Planning and Installing the IBM eServer X3 Architecture Servers

An IBM Redbook Publication
IBM Redbook Form Number: SG24-6797-00
ISBN: 0738492620
ISBN: 9780738492629
Publication Date: 04-Jan-2006
Last Update Date: 14-Mar-2006
Find Similar Download

Related People

David Watts - Author [+3] [-3]
Sergey Gorbas - Author
Christian Schroeder - Author
Joshua Young - Author

Abstract

The IBM eServer X3 Architecture servers are the new third-generation Enterprise X-Architecture servers from IBM. The xSeries 366 was announced in February 2005, the xSeries 460 was announced in May, and the xSeries 260 was announced in August. These X3 Architecture servers are ideal for random commercial workloads with high processor and memory bandwidth requirements.

Although aimed at different client sets, they share common components and options. The x460 is targeted at clients who need to implement a large single image "scale-up" configuration of up to 32 processors and 512 GB of RAM. The x366 is targeted at the high-performance environment where rack space is a premium. The x260 is for clients who need high performance but also require large amounts of internal disk storage.

This IBM Redbooks publication provides a detailed technical description of the three servers and explains how to plan, install, configure, and manage these high-performance servers running 32-bit and 64-bit versions of Windows Server 2003, Red Hat Enterprise Linux, SUSE Linux Enterprise Server, and VMware ESX Server.

Language

English

Table of Content

Chapter 1. Technical overview
Chapter 2. Positioning
Chapter 3. Hardware planning
Chapter 4. Operating system installation
Chapter 5. Management
IBM
ibm.com/redbooks
Planning and Installing
the IBM Eserver X3
Architecture Servers
David Watts
Sergey Gorbas
Christian Schroeder
Joshua Young
Covers the xSeries 460, xSeries 366,
and xSeries 260
Learn the technical details of
these high-performance servers
How to configure, install,
and manage the servers
Front cover


Planning and Installing the IBM Eserver X3
Architecture Servers
January 2006
International Technical Support Organization
SG24-6797-00

© Copyright International Business Machines Corporation 2006. All rights reserved.
Note to U.S. Government Users Restricted Rights -- Use, duplication or disclosure restricted by GSA ADP
Schedule Contract with IBM Corp.
First Edition (January 2006)
This edition applies to the following servers:
IBM Eserver xSeries 460, machine type 8872
IBM Eserver xSeries MXE-460, machine type 8874
IBM Eserver xSeries 366, machine type 8863
IBM Eserver xSeries 260, machine type 8865
Note: Before using this information and the product it supports, read the information in
“Notices” on page vii.

© Copyright IBM Corp. 2006. All rights reserved.
iii
Contents
Notices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Trademarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .viii
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .ix
The team that wrote this redbook. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .ix
Become a published author . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xi
Comments welcome. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xi
Chapter 1. Technical overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 The X3 Architecture servers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Key features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Multi-node capabilities and partitioning. . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 IBM XA-64e third generation chipset. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.5 Scalable implementation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.6 Current models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.6.1 x460 and MXE-460 models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.6.2 x366 models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.6.3 x260 models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.7 Front and rear layouts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.7.1 x460 and MXE-460. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.7.2 x366 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.7.3 x260 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.8 Processors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.8.1 Models with dual-core processors. . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.8.2 667 MHz front-side bus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.8.3 Hyper-Threading. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.8.4 Intel 64-bit Extension Memory Technology (EM64T) . . . . . . . . . . . . 23
1.8.5 The benefit of 64-bit computing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.9 XceL4v cache . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
1.10 Memory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
1.11 Serial Attached SCSI. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
1.12 PCI-X slots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
1.13 Remote Supervisor Adapter II SlimLine . . . . . . . . . . . . . . . . . . . . . . . . . 36
1.14 Baseboard Management Controller . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
1.15 Light path diagnostics and PFA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Chapter 2. Positioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.1 Positioning the servers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.1.1 x366 versus x360 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

iv
Planning and Installing the IBM Eserver X3 Architecture Servers
2.1.2 x460 versus x445 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.1.3 x260 versus x255 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.2 Focus market segments and target applications. . . . . . . . . . . . . . . . . . . . 52
2.2.1 x260 target applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
2.2.2 x366 target applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.2.3 x460 target applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.3 Server consolidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
2.3.1 Why consolidate servers?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
2.3.2 Types of server consolidation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
2.3.3 Benefits of server consolidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
2.4 XpandOnDemand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
2.5 Positioning the x460 and the MXE-460. . . . . . . . . . . . . . . . . . . . . . . . . . . 66
2.6 Application scalability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
2.6.1 Enterprise application scaling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
2.6.2 SAP product offering. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
2.6.3 Oracle/J.D. Edwards product offering. . . . . . . . . . . . . . . . . . . . . . . . 73
2.6.4 Siebel product offering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
2.6.5 Oracle/PeopleSoft product offering. . . . . . . . . . . . . . . . . . . . . . . . . . 77
2.6.6 Microsoft SQL product offering. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
2.6.7 DB2 Universal Database. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
2.7 Scale-up versus scale-out. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
Chapter 3. Hardware planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
3.1 Processor subsystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
3.1.1 Dual-core upgrades. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
3.1.2 Processor configuration options . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
3.2 Memory subsystem. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
3.2.1 Memory mirroring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
3.2.2 Hot-swap memory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
3.2.3 Hot-add memory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
3.2.4 Memory ProteXion: redundant bit steering . . . . . . . . . . . . . . . . . . . 101
3.2.5 Memory configuration in BIOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
3.3 Multi-node configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
3.3.1 Cabling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
3.3.2 Scalable system setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
3.3.3 Partitioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
3.3.4 Export regulations for 32-core complexes. . . . . . . . . . . . . . . . . . . . 118
3.4 RSA II SL and BMC initial setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
3.4.1 RSA II SlimLine setup. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
3.4.2 BMC IP address setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
3.5 Storage options. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
3.5.1 Serial Attached SCSI (SAS) subsystem . . . . . . . . . . . . . . . . . . . . . 122
3.5.2 ServeRAID-8i . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

Contents
v
3.5.3 Internal storage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
3.5.4 External storage options. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
3.6 Power considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
3.6.1 Working in 110V environments. . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
3.7 Performance tuning and optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
3.7.1 Optimal memory module installation. . . . . . . . . . . . . . . . . . . . . . . . 129
3.7.2 Memory settings in BIOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
3.7.3 CPU settings in BIOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
3.7.4 PCI adapter placement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
3.8 ServerProven . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
3.9 Solution Assurance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
3.9.1 Solutions Assurance Reviews. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Chapter 4. Operating system installation. . . . . . . . . . . . . . . . . . . . . . . . . 135
4.1 Supported operating systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
4.1.1 Operating systems scalability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
4.1.2 Hyper-Threading support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
4.2 Updating BIOS and firmware. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
4.3 Available memory in multi-node configurations. . . . . . . . . . . . . . . . . . . . 141
4.4 Microsoft Windows Server 2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
4.4.1 Installing Windows Server 2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
4.5 Datacenter offerings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
4.6 Microsoft Windows 2000 Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
4.7 Red Hat and SUSE LINUX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
4.7.1 Additional information for SLES 9. . . . . . . . . . . . . . . . . . . . . . . . . . 155
4.8 VMware ESX Server 2.5.1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
4.8.1 Support for applications running on ESX Server. . . . . . . . . . . . . . . 157
4.8.2 Pre-install information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
4.8.3 Installing ESX Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Chapter 5. Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
5.1 IBM Director . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
5.2 Remote Supervisor Adapter II SlimLine . . . . . . . . . . . . . . . . . . . . . . . . . 164
5.2.1 Installing the RSA II SlimLine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
5.2.2 Connectivity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
5.2.3 Network settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
5.2.4 Web interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
5.2.5 Updating firmware. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
5.2.6 Installing the device driver. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
5.2.7 Remote console and media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
5.2.8 TCP/UDP ports used by the RSA II SlimLine . . . . . . . . . . . . . . . . . 176
5.2.9 MIB files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
5.3 Baseboard Management Controller . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177

vi
Planning and Installing the IBM Eserver X3 Architecture Servers
5.3.1 Connectivity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
5.3.2 Updating the BMC firmware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
5.3.3 Configuring the BMC in BIOS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
5.3.4 Event Log . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
5.3.5 Remote control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
5.3.6 Installing the BMC device drivers . . . . . . . . . . . . . . . . . . . . . . . . . . 182
5.3.7 Ports used by the BMC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
5.4 Integrating the service processors with IBM Director . . . . . . . . . . . . . . . 187
5.4.1 Adding the service processor to IBM Director. . . . . . . . . . . . . . . . . 187
5.4.2 Adding users. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
5.5 OSA SMBridge utility. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
5.5.1 Configuring BIOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
5.5.2 Installation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
5.5.3 Connecting via the telnet server . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
5.5.4 Configuring Windows Server 2003 to support SOL . . . . . . . . . . . . 206
5.5.5 Configuring Red Hat Linux to support SOL. . . . . . . . . . . . . . . . . . . 212
5.5.6 Configuring SUSE LINUX to support SOL . . . . . . . . . . . . . . . . . . . 215
5.5.7 Connecting via the command-line interface . . . . . . . . . . . . . . . . . . 217
5.6 Predictive Failure Analysis (PFA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
5.7 IBM Dynamic System Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
5.7.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
5.7.2 Usage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
5.8 Partition management. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
Abbreviations and acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
Related publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
IBM Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
Other publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
xSeries 460 and MXE-460 publications . . . . . . . . . . . . . . . . . . . . . . . . . . 230
xSeries 366 publications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
xSeries 260 publications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
Online resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
IBM Web pages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
Intel Web pages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
Microsoft Web pages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
Others. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
Help from IBM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237

© Copyright IBM Corp. 2006. All rights reserved.
vii
Notices
This information was developed for products and services offered in the U.S.A.
IBM may not offer the products, services, or features discussed in this document in other countries. Consult
your local IBM representative for information on the products and services currently available in your area.
Any reference to an IBM product, program, or service is not intended to state or imply that only that IBM
product, program, or service may be used. Any functionally equivalent product, program, or service that
does not infringe any IBM intellectual property right may be used instead. However, it is the user's
responsibility to evaluate and verify the operation of any non-IBM product, program, or service.
IBM may have patents or pending patent applications covering subject matter described 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:
IBM Director of Licensing, IBM Corporation, North Castle Drive Armonk, NY 10504-1785 U.S.A.
The following paragraph does not apply to the United Kingdom or any other country where such provisions
are inconsistent with local law: INTERNATIONAL BUSINESS MACHINES CORPORATION PROVIDES
THIS PUBLICATION "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED,
INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF NON-INFRINGEMENT,
MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Some states do not allow disclaimer
of express or implied warranties in certain transactions, therefore, this statement may not apply to you.
This information could include technical inaccuracies or typographical errors. Changes are periodically made
to the information herein; these changes will be incorporated in new editions of the publication. IBM may
make improvements and/or changes in the product(s) and/or the program(s) described in this publication at
any time without notice.
Any references in this information to non-IBM Web sites are provided for convenience only and do not in any
manner serve as an endorsement of those Web sites. The materials at those Web sites are not part of the
materials for this IBM product and use of those Web sites is at your own risk.
IBM may use or distribute any of the information you supply in any way it believes appropriate without
incurring any obligation to you.
Information concerning non-IBM products was obtained from the suppliers of those products, their published
announcements or other publicly available sources. IBM has not tested those products and cannot confirm
the accuracy of performance, compatibility or any other claims related to non-IBM products. Questions on
the capabilities of non-IBM products should be addressed to the suppliers of those products.
This information contains examples of data and reports used in daily business operations. To illustrate them
as completely as possible, the examples include the names of individuals, companies, brands, and products.
All of these names are fictitious and any similarity to the names and addresses used by an actual business
enterprise is entirely coincidental.
COPYRIGHT LICENSE:
This information contains sample application programs in source language, which illustrates programming
techniques on various operating platforms. You may copy, modify, and distribute these sample programs in
any form without payment to IBM, for the purposes of developing, using, marketing or distributing application
programs conforming to the application programming interface for the operating platform for which the
sample programs are written. These examples have not been thoroughly tested under all conditions. IBM,
therefore, cannot guarantee or imply reliability, serviceability, or function of these programs. You may copy,
modify, and distribute these sample programs in any form without payment to IBM for the purposes of
developing, using, marketing, or distributing application programs conforming to IBM's application
programming interfaces.

viii
Planning and Installing the IBM Eserver X3 Architecture Servers
Trademarks
The following terms are trademarks of the International Business Machines Corporation in the United States,
other countries, or both:
AIX®
BladeCenter®
Chipkill™
DB2 Universal Database™
DB2®
DFS™
Eserver®
Eserver®
i5/OS®
Informix®
IBM®
iSeries™
Lotus®
Netfinity®
NetView®
PartnerWorld®
PowerPC®
Predictive Failure Analysis®
pSeries®
Redbooks™
Redbooks (logo) ™
ServerGuide™
ServerProven®
ServeRAID™
ServicePac®
Tivoli Enterprise™
Tivoli®
TotalStorage®
WebSphere®
X-Architecture™
xSeries®
zSeries®
The following terms are trademarks of other companies:
Opteron, AMD Opteron, AMD, the AMD Arrow logo, and combinations thereof, are trademarks of Advanced
Micro Devices, Inc.
IPX, Java, J2EE, and all Java-based trademarks are trademarks of Sun Microsystems, Inc. in the United
States, other countries, or both.
Axapta, Microsoft, Windows NT, Windows, and the Windows logo are trademarks of Microsoft Corporation in
the United States, other countries, or both.
i386, Intel, Itanium, Pentium, Xeon, Intel logo, Intel Inside logo, and Intel Centrino logo are trademarks or
registered trademarks of Intel Corporation or its subsidiaries in the United States, other countries, or both.
UNIX is a registered trademark of The Open Group in the United States and other countries.
Linux is a trademark of Linus Torvalds in the United States, other countries, or both.
Other company, product, and service names may be trademarks or service marks of others.

© Copyright IBM Corp. 2006. All rights reserved.
ix
Preface
The IBM Eserver® X3 Architecture servers are the new third-generation
Enterprise X-Architecture™ servers from IBM®. The xSeries® 366 was
announced in February 2005, the xSeries 460 was announced in May, and the
xSeries 260 was announced in August. These X3 Architecture servers are ideal
for random commercial workloads with high processor and memory bandwidth
requirements.
Although aimed at different client sets, they share common components and
options. The x460 is targeted at clients who need to implement a large single
image “scale-up” configuration of up to 32 processors and 512 GB of RAM. The
x366 is targeted at the high-performance environment where rack space is a
premium. The x260 is for clients who need high performance but also require
large amounts of internal disk storage.
This IBM Redbook provides a detailed technical description of the three servers
and explains how to plan, install, configure, and manage these high-performance
servers running 32-bit and 64-bit versions of Windows® Server 2003, Red Hat
Enterprise Linux®, SUSE Linux Enterprise Server, and VMware ESX Server.
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, Raleigh Center.
David Watts is a Consulting IT Specialist at the IBM ITSO Center in Raleigh.
He manages residencies and produces Redbooks™ on hardware and
software topics related to IBM Eserver xSeries systems and associated
client platforms. He has authored over 30 Redbooks and Redpapers. He
holds a Bachelor of Engineering degree from the University of Queensland
(Australia) and has worked for IBM for over 15 years. He is an IBM Eserver
Certified Specialist for xSeries and an IBM Certified IT Specialist.
Sergey Gorbas is an IBM ^® xSeries Technical Sales Team leader in
Russia. He has been working for IBM for 11 years. He has more than nine years
of experience in xSeries and Netfinity® servers, including pre-sales and
post-sales technical support. He holds a degree in Computer Engineering from
the Moscow Engineering Physics Institute in Moscow. He is an IBM^
Certified Specialist for xSeries and his areas of expertise include xSeries
systems management hardware and software, including IBM Director,

x
Planning and Installing the IBM Eserver X3 Architecture Servers
IBM ^ Cluster 1350, Microsoft® clustering, Microsoft Datacenter
solutions, and IBM DS Storage Servers.
Christian Schroeder is a Support Specialist in Germany. He has been working
in this function and with IBM for six years. His areas of expertise include
IBM ^ xSeries, BladeCenter®, systems management hardware, and
Windows operating systems.
Joshua Young is a university intern at the International Technical Support
Organization. He has four years of experience in Information Technology related
fields. He will earn a degree in Information Sciences and Technology from
Pennsylvania State University. His skill areas include server environments,
operating systems, team management, and problem-based learning. He also
co-authored the IBM Eserver xSeries 260 Solution Assurance Product Review
Guide, REDP-4007.
Figure 1 The team (l-r): Sergey, David, Josh, and Christian
Thanks to the following people for their contributions to this project:
Daniel Ghidali, IBM Center for Microsoft Technologies, Kirkland
Michael Lee, IBM Center for Microsoft Technologies, Kirkland
Silvio Erdenberger, xSeries Pre-Sales Support Specialist, Germany
Ralph Begun, xSeries Engineering, Raleigh
Gregg McKnight, xSeries Performance Lab, Raleigh

Preface
xi
Become a published author
Join us for a two- to six-week residency program! Help write an IBM Redbook
dealing with specific products or solutions, while getting hands-on experience
with leading-edge technologies. You'll team with IBM technical professionals,
Business Partners, and clients.
Your efforts will help increase product acceptance and client satisfaction. As a
bonus, you'll develop a network of contacts in IBM development labs, and
increase your productivity and marketability.
Find out more about the residency program, browse the residency index, and
apply online at:
ibm.com/redbooks/residencies.html
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 e-mail to:
redbook@us.ibm.com
Mail your comments to:
IBM Corporation, International Technical Support Organization
Dept. HZ8 Building 662
P.O. Box 12195
Research Triangle Park, NC 27709-2195

xii
Planning and Installing the IBM Eserver X3 Architecture Servers

© Copyright IBM Corp. 2006. All rights reserved.
1
Chapter 1.
Technical overview
The IBM Eserver X3 Architecture is the culmination of many years of research
and development and has resulted in what is currently the fastest processor and
memory controller in the Intel® processor marketplace. With support for up to 32
Xeon® MP processors and over 20 GBps of memory bandwidth per 64 GB of
RAM (up to a maximum of 512 GB), the xSeries servers that are based on the X3
Architecture offer maximum performance and broad scale-up capabilities.
Topics in this chapter are:
1.1, “The X3 Architecture servers” on page 2
1.2, “Key features” on page 3
1.3, “Multi-node capabilities and partitioning” on page 4
1.4, “IBM XA-64e third generation chipset” on page 8
1.5, “Scalable implementation” on page 9
1.6, “Current models” on page 12
1.7, “Front and rear layouts” on page 15
1.8, “Processors” on page 18
1.9, “XceL4v cache” on page 27
1.10, “Memory” on page 28
1.11, “Serial Attached SCSI” on page 31
1.12, “PCI-X slots” on page 34
1.13, “Remote Supervisor Adapter II SlimLine” on page 36
1.14, “Baseboard Management Controller” on page 37
1.15, “Light path diagnostics and PFA” on page 38
1

2
Planning and Installing the IBM Eserver X3 Architecture Servers
1.1 The X3 Architecture servers
The three servers based on the X3 Architecture are the x460, x366 and x260.
The have a common set of technical specifications and features, but the key
differences are:
The
xSeries 460
is the flagship server with the following characteristics:
Each chassis occupies 3U of rack
space and supports four CPUs
and 64 GB of RAM.
Six hot-swap drive bays and six
266 MHz PCI-X 2.0 hot-swap
slots.
Targeted at eight-way and above
configurations where effective scale-up options are essential.
Up to eight systems can be connected together to form one single 32-way
complex with up to 512 GB RAM.
The
xSeries 366
is a high-performance four-way server with the following
characteristics:
The same mechanical design as
the x460, offering up to 64 GB of
RAM and up to four Xeon MP
processors.
Targeted at two-way and four-way
high performance commercial
computing, such as database,
e-mail, and e-commerce
applications.
The
xSeries 260
is also a high-performance four-way server:
Same central electronics as the
x366 and x460.
Larger 7U chassis to house up to
12 hot-swap disk drives and a
full-height internal tape drive.
Targeted at two-way and four-way
high performance commercial
computing applications where
more internal disk storage is
required.

Chapter 1. Technical overview
3
1.2 Key features
The x460, x366, and x260 have a number of common features
IBM ^ X3 Architecture, featuring the XA-64e third-generation chipset.
Common system boards: the CPU/memory board, the I/O board, and the
PCI-X board.
Up to four Intel Xeon MP processors. These Processors support 64-bit
addressing with the Intel Extended Memory 64 Technology (EM64T)
architecture.
Existing or planned support for dual-core processors. The x460 and x366 now
support the Xeon 7020 and 7040 (“Paxville”) processors.
Up to 64 GB of RAM, using high performance PC2-3200 ECC DDR2 DIMMs.
Active Memory with Memory ProteXion, memory mirroring, memory hot-swap
and hot-add, and ChipKill.
Six full-length 64-bit 266 MHz PCI-X 2.0 Active PCI slots.
Integrated Adaptec AIC-9410 serial-attached SCSI (SAS) controller.
Support for internal RAID arrays using an optional ServeRAID™-8i adapter.
ServeRAID-6M is also supported for external SCSI storage with the EXP400
enclosure.
Integrated dual-port Broadcom 5704 PCI-X Gigabit Ethernet.
Integrated Baseboard Management Controller. Remote Supervisor Adapter II
SlimLine adapter standard (x460) or optional (x366 and x260).
Support for the IBM Integrated xSeries Adapter for iSeries™ (IXA) for a direct
high speed link to an iSeries server (x460 and x366 only).
Hot-swap fans and power supply.
Light path diagnostics to identify any failed components.
Three-year warranty on-site, nine hours per day, five days per week, with a
next business day response.

4
Planning and Installing the IBM Eserver X3 Architecture Servers
The key features that differ between the three servers are shown in Table 1-1.
Table 1-1 Key feature differences
1.3 Multi-node capabilities and partitioning
The x460 offers the ability to expand the server by connecting multiple chassis
together to form a larger complex. These chassis (or
nodes
) can be either other
x460s or MXE-460 modular expansion enclosures.
xSeries 460 xSeries 366 xSeries 260
Processors Intel Xeon MP “Potomac”
or Xeon 7020/7040
“Paxville” processors
Intel Xeon MP “Cranford”
or Xeon 7020/7040
“Paxville” processors
Intel Xeon MP “Cranford”
processors
Installed / max
processors
2 / 4 1 / 4 1 / 4
Memory standard /
max
2 / 64 GB 2 / 64 GB 1 or 2 / 64 GB
Largest
configuration
Eight nodes (32-way) One node (4-way) One node (4-way)
Rack height 3U 3U 7U
Tower-to-rack
conversion
No No Yes
Power supplies 2x 1300W supplies
(650W at 110V), both
standard
1300W supplies (650W
at 110V), one standard,
one optional
2x 775W supplies / 4
(220V or 110V), both
standard
Remote Supervisor
Adapter II SlimLine
Standard Optional Optional
Hot-swap disk drive
bays
Six (2.5” bays, SAS) Six (2.5” bays, SAS) Six standard, additional
six optional (3.5” bays)
Optical media 8x DVD-ROM 8x DVD-ROM 40x CD-ROM
Diskette drive Optional (external USB) Optional (external USB) Standard (internal)
Tape drive bay No No Two half-high 5.25” bays,
can be used as a single
full-height bay

Chapter 1. Technical overview
5
The MXE-460 modular expansion enclosure is a system used to extend an x460
configuration. Like the x460, it contains microprocessors, memory, disks, and
PCI-X adapters. However, unlike the x460, the MXE-460 can only be used to
expand an x460 configuration.
The MXE-460 is functionally identical to the x460 and supports the same
hardware options. The key differences between the x460 and MXE-460 are:
The MXE-460 is for expansion purposes and cannot be used as the primary
node of a multi-node complex, and it cannot be used as the primary node in a
partition.
The MXE-460 does not have a DVD-ROM drive installed.
The MXE-460 has no processors installed as standard.
The MXE-460 has no memory installed (although two memory cards are
installed).
An x460 server can be configured together with one, three, or seven MXE-460s
to form a single 8-way, 16-way, or 32-way complex. This is shown in Figure 1-1
on page 6. Other multi-node configurations are not supported.
Note: The x260 and x366 cannot expand beyond the single chassis. Only the
x460 and MXE-460 support multi-node configurations.
Important: All configurations with 32 CPU cores (that is, 16-way for dual-core
x460 systems) will require US government approval regardless of where the
complex will be installed.

6
Planning and Installing the IBM Eserver X3 Architecture Servers
Figure 1-1 The four multi-node configurations supported
Additional x460s can substitute for the MXE-460s in Figure 1-1. While this does
increase the cost of the complex, it does mean you can implement partitioning.
This is discussed further in 3.3.3, “Partitioning” on page 116.
multi-node complexes support partitioning on node boundaries. This means, for
example, you can logically partition up your 16-way as two 8-way systems, while
still leaving the complex cabled as a 16-way. This increases flexibility. You can
reconfigure the complex using the Web interface without changing the systems
or cabling.
However, as discussed in 3.3.3, “Partitioning” on page 116, every partition must
have an x460 as the primary node. This restriction means that clients must
decided up front if they wish to use the partitioning feature and order an
appropriate combination of x460 and MXE-460 systems.
For more information about multi-node complexes, see 3.3, “Multi-node
configurations” on page 104.
A multi-node configuration can be partitioned using the RSