. Once you have
launched the file, you can press F1 for help using
smit,
see

Figure 17-1. For additional
on-line help using
smit
, refer to chapter 7 of the Redbook
IBM Certification Study Guide
pSeries AIX System Support
, SG24-6199.

Chapter 17. Managing and monitoring the ESS
277
Figure 17-1 AIX smitty panel
Being familiar with the Logical Volume Manager (LVM) will be of great benefit to you. Among,
many things LVM allows you to mirror, stripe, partition, create, grow, and delete volume
groups and file systems on the disks.
With the
diag
command you can invoke a standard AIX diagnostics menu. This menu lists
these aids that allow you to manage external disk operations.
A variety of scripts exist that have been developed for disk management.
17.3.2 HP-UX and Sun Solaris
HP and Sun can utilize SAM and Veritas Volume Manager. Solitus Disk Suite is only available
for Solaris. These tools are and additional purchase and are used separately from the
operating system management functions. These tool allows disk and data management
functions. Figure 17-2 shows an example of the VERITAS Volume Manager GUI tool, that
allows the administrator to manage the volumes and disk groups. This tool enables you to
create and delete filesystems. Set sizes and permissions. Manage and monitor the ESS disks
presented to the host on the host operating system level.
Figure 17-2 Veritas Volume Manager GUI

278

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
17.3.3 OS/400
There are several tools and commands on iSeries you can use for managing and monitoring
disks. The Dedicated Service Tools (DST) is part of the service function used to service the
system when the operating system is not running. With this function you can add new disks,
move the disks between auxiliary storage pools, and perform other functions (see
Figure 17-3).
Figure 17-3 DST main menu
System Service Tools (SST)
The SST is part of the service function that can be used when the system is up and running.
When the command STRSST base menu is displayed, you can select option
Work with disk
units
. After that you can select these options:

Display disk configuration

Work with disk configuration

Work with disk unit recovery
Operations Navigator
The Operations Navigator provides a Windows interface to common iSeries management
functions. Operations Navigator is supported on Windows 95, Windows 98, Windows NT
4.0.and Windows 2000.
Operations Navigator can be used to configure and manage ASPs comprised of both internal
and external (ESS) disk units. However, as of V5R1, Operations Navigator provides a
graphical display of only its internal disk units, and not of ESS hardware LUNs.
Other common management functions include Basic Operations, TCP/IP Configuration, Job
Management, Users and Groups, and Database Management. Operators configure AS/400
systems to be managed. Operations Navigator will connect directly to the AS/400 from the
Windows workstation using TCP/IP to provide management of that single AS/400 system.
Management Central is an extension of Operations Navigator that is accessed through a
toolbar button on the Main Operations Navigator window. Management Central allows
operators to monitor multiple AS/400 systems with the same ease that Operations Navigator
allows them to perform other basic management functions on a single system.

Use Dedicated Service Tools (DST)
Select one of the following:
>1. Perform an IPL
>2. Install the operating system
>3. Work with Licensed Internal Code
>4. Work with disk units
>5. Work with DST environment
>6. Select DST console mode
>7. Start a service tool
>8. Perform automatic installation of the operating system
>9. Work with save storage and restore storage
10. Work with remote service support
11. Work with system partitions
Selection
4
F3=Exit F12=Cancel

Chapter 17. Managing and monitoring the ESS
279

You can define, for example, one monitor to watch several metrics that correspond to
response time, and another monitor could keep track of communications lines. Metrics
include:
– CPU utilization
– Interactive response time
– Transaction rate
– Batch logical database I/O
– Disk arm utilization
– Disk storage
– Disk I/O processor (IOP) utilization
– Communications IOP utilization
– Machine pool faults
– User pool faults
– Communications line utilization
– LAN utilization
With the Operations Navigator installed and configured, you can refer to the Operations
Navigator online help for specific task-based information. Also you can visit the iSeries Info
Center Web site by clicking this option at:
http://publib.boulder.ibm.com/pubs/html/as400/v5r1/ic2924/index.htm
At the Info Center you will find many examples on how to exploit the functions that Operations
navigator provides for managing and monitoring.
Once you have the basics down, you can explore the advanced features. Some of these
include: changing the collection interval to collect metric data less frequent, ignoring data
spikes by extending the threshold duration, setting threshold triggers and reset values, and
performing automation.
After you have your monitors defined, you simply press the start button and select one or
more system groups or individual endpoint systems to run your monitor. The monitors window
will give you up-to-date status information for each of your defined monitors. You will be able
to see which monitors have started, any monitored systems that have failed, and which
monitors have reached a threshold. A monitor may run for minutes, hours, days, or until you
stop it. Even though a monitor is running, you are free to perform other functions in
Management Central, Operations Navigator, or on your PC. You may also quit your
Operations Navigator session or even shut down your PC and your monitors will continue to
run. The next time you enter Management Central you can check the status of the monitors
you are running.
Monitor graph
When you open a monitor that you started, you will see the monitor graph window. The
monitor graph window shows a separate real-time data collection graph for each metric
defined in the monitor. If the monitor is collecting metric data from multiple endpoint systems
or system groups, all systems will be displayed in each metric graph.
Events and automation
If you defined your monitor to watch for thresholds, you can also set actions for when the
threshold is triggered and reset. Actions on the PC include: opening the graph window,
opening the event log window, sounding an alarm, and performing automation. Actions on the
AS/400 include logging an event and performing automation.

280

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
If log an event is set, you will be able to see these in the event log. You can filter the event log
to see all events for a particular system or group, for a particular metric, or for all the events
associated with a specific monitor. Each event contains helpful details like the monitor
threshold value, the actual metric data collected value, system that sent the event, and
whether any automation occurred.
17.3.4 Windows Disk Administrator
The Disk Administrator is a graphical tool for managing disks in a Windows NT environment.
This tool encompasses and extends the functionality of character-based disk management
tools.
With the disk administrator you can do the following:

Create and delete partitions on a hard disk and logical drives within an extended partition.

Format and label volumes.

Read status information about disks such as the partition sizes and the amount of free
space that is available for creating additional partitions.

Read status information about Windows NT volumes such as the drive-letter assignment,
volume label, file system type, and size.

Make and change drive-letter assignments for hard disk volumes.

Create and delete volume sets.

Extend volumes and volume sets.

Create and delete stripe sets.
For additional information, please refer to your Windows Disk Administrators users guide.
17.3.5 Linux
With the Linux source code being freely available, several companies have developed
different distributions of Linux. A distribution is a complete system, the key component of
which is the Linux kernel. Other utilities, services, and various applications can be included as
well, depending on the distribution and the intended use. You may want to check the tools
available with the distribution you are using.
Bonnie
Bonnie is a performance measurement tool written by Tim Bray. Bonnie performs a series of
tests on a file of selectable size (100MB default). Read and write access of different kinds are
performed during the verifications. For each, the processed bytes per second and the CPU
usage are measured. If you are interested in Bonnie, you will find more details on:
http://www.textuality.com/bonnie/advice.html
17.3.6 NUMA-Q
ConfigApp, Advanced Detection Availability Manager (ADAM) and VERITAS Volume
Manager provide disk and data management tools.
Visit the following Web sites for more details on NUMA-Q:
http://techdocs.beaverton.ibm.com/
http://techdocs.beaverton.ibm.com/docs/ncrnab00/ch_5.htm#SE178140

Chapter 17. Managing and monitoring the ESS
281
The Advanced Detection Availability Manager (ADAM) subsystem provides NUMACenter
systems management. ADAM is composed of the ADAM Server, the ADAM hardware, the
ADAM software stack, and the ADAM Console. ADAM provides device access, management
of servers and SAN elements, single point of access, resource monitoring, reconfiguration of
components regardless of vendor, error and event consolidation, auto discovery of devices,
and phone-out capability. For additional information on installation and use of ADAM, visit the
Web at:
http://techdocs.beaverton.ibm.com/docs/ncagaa07/ncagaa07.htm

282

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment

© Copyright IBM Corp. 2002
283
Chapter 18.
ESS Copy Services for open
systems
PPRC and FlashCopy are ESS functions used for disaster recovery, data backup, creation of
test data, and for data migration. In this chapter we examine some considerations that will
help you when planning for the implementation of these functions in your open systems
environment. For detailed information and recommendations on how to implement and use
this powerful ESS functions you must refer to the IBM redbook
Implementing ESS Copy
Services on UNIX and Windows NT/2000,
SG24-5757.
18

284

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
18.1 ESS Copy Services: considerations
ESS Copy Services provides a Web-based interface for setting up and managing
Peer-to-Peer Remote Copy (PPRC) and FlashCopy. PPRC operates at the volume level from
one SCSI or Fibre Channel bus target/LUN group (LSS), to the volume level within another
LSS. FlashCopy also operates at the volume level, however the source and target volumes
are within the same SCSI or Fibre Channel bus target/LUN group (LSS). CLI is provided to
allow these functions to execute on supported host systems at the command line level. For a
current list of CLI supported hosts please refer to Table 13-1 on page 207.
Before you can use the PPRC or FlashCopy functions you must have the appropriate feature
codes installed on the ESS:

The FlashCopy license feature ordered must be equal to or greater than the total capacity
of the ESS

The PPRC license feature must be equal to or greater than the total capacity of the ESS.
The PPRC feature must also be purchased and installed on both the primary and
secondary ESS
You should consider the following when planning:

Only one FlashCopy at a time can be active on a volume, however, you can perform a
PPRC concurrently with FlashCopy on the same volume. The designated FlashCopy
target volume cannot be a primary volume in a PPRC volume pair.

The primary and secondary volumes must reside within an ESS, for both PPRC and
FlashCopy. You cannot FlashCopy from an ESS to different external device type.

You need to manage PPRC using the ESS Copy Services Web interface, therefore,
Ethernet, and TCP/IP connectivity is needed between the two participating ESS
subsystems, primary and secondary, and the Web browser initiating and managing the
PPRC activities.

The source and target logical volumes must be the same, or larger in size.

If you want to use PPRC or FlashCopy to do a hardware copy of the disks attached to the
iSeries, you must mirror the LSU from the internal drive into the ESS to ensure the whole
single level storage is copied.

There is a maximum of 2000 volumes allowed for a ESS Copy Services server. This
number includes all the primary and secondary PPRC volumes plus all the source and
target FlashCopy volumes.
18.2 Preliminary setup
In order to use ESS Copy Services you must configure one ESS to be the primary ESS Copy
Services server. All information related to the copy services is stored in this ESS, such as
volumes and their state, ESCON connectivity between ESS, and much more.
On each ESS that is configured to use ESS Copy Services there is a client running.
Whenever the copy services configuration changes this client notifies the ESS Copy Services
server of the changes.
Optionally, there could be one ESS defined as the backup ESS Copy Services server. In case
the primary ESS Copy Services server is lost, the backup server could be used for controlling
the ESS Copy Services. Once the primary ESS is up again, the backup server will notify all
clients, so the clients can be switched back to the primary ESS Copy Services server.

Chapter 18. ESS Copy Services for open systems
285
We recommend that you have the primary and backup ESS Copy Services servers on
different sites if your ESS Storage Network spans multiple locations. In case the entire side
with the primary ESS Copy Services server is going down for whatever reason, the backup
ESS Copy Services server can keep copy services alive. You may also consider to run the
primary ESS Copy Services server on the remote site ESS cluster.
The information on the primary and backup ESS Copy Services server has to be specified on
each cluster of all the Enterprise Storage Servers that are going to be used for ESS Copy
Services. We recommend that you specify a backup for ESS Copy Services whenever
possible. This task is done by an IBM SSR with the Master Console.
18.3 PPRC
Peer-to-Peer Remote Copy (PPRC) provides a synchronous copy, mirroring (RAID 1) of a
source LUN to a secondary LUN. Data updates to the source device are also sent to the
secondary device. The host is notified that the update has been completed when both
primary and secondary ESS have the update safely in both NVS and cache.
PPRC will operates only over ESCON connections, and therefore, even if it is being used in
an open systems environment, ESCON adapters and cabling are required.
PPRC is supported at distances over 103 km but requires IBM approval, which can be
requested by submitting a Request for Price Quotation (RPQ).
ESTABLISH/RESYNC with INRANGE allows for new implementations:
Enhancements to the INRANGE 9801 SNS exploit the PPRC Establish Copy design, allowing
data to stream across the network without unnecessary intermediate messages (protocols)
for data transmission over long distances. These enhancement enable PPRC solutions with
the INRANGE 9801 for data movement and migration; remote backup (in combination with
FlashCopy); and long distance disaster recovery in two ways: sending log files, or
implementing RESYNCH/SUSPEND loops (if the primary site can tolerate application
quiesce).
18.3.1 Rules for configuring PPRC links
In this section we provide rules for configuring PPRC links.
Establishing ESCON connections between two ESS subsystems
These are some requirements to observe:

A primary LSS can be connected via ESCON links to up to four secondary LSSs.

A secondary LSS can be connected to any number of primary LSSs, limited by the
number of ESCON links available.

PPRC links are unidirectional, because the ESCON port at the primary ESS is
reconfigured to act like an ESCON channel in a host S/390 processor. The primary
ESCON port is dedicated to PPRC.

The ESCON protocol has been streamlined with less handshaking and larger frames
transmitted between ESS.
An ESCON PPRC link can be used only to transmit data from the primary storage control to
the secondary. If you want primary and secondary volumes on each of two ESS, you need
ESCON PPRC links in each direction. The number of links needed in each direction depend
on the total write activity to all the primary devices in each ESS.

286

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
18.3.2 Planning for PPRC
PPRC is possible only between Enterprise Storage Servers. Other disk storage units that
support PPRC can also communicate to the same type of unit only.
You need to have the PPRC feature purchased and PPRC-capable microcode activated on all
ESS units that will be used for PPRC.
PPRC operates at a volume level from one LSS to another LSS.That means you need to have
the target volumes available on the secondary ESS, and you need to identify the LSSs where
the primary and secondary volumes are located.
ESCON connections have to be configured between the units. There can be up to eight
ESCON links between the subsystems. A primary ESS can communicate with up to four
secondary ESSs. A secondary ESS can be connected to any number of ESS primary
subsystems.
You will need to purchase ESCON cables and possibly some other equipment, depending on
the distance between the primary and the secondary ESS. However, each of the ESS Copy
Services servers can only control two ESSs. Therefore, to have one primary ESS
communicate with more than one secondary ESS, this will mean that multiple ESS Copy
Services servers will have to be configured.
The ESS units involved in PPRC must be connected with their standard Ethernet and TCP/IP
to the units that are the primary and backup ESS Copy Services servers. All ESS cluster
hostnames, including its own, must be added to the cluster hostname list, the /etc/hosts file,
during installation. This is configured by the IBM SSR during the installation of the PPRC
feature.
A browser for the Web interface has to be installed on the machines that will be used to
control PPRC with the ESS Copy Services Web Interface.
If you plan to use CLI, install the Java Developers Kit (JDK) JDK version 1.1.8 on the
machines that will run the CLI commands.
Host Systems Attachment Guide,
SC26-7296 for
the latest recommended JDK revision.
You need to check that the CRIT=YES - Light or Heavy options have been set correctly in the
ESS VPD by your hardware specialist.
You need to have the hardware engineer check that all the ESS cluster hostnames were
added to the hostname list during installation. Otherwise the cluster hostname will not be
found when defining the ESS Copy Services primary.
When planning your secondary ESS volume layout for PPRC, optimize your disk capacity. It is
important to realize that the capacity needed on the secondary ESS for disaster recovery may
not have to be initially as large as the primary ESS. A disaster recovery plan (DRP) requires
significant investment financially in technology, people, and process. Every company will be
different, but the I/T components of a disaster recovery plan are essentially driven by the
applications and data you require for business continuity, should a disaster occur. Some
applications and data will be more critical than others. An organization will typically require its
core business systems to be available in a short time, whereas less critical systems quite
possibly could be restored over a number of days.

Chapter 18. ESS Copy Services for open systems
287
Bearing in mind that the disk space you need for PPRC secondary volumes is real disk
space, size your secondary ESS based on your critical business requirements, possibly with
some headroom for applications of intermediate importance. Create PPRC pairs for the
critical data so that is copied in real time. Then, if a disaster happens, you will have the core
systems available on the secondary copies. After the initial recovery priorities have been
handled, you can add more disk ranks for the applications of lower importance and restore
them from tape.
18.3.3 PPRC Read from Secondary option
This function allows a host to read from a Secondary PPRC volume, once the PPRC pair has
reached full duplex state. This function is enabled on open system volumes only, and the ESS
LIC level required is 1.5. This option can be enabled when you select the copy options (see
Figure 18-1).
Figure 18-1 ESS Copy Services: Read from Secondary allowed
The following rules apply:

Read from Secondary will not be allowed unless designated in the Establish PPRC pair
command.

Read from Secondary will not be allowed until the pair has transitioned to full duplex state.
The pair must have reached full duplex state in order for Read from Secondary to be
successful. (if the pair suspends after reaching full duplex state, Read from Secondary will
still be allowed).

Read from Secondary permission may be “added” or “removed” by issuing another
Establish Pair command with the desired Read from Secondary setting.

Read from Secondary is only allowed on open volumes. If this request is made on a S/390
volume, the indication will be stored, but S/390 read commands will still be rejected.
Note:
The value of Read from Secondary is that it allows a server attached to a PPRC
secondary device to discover the physical configuration on the PPRC secondary
volumes. This will reduce the amount of time to switch over to a recover site after a
disaster while PPRC is active. You must then terminate PPRC and launch the
application from the secondary site.

288

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
18.3.4 How to invoke PPRC
There are two different methods of using the ESS Copy Services in the open systems
environment:

Using the ESS Copy Services Web interface

Using the Java-based Command Line Interface (CLI)
To compliment the information presented in this section, please refer to the following
publications:
IBM TotalStorage Enterprise Storage Server Copy Services Command-Line
Interface Reference,
SC26-7434 and
Implementing ESS Copy Services on UNIX and
Windows NT/2000,
SG24-5757.
ESS Copy Services Web interface
The ESS Copy Services function that runs in the ESS provides a Web browser interface that
can be used to control the PPRC functions in the open systems environments. You will be
able to establish PPRC pairs in three different ways:

From the
Volumes
panel — based on volumes

From the Storage Servers panel — based on entire logical subsystems

From the Task panel — once a task for PPRC is created
The ESS Copy Services panel shows volumes (LUNs) by their ESS internal serial numbers.
You will first have to find out what is the serial number of the volumes you intend to copy
(source and target) from your operating system, in order to identify the volume on the ESS
Copy Services
Volume
panels.
To invoke PPRC, you need to perform the following tasks from the ESS Copy Services main
menu:
1.With the PATHS panel, establish a logical path between the primary ESS LSS and host
adapter, and the target LSS and its host adapter.
2.Use the VOLUMES panel to find and select the source and then target volume PPRC
pairs and Establish, Suspend, or Terminate the data transfer. You can optionally establish
both paths and volume pairs from this panel.
3.Alternatively, you can select the CONTROL UNIT panel to initiate or remove PPRC
relationships between all volumes on an LSS.
4.You can save previously defined PPRC path and pair definitions as tasks. Using the
TASKS panel, you can select and run the pre-saved set of tasks.
During steps 1, 2 and 3, once you have selected the resources that you are working with, a
right-click on the target resource starts a wizard to guide you through the selection of the
appropriate PPRC functions.
To assist you in removing unrequired resource information from the panels, a FILTER button
is provided to enable display of selected resources. For example, show only:

S/390 or open systems volumes

Source or target volumes

Physical or logical ESS
Each panel also has an INFORMATION button that will display PPRC status and other
general information about the selected LSSs, volumes or paths.
Detailed information regarding each panel described above can be found in the
Web Interface
Users Guide
, SC26-7346.

Chapter 18. ESS Copy Services for open systems
289
Command Line Interface (CLI)
The CLI interface allows administrators to execute Java-based ESS Copy Services
commands from a command line. This command line interface is currently available for the
following operating systems: AIX, Sun Solaris, HP-UX, Windows NT and 2000, Intel-based
Linux (See Section 13.5, “Command Line Interface (CLI)” on page 209 for information on
operating system support).
18.4 FlashCopy
FlashCopy makes a single point in time (T0) copy of a LUN. The target or copy LUN is almost
immediately available after the command has been processed. For detailed information on
how FlashCopy works and for considerations when planning for FlashCopy implementations
we recommend you refer to the publication
Implementing ESS Copy Services on UNIX and
Windows NT/2000,
SG24-5757.
18.4.1 Invoking FlashCopy
There are two different methods of using the ESS Copy Services in the open systems
environment:

Using the ESS Copy Services Web interface:
The ESS Copy Services function
provides a Web browser interface that can be used to control the FlashCopy functions in
the open systems environment. For this situation you will be establishing a FlashCopy pair
using the
Volumes
panel, or using the
Task
panel that the Web browser presents you.
Detailed information regarding the use of each panel can be found in the
Web Interface
Users Guide
, SC26-7346.

Using the Java-based Command Line Interface (CLI):
The CLI interface allows
administrators to execute Java-based ESS Copy Services commands from a command
line. For this situation you will be starting a pre-defined task from the command line or
from your own customized scripts. This command line interface is currently available for
the following operating systems: AIX, Sun Solaris, HP-UX, Windows NT and 2000,
Intel-based Linux (See Section 13.5, “Command Line Interface (CLI)” on page 209 for
information on operating system support).
To compliment the information presented in this section, please refer to the following
publications:
IBM TotalStorage Enterprise Storage Server Copy Services Command-Line
Interface Reference,
SC26-7434 and
Implementing ESS Copy Services on UNIX and
Windows NT/2000,
SG24-5757.
18.4.2 Invoking FlashCopy scenario for AIX
The first step requires you to identify the internal serial number of the logical source and
target disks on the host. For example, the following command is issued on AIX:
lscfg -vl hdisk#| grep Serial
Here is an example of the returned output:
Serial Number...............30115234
The first 3 digits represent the LUN’s hex ID. The last 5 digits represent the serial number of
the ESS. The first 3 digits are what you are looking for to identify the Source LUN serial
number.

290

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
We suggest that you create a LUN mapping table that represents the layout of the ESS, such
as in Figure 15-2 on page 237. This will help you strategically see at a glance the LSSs and
individual LUNs.
To invoke FlashCopy, use the VOLUMES panel to find and select the source and target
volumes. Left-click on the LUN you have identified as the source (the volume being copied
from). Right-click on the LUN that you have identified as the target (the volume being copied
to). Right-click one more time to execute the wizard window.
Once presented with the wizard window, select
Establish FlashCopy pair
or
Withdraw
FlashCopy pair
. Figure 18-2 represents establishing a FlashCopy pair. Click
Next
to go to
the next panel. Figure 18-3 shows the wizard window with No Copy Background selected.
This means that the data will only be accessible to the target LUN as long as the pair remains
established. In the first wizard window, notice on the ESS Master Console, that if the button
appears to be pushed in, then it is selected. See Figure 18-2.
Figure 18-2 ESS Copy Services: First wizard window
From the ESS Copy Services panel, choose:
Volumes -> LSS -> source LUN
(with a left-click)
-> target LUN
(with a right click)
->
(right-click again)
-> Establish FlashCopy pair -> Next ->
choose copy option (see
Figure 18-3)
-> Next ->
name the task (see Figure 18-4), and either
Run
or
Save
the task.
Figure 18-3 Establishing a pair with No Copy background

Chapter 18. ESS Copy Services for open systems
291
Figure 18-4 Wizard window with example task name and description
As with PPRC, you can save your FlashCopy definitions as tasks and run them at any time by
using the
Tasks
panel. You can also
Filter
your displays and use the
Information
button.
Figure 18-4 shows an example task name and description.
Once a relationship has been established between a FlashCopy source and target volume, a
background task commences that copies the entire source volume to the target. Using the
ESS Copy Services Web interface panels, you can suppress this copy task by specifying a
NOCOPY option.
If NOCOPY has been specified, any data about to be updated on the source volume is first
copied to the target volume. Hence, the target volume only contains pre-updated data, not a
complete volume copy. The T0 copy of the source is still available for use as long as the
source target relationship exists.
This relationship must be terminated (WITHDRAW) using the ESS Copy Services Web
interface panels. If NOCOPY was not specified, the relationship ends automatically once the
background source copy has been completed.
We will now summarize these operations.

292

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
To invoke a no-copy background FlashCopy pair
1.Log into ESS Copy Services.
2.From the ESS Copy Services panel, select the
Volumes
panel.
3.Click the desired LSS (from the source
Select Focus here
, drop-down box).
4.Left-click on the source LUN.
5.Right-click on the target LUN.
6.Right-click again.
7.The wizard window will now appear. Select
Establish

FlashCopy pair
.
8.Click
Next
.
9.Select,
Do not perform background copy if checked
.
10.Click
Next
.
11.The wizard window will now appear. Enter in the task name and task description.
12.Click
Save,
to run it at another time, or from the command line on the target, or click
Run
,
to run it immediately.
To invoke a withdraw
For every FlashCopy pair that you establish with the No Copy background option, you must
also create a task to withdraw that pair.
1.Log into ESS Copy Services.
2.From the ESS Copy Services panel, select the
Volumes
panel.
3.Click the desired LSS.
4.Left-click on the source LUN.
5.Right-click on the target LUN.
6.Right-click again.
7.The wizard window will now appear. Choose
Withdraw the Established Pair
.
8.Click
Next
.
9.Enter the task name and task description.
10.Click
Save,
to run it at another time, or from the command line on the target, or click
Run
,
to run it immediately.
For detailed instructions on setting up FlashCopy pairs and withdrawing them, refer to the
IBM publication
IBM TotalStorage Enterprise Storge Server Web Interface User’s Guide,
SC26-7346.

© Copyright IBM Corp. 2002
293
Appendix A.
ESS configuration planning
process
This appendix describes the procedures and support available for the IBM Field Technical
Support Specialist (FTSS) and IBM Business Partners (BP). This supplement is designed to
help FTSSs and BPs to support the customers in the installation planning of the ESS. We
cover the following topics:

Considerations for planning the storage capacity for each host

Where you should start:
– Gathering the information to submit the ESS Configuration worksheets
– Gathering information to complete the Communication Resources worksheets

Where you can find the reference material

Submitting the ESS Configuration worksheet
A

294

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Considerations for planning
You should ask yourself the following questions when planning the total storage capacity of
the ESS:

How many TB of usable disk space do you have? How many disk groups do you have?

What is the total capacity in GB of each disk group. Remember that with the disk drive
intermix, your disk groups will vary in size. You can have 8 packs of 9.1, 18.2, 36.4, and
72.8 GB sizes. Also remember that you can upgrade the installed eight-packs to the 36.4
GB and 72.8 GB higher capacity disk drives.

What size LUNs should you carve out of each disk group to meet the needs of the
applications data requirements?

How much disk space will you need for Flash Copy? A good rule of thumb is double the
amount of disk space that you will be using to flash and designating it for FlashCopy.
Remember that the target LUN in the ESS must be in the same LSS and be of equal or
greater size than the source LUN that you will be flashing.

How should you lay out LUNs for best performance and FlashCopy.

What type of hosts are you connecting to the ESS and how many do you have?

How many host adapter ports will you be connecting to the ESS or SAN Fabric switch?

Will you be using SDD?

Will you be using PPRC or XRC copy services?

Should you install CLI?

What is the order number, feature code, and serial number of the ESS?
Where should you start?
After considering the questions asked previously, you should start gathering information
about each topic.
You should contact your IBM Field Technical Support Specialist (FTSS) or IBM Business
Partner (BP).
Gathering information to submit the ESS Configuration worksheets
You should download the ESS Configuration worksheets from the web at the following site:
http://ssddom02.storage.ibm.com/disk/ess/documentation.html
Click on the
ESS Configuration Planner
and look under Appendix A, B, C and D.
Note:
If your ESS is coming with a mix of drive sizes, then the feature code will not be
valid. You instead need the order number readily accessible. The minimum quantity of
Eight Packs per order is two.

Appendix A. ESS configuration planning process
295
Obtain the order number, feature code and Serial number of the ESS from the Sales
Representative, Project Manager, or Solutions Architect. It would be helpful for you to:
1.Involve the Project office in San Jose to help with the overall ESS Plan.
2.Gather the names of the resources you need to help you configure the ESS.
3.Identify the IBM SSR
4.Identify the storage specialist.
5.Identify the various team members, such as Disk Admin/Sys Admin, SAP/Basis, DBA
members and Network team.
6.Receive the file system requirements.
Completing the Communication Resources worksheets
You should download the Communication Resource worksheets from the Web at the following
site:
http://ssddom02.storage.ibm.com/disk/ess/documentation.html
Click on the
Introduction and Planning Guide
and look under Appendix A.
You should decide on connecting through a private or public network or LAN.
You will probably need to gather the information from all of the team members. You will need
IP addresses, phone numbers, etc. Refer to the worksheets for the information needed.
After completing the Communications worksheets you must E-mail or send them to the IBM
SSR that will be installing the ESS.
Where you can find the reference material
You can find information on all of the above questions and topics by visiting the Web site:
http://ssddom02.storage.ibm.com/disk/ess/documentation.html
Topics and subjects include the following:

ESS overview

ESS spec Sheet

Host systems Attachment Guide

IBM TotalStorage Redbooks

Introduction and Planning Guide

SCSI Command Reference Guide

System/390 and open systems

Users Guide

Web Interface User’s Guide

ESS Copy Services Command-Line Interface User’s Guide

ESS Configuration Planner (Includes a chart with various feature codes and capacities).

Sample storage configuration worksheets.

296

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
Submitting the ESS Configuration worksheets
This is the process for the Americas and it may differ from geography to geography.
Configuration Worksheet Assistance details
The Configuration HelpLine is available to complete the Logical Configuration for the ESS
install. Follow these steps to contact the Configuration HelpLine:
1.Call (800) 237-5511.
2.Select Option #0 (there is a long menu, so try it as soon as you get to this spot).
3.The Agent will ask for the following information:
a.The customer number or order number and serial number of the ESS.
b.The caller's name and telephone number (also be sure to give them the email address
of the SSR who is to receive the completed configuration.)
c.The system or component type: Say “Shark” or “ESS” (these are the “trigger words”).
d.The severity of the problem / question: Reply “Sev 1” if you need immediate response,
otherwise please use “Sev 2”.
e.The topic or nature of the question: Reply “Logical Configuration Help”.
4.The Agent will set up a PMR number and Branch number, assigned to the question
(Important! These may be used later to check on the status of the question).
5.The Agent will forward the request for a call-back to the Toronto Configuration Help Desk.
6.If you can't access the 800 number, you can use the following tieline or outside line
numbers: 8/367-5079, 770-858-5079.
7.After receiving the PMR number, the Advanced Logical Configuration Team member must
fill out the Web Configuration Worksheets at:
http://ssddom01.storage.ibm.com/ess/sirtstor.nsf/Welcome/?OpenForm

You must have a PMR opened with the IBM Support Center in order to proceed with
submitting the worksheets to Toronto.

The worksheets are then converted to a different format for the IBM SSR to use for the
ESS installation.

The completed worksheets are either returned to you or the IBM SSR to do the logical
configuration.

The completed communication resource worksheets are also used by the IBM SSR.

You will need to provide the SSR and Systems Administrator with cabling diagrams for the
host adapter attachments to the ESS.

© Copyright IBM Corp. 2002
297
Related publications
The publications listed in this section are considered particularly suitable for a more detailed
discussion of the topics covered in this redbook.
IBM Redbooks
For information on ordering these publications, see “How to get IBM Redbooks” on page 298.
IBM TotalStorage Enterprise Storage Server,
SG24-5465
Implementing Linux with IBM Disk Storage,
SG24-6261
IBM StorWatch Enterprise Storage Server Expert Hands-on Usage Guide
, SG24-6102
Implementing Fibre Channel Attachment on the ESS,
SG24-6113
ESS Solutions for Open Systems: Compaq Alpha Server, HP, and Sun,
SG24-6119
iSeries in Storage Area Network A Guide to Implementing FC Disk and Tape with iSeries,
SG24-6220
Implementing ESS Copy Services on UNIX and Windows NT/2000,
SG24-5757
Implementing ESS Copy Services on S/390,
SG24-5680
Planning for IBM Remote Copy,
SG24-2595
Referenced Web sites
These Web sites are also relevant as further information sources:

ESS Web site
http://www.storage.ibm.com/hardsoft/products/ess/ess.htm

ESS manuals
http://ssddom02.storage.ibm.com/disk/ess/documentation.html

ESS supported servers
http://www.storage.ibm.com/hardsoft/products/ess/pdf/1012-01.pdf

SDD support:
http://ssddom01.storage.ibm.com/techsup/swtechsup.nsf/support/sddupdates

Host System Attachment Guide
http://www.storage.ibm.com/hardsoft/products/ess/pubs/f2ahs04.pdf

298

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
How to get IBM Redbooks
You can order hardcopy Redbooks, as well as view, download, or search for Redbooks at the
following Web site:
ibm.com/redbooks
You can also download additional materials (code samples or diskette/CD-ROM images) from
that site.
IBM Redbooks collections
Redbooks are also available on CD-ROMs. Click the CD-ROMs button on the Redbooks Web
site

for information about all the CD-ROMs offered, as well as updates and formats.

© Copyright IBM Corp. 2002
299
Special notices
References in this publication to IBM products, programs or services do not imply that IBM
intends to make these available in all countries in which IBM operates. Any reference to an
IBM product, program, or service is not intended to state or imply that only IBM's product,
program, or service may be used. Any functionally equivalent program that does not infringe
any of IBM's intellectual property rights may be used instead of the IBM product, program or
service.
Information in this book was developed in conjunction with use of the equipment specified,
and is limited in application to those specific hardware and software products and levels.
IBM may have patents or pending patent applications covering subject matter in this
document. The furnishing of this document does not give you any license to these patents.
You can send license inquiries, in writing, to the IBM Director of Licensing, IBM Corporation,
North Castle Drive, Armonk, NY 10504-1785.
Licensees of this program who wish to have information about it for the purpose of enabling:
(i) the exchange of information between independently created programs and other programs
(including this one) and (ii) the mutual use of the information which has been exchanged,
should contact IBM Corporation, Dept. 600A, Mail Drop 1329, Somers, NY 10589 USA.
Such information may be available, subject to appropriate terms and conditions, including in
some cases, payment of a fee.
The information contained in this document has not been submitted to any formal IBM test
and is distributed AS IS. The use of this information or the implementation of any of these
techniques is a customer responsibility and depends on the customer's ability to evaluate and
integrate them into the customer's operational environment. While each item may have been
reviewed by IBM for accuracy in a specific situation, there is no guarantee that the same or
similar results will be obtained elsewhere. Customers attempting to adapt these techniques to
their own environments do so at their own risk.
Any pointers in this publication to external Web sites are provided for convenience only and
do not in any manner serve as an endorsement of these Web sites.
The following terms are trademarks of other companies:
C-bus is a trademark of Corollary, Inc. in the United States and/or other countries.
Java and all Java-based trademarks and logos are trademarks or registered trademarks of
Sun Microsystems, Inc. in the United States and/or other countries.
Microsoft, Windows, Windows NT, and the Windows logo are trademarks of Microsoft
Corporation in the United States and/or other countries.
PC Direct is a trademark of Ziff Communications Company in the United States and/or other
countries and is used by IBM Corporation under license.
ActionMedia, LANDesk, MMX, Pentium and ProShare are trademarks of Intel
Corporation in the United States and/or other countries.

300

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
UNIX is a registered trademark in the United States and other countries licensed exclusively
through The Open Group.
SET, SET Secure Electronic Transaction, and the SET Logo are trademarks owned by SET
Secure Electronic Transaction LLC.
Other company, product, and service names may be trademarks or service marks of others.

© Copyright IBM Corp. 2002
301
Index
Numerics
3390 25, 107
6+P+S, 7+P array 15
72.8 GB disk drive 27
8-pack 15, 23
9337 20
A
AC power supplies 61
accessing the ESS Expert 88
Add Volumes (1 of 2) panel 252
Add Volumes (2 of 2) panel 254
Add Volumes panel 134
adding FB volumes 251
adding volumes to an LCU 134
addressing 16
affinity 16
AIX
host setup tasks 212
logical volume manager (LVM) 270
sample data migration procedure 267
SMIT 276
alias management 154
array
6+P+S / 7+P configurations 15, 105
capacity 20, 107
AS/400
data migration 268
host setup tasks 221
logical volumes 20
assigning
Fibre Channel host adapters 261
LUNs 263
SCSI host adapters 260
availability
cache considerations 58
clusters considerations 55
configuration 72
DC power supplies 61
device adapters 59
disk drives 59
ESCON configuration 74
host adapters 54
maintenance strategy 63
power and cooling 60, 77
power interruptions 61
remote copy 76
SDD 73
service actions 65
upgrades 64
B
backup and restore for data migration 269
C
cables 33
cache 5, 28
availability considerations 58
Call Home 38, 51, 68
capacity 19
managing with the ESS Expert 88
options 18
capacity intermix 15, 23
capacity planning 14, 19
carving LUNs 261
CKD 8
custom volumes 107
data capacity 20
logical volumes 21, 106
LSS mapping 103, 106
standard volumes 107
storage configuration 102
CLI 289
clusters 5
availability considerations 55
Ethernet adapters 63
failback 58
failover 57
CNTLUNIT UNIT=2105 146
Command Line Interface, see CLI
command tag queueing 6
Compaq Open VMS and Tru64 4.x
host setup tasks 214
Concurrent Copy 6, 166, 199
configuration for availability 72
configuration rules 15
configuration worksheets 116
Communication Resources worksheet 117, 295
Configure Disk Groups panel 129
Configure Host Adapter Ports panel 247
Configure LCU panel 124
configuring AdvFS 216
connecting to the ESS Specialist 82
Control Unit Initiated Reconfiguration, see CUIR
cooling and airflow 38
cooling fans 62
replacement 67
copy for data migration 269
copy services
considerations 21
CRR 180, 183
CUIR 8
configuration 117
software support 100
using 178
custom logical configuration 118
custom volumes 7, 160
configuration 107
planning for performance 29

302

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
D
daisy chaining 32
data migration
across ESSs 174, 273
IBM services 175, 273
iSeries and AS/400 systems 268
open systems 265
sample procedure for AIX 267
UNIX systems 269
Windows NT and 2000 273
zSeries servers 171
data set considerations for large volumes 162
DC power supplies
availability considerations 61
replacement 66
device adapters 5, 15
availability considerations 59
disk groups 237
device eligibility for dynamic management 156
devices 16
DEVSERV command 72
DEVSERV PATHS command 188
DEVSERV QPAVS command 160, 189
DFSMS
requirements 165
DFSMS Optimizer 180, 183
DFSMSdss 172
DFSMShsm 173
DFSMShsm ABARS 173
disk drives 5
72.8 GB capacity hdd 27, 106
availability considerations 59
capacity 27
capacity intermix 15, 23
hot pluggable 60
replacement 65
upgrade to higher capacity 19, 294
disk groups 8, 103–104
allocation 24
configuration with the ESS Specialist 129
configuring for FB storage 241
configuring for RAID-5 259
considerations 105
deletion 242
relation to device adapters 237
selection during storage allocation 253
DISPLAY MATRIX command 192
documentation 10
dual line cords 34
dump and restore for data migration 269
dynamic alias management
device eligibility 156
enabling 159
implementation 158
dynamic workload management 155
DYNIX/ptx
host setup tasks 233
E
effective data capacity 19
eight-pack 15, 23
electrical power 34
E-mail 67, 69
configuration 51
Emergency Power Off (EPO) switch 61
Emulex adapter 225, 230
enabling dynamic alias management 159
EREP 67, 72, 180, 184
error messages 70
ESCON 8
addressing 16
configuration for availability 74
host adapter 54
host attachment 33
interface 4
logical paths and devices 16, 148
ESCON directors 75
ESS
functions and features 6
high availability features 53
maintenance strategy 63
major components 4
overview 4
scalability 7
ESS Copy Services 6
Concurrent Copy 6, 166
considerations 21, 199, 284
FlashCopy 21
network configurations 45
open systems 283
OS/390 and z/OS support 166
overview 21
PPRC 21
PPRC configuration 200, 285
primary and backup servers 21, 284
TCP/IP configuration 50
VSE/ESA 168
Web interface 169, 200, 286, 288–289
XRC 21, 166
z/VM and VM/ESA 168
zSeries servers 197
ESS Expert 87, 181, 276
initial panels 88
overview 88
ESS Master Console 39, 83
connections 39
differences with ESSNet 42
panels 40
SSR use 63
ESS Specialist 79, 178, 235, 239
Add Volumes (1 of 2) panel 252
Add Volumes (2 of 2) panel 254
Add Volumes panel 134
authorizations 85
Configure Disk Groups panel 129
Configure Host Adapter Ports panel 138, 247
Configure LCU panel 124
configuring disk groups 129

Index
303
configuring PAVs 126
configuring S/390 storage 102
Fixed Block Storage panel 241
messages 120
Modify Host Systems panel 243
Modify PAV Assignments panel 137
Modify Volume Assignment panel 256
open systems 236
overview 80
panels 83
panels for CKD storage allocation 118
prerequisites 81
S/390 configuration process 114
S/390 Storage panel 122
setup 50
step-by-step example 140
Storage Allocation panel 105, 121
task sequence for FB storage allocation 239
ESS Web interface 81, 169, 286, 288
Ethernet client connection 43
ESSNet console 39
Ethernet private LAN 44
F
failover/failback of clusters 57
fans 62
fast write 6
FB logical volumes and Fibre Channel LUNs 239
FB logical volumes and SCSI LUNs 239
FBA 8
adding volumes 251
LSSs 236
FCP
addressing 16
assigning host with the ESS Specialist 260
host adapter 54
host adapter ports configuration 250
host attachment 32
interface 4
LUN 25–26
Fibre Channel protocol, see FCP
Fibre Channel/FICON host adapter 4
FICON 8
addressing 16
attachment example 153
characteristics 147
host adapter 54
host attachment 33, 147
interface 4
logical paths and devices 16, 148
RMF Channel Path reporting 182
software support 98
Fixed Block Architecture, see FBA
Fixed Block Storage panel 241
FlashCopy 6, 21, 166, 289
considerations 199, 284
invoking in AIX 289
invoking in open systems 289
software support 98
TPF 98, 169
zSeries servers 198
flexible configurations 23
floor loading 35, 37
H
Hardware Configuration Definition, see HCD
HCD 106, 146, 148
panels 149
hdisk 7
Host adapter 8
host adapter ports configuration
Fibre Channel ports 250
SCSI ports 247
host adapters 15
availability considerations 54
ESCON 54
FCP 54
Fibre Channel
assigning LUNs 263
Fibre Channel/FICON 4
FICON 54
planning for performance 27
SCSI 54
assigning LUNs 263
host attachments 25, 32
ESCON 33
FCP 32
FICON 33
SCSI 32
host name 245
host setup
AIX 212
HP-UX 219
Linux for Intel-based processors 226
Linux for zSeries 169
Novell NetWare 230
NUMA-Q 233
open systems 211
Sun Solaris 231
TPF 168
VSE/ESA 168
z/OS and OS/390 146
z/VM and VM/ESA 167
host setup tasks
Compaq Open VMS and Tru64 4.x 214
iSeries and AS/400 221
xSeries and Intel based servers 224
host view
SCSI server 17
zSeries server 16
HP-UX
host setup tasks 219
monitoring and managing 277
I
I/O adapters
AIX server 213
I/O Priority Queuing 6
IBM StorWatch Enterprise Storage Server Expert, see

304

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
ESS Expert
IBM TotalStorage Enterprise Storage Server Master Con-
sole, see ESS Master Console
IBM TotalStorage Enterprise Storage Server Specialist,
see ESS Specialist
IDCAMS 180
LISTDATA command 185
RAID Rank Counters report 186
SETCACHE command 187
Subsystem Counters report 185
IEA435I message 194
IEA480E SIM message 194
implementing dynamic alias management 158
implementing large volumes 161
informational messages 69
INRANGE channel extender 22, 285
Intel-based servers
host setup tasks 224
intemixed capacity 15
interleaved 21, 29, 106
invoking FlashCopy 289
invoking FlashCopy example 289
invoking PPRC 200, 288
IOCP 146, 148
IODF 106
iSeries
data migration 268
host setup tasks 221
logical volumes 20
J
JBOD 8, 19, 25
K
kernel configuration in Linux 227
L
large volume support, see LVS
LCU 102
adding volumes with the ESS Specialist 134
address layout 112
number of devices 103
LIC 5, 80
Linux
host setup tasks for Intel based processors 226
host setup tasks for zSeries 169
software support 98
Local TCP/IP Network 45
logical configuration 23
CKD storage 114
configuration worksheets 116
custom configuration 118
FB storage 239
sample procedure 259
step-by-step example 140
Logical Control Units, see LCU
logical device mapping 24
logical paths 16
Logical Subsystem, see LSS
logical volume 21
CKD 106
iSeries and AS/400 20
per array 109
size 20
LSS 8, 16
CKD 106
configuration rules 15
FBA 236
LCU mapping 102
mapping 23, 238
monitoring 185
physical mapping 24
planning for performance 28
LUN 7, 16
assigning to Fibre Channel host adapters 263
assigning to SCSI host adapters 263
assignment in AIX 213
carving 261
SCSI / FCP 26
shared LUNs 73
LVS 7
characteristics 161
data set considerations 162
implementation 161, 164
PPRC, XRC and FlashCopy 164
software support 99
M
Machine reported product data, see MRPD
major components 4
managing tools for S/390 178
mapping
logical device 24
LSS 23, 103
messages
error 70
informational 69
SIM 71
microcode updates 65
migrating data
iSeries environment 268
open systems 266
Windows NT and 2000 273
zSeries servers 171
migration
using AIX LVM mirroring 270
MIH interval 164
model capacity 18
Modify Host Systems panel 243
Modify PAV Assignments panel 137
Modify Volume Assignments panel 256
modifying an LCU 128
MoST 64
MRPD 7
MVS messages and codes 193
MVS system commands 180
DEVSERV PATHS command 188
DEVSERV QPAVS command 189

Index
305
DISPLAY MATRIX command 192
N
navigating through the ESS Expert 89
non-interleaved 21, 29
Novell NetWare
host setup tasks 230
NUMA-Q
host setup tasks 233
management tools 280
NVS 5
O
open systems
data migration 266
host setup 211
managing and monitoring tools 276
open systems storage allocation
sample procedure 259
Operations Navigator for OS/400 278
OS/400 20
monitoring and managing 278
P
pager 67, 71
Parallel Access Volumes, see PAV
PAV 29, 110
alias ratios 111
base and alias device type 146
configuring with the ESS Specialist 126
for z/VM and VM/ESA 167
LCU addressing 103
performance
managing with ESS Expert 88
PAVs 111
planning 26
physical installation
considerations 32
cooling 38
electrical power 34
floor loading 35, 37
planning
capacity 14
data migration in zSeries 172
for the future 26
logical configuration 23
performance 26
physical installation considerations 32
PPRC 286
the configuration 294
planning for availability
cache 58
clusters 55
device adapters 59
disk drives 59
host adapters 54
power and cooling 60
power interruptions 61
remote copy 76
SDD 73
platform support 8
point-in-time copy 21
port logins 16
power and cooling
availability considerations 60, 77
power interruptions 61
PPRC 6, 21, 26, 77, 166
configuration guides 200, 285
considerations 33, 199, 284
for data migration 174, 273
invoking in open systems 288
invoking in S/390 200
large volume support 164
read from a secondary 287
TPF 98, 169
zSeries servers 198
primary and backup servers 21
TCP/IP configuration 50
Q
Qlogic adapter 225, 230
QUERY PAV command 167
R
RAID-5 7, 19, 25, 29, 59, 104
array configuration 105
CKD volumes per array 109
configuring disk groups 259
for availability 60
rank 8, 104
read from secondary 287
Redbooks Web site 298
Contact us xxii
Remote Support Facility 38, 51, 68
RESYNC 22
RMF 180
Device Activity report 181
Workload Activity report 182
S
S/390 200
FlashCopy 198
host view 16
managing tools 178
monitoring tools 180
planning for performance 27
PPRC 198
S/390 Storage panel 122
scalability 7
SCSI
addressing 16
assigning host with the ESS Specialist 259
command tag queueing 6
daisy chaining 32
host adapter 54
host adapter ports configuration 247

306

Implementing the IBM TotalStorage Enterprise Storage Server in Your Environment
host attachment 32
host view 17
interface 4
LUN 25–26
target id 32, 239
SDD 28
selecting a Host Adapter port 251
sequential detect 7
service actions 65
service definitions for WLM 156
SIM messages 67, 71, 117, 180, 184
IEA480E 194
SMF 180, 184
SMIT 276
SNMP 67, 69
software support
CUIR 100
FICON 98
FlashCopy 98
Large Volume Support 99
Linux 98
TPF 97
VSE/ESA 97
z/VM and VM/ESA 96
zSeries servers 95
spare disk 15
SSID 125
SSR 8
standard volumes 107
Standard volumes to auto-allocate 130
Storage Allocation panel 105, 121
storage capacity options 18
storage consolidation 172
submitting the Configuration worksheets 296
Subsystem Device Driver, see SDD
Sun Solaris
host setup tasks 231
monitoring and managing 277
supported servers 10
System Service Tool for OS/400 278
Systems Managed Storage data migration 172
T
target ID 32, 239
terminology 7
tools for monitoring S/390 180
TPF
host setup tasks 168
native FICON support 98
PPRC and FlashCopy 98, 169
TPF software support 97
U
undefining an LCU 128
UNIX
commands for migrating data 269
data migration 269
upgrades
availability considerations 64
using the ESS Expert 88
V
vital product data, see VPD
vpath 7
VPD 7
VSE/ESA
data migration 174
ESS Copy Services 168
host setup tasks 168
software support 97
W
Web client
connection 44
requirements 43
Web interface 200, 286, 288–289
Web sites 9
Welcome panel 84
WLM 29, 111
alias management 154
Workload Manager, see WLM
WWPN 26, 251
defining in the ESS Specialist panel 245
finding in AIX 213
X
XRC 6, 21, 26, 77, 166, 198
data migration 173
xSeries servers
host setup tasks 224
Z
z/OS and OS/390
data migration 172
HCD and IOCP 146
host setup tasks 146
z/VM and VM/ESA
data migration 174
ESS Copy Services 168
host setup tasks 167
PAV implementation 167
software support 96
zSeries servers
data migration 171
ESS monitoring and management 177
host setup tasks 145
software support 95

To determine the spine width of a book, you divide the paper PPI into the number of pages in the book. An example is a 250 page book
using Plainfield opaque 50# smooth which has a PPI of 526. Divided 250 by 526 which equals a spine width of .4752". In this case, you
would use the .5” spine. Now select the Spine width for the book and hide the others: Special>Conditional
Text>Show/Hide>SpineSize(-->Hide:)>Set
(0.5” spine)
0.475”<->0.873”
250 <-> 459 pages
IBM TotalStorage Enterprise Storage Server: Implementing the ESS in Your Environment

To determine the spine width of a book, you divide the paper PPI into the number of pages in the book. An example is a 250 page book
using Plainfield opaque 50# smooth which has a PPI of 526. Divided 250 by 526 which equals a spine width of .4752". In this case, you
would use the .5” spine. Now select the Spine width for the book and hide the others: Special>Conditional
Text>Show/Hide>SpineSize(-->Hide:)>Set


®
SG24-5420-01 ISBN 0738424390
INTERNATIONAL
TECHNICAL
SUPPORT
ORGANIZATION
BUILDING TECHNICAL
INFORMATION BASED ON
PRACTICAL EXPERIENCE
IBM Redbooks are developed
by the IBM International
Technical Support
Organization. Experts from
IBM, Customers and Partners
from around the world create
timely technical information
based on realistic scenarios.
Specific recommendations
are provided to help you
implement IT solutions more
effectively in your
environment.
For more information:
ibm.com/redbooks
IBM TotalStorage
Enterprise Storage Server
Implementing the ESS in Your Environment
Learn how to install,
tailor, and configure
the ESS
Plan for migrating
data and changing
configurations
Know the tools for
monitoring and
managing the ESS
Since the initial availability of the IBM TotalStorage Enterprise
Storage Server (ESS) there have been many modifications and
additional functionality. The original ESS Implementation Guide
was written in November 1999. The 2002 version is an updated
guide that includes all of the changes since then.
This IBM Redbook is a guide for the installation, implementation,
and administration activities of the ESS in both the S/390 and
open systems environments. It will help you plan and accomplish
the installation, tailoring, and configuration of the ESS in your
environment. It explains how you can use the functions available
for the ESS to efficiently manage your disk storage data as well
as the ESS, once it is operative.
We cover the latest announcements on the ESS: disk capacity
intermix; 72.8 GB capacity disk drive; flexible configurations;
Control Unit Initiated Reconfiguration (CUIR) support; large
volume support (LVS); read from secondary; ESS Master Console;
Subsystem Device Driver (SDD) and the Command Line Interface
(CLI) support for additional operating systems; INRANGE Channel
Extender support; TPF support for PPRC and FlashCopy.
We also provide information on the new models F10 and F20;
FICON native host attachment; new Fibre Channel/FICON host
adapters (short wave and long wave); Linux support for
Intel-based servers and zSeries servers; iSeries and AS/400
support for copy services; and new cache options.
Back cover
First page image
Due to scheduled site updates, ip.com and all related websites (including Prior Art publishing as well as Pantros IP) will be unavailable on Sunday, September 21st, starting at 2:00PM EDT and lasting for up to 4 hours.

We thank you for your patience and apologize in advance for any inconvenience this may cause.
We are pleased to offer a download of this document free of charge.
Files available for download:
  • a representative PDF of the primary file (contains all the relevant information for most users)
To obtain the file, please enter the "captcha" below and click the Download button.
Avoid entering CAPTCHAs! Sign In or Create a Free Account.

Challenge image
  • Please enter letters and numbers only; no spaces.
  • Cannot read this one? Click the image.
  • Difficulty with captchas? Contact us with the URL of this page and we will email it to you.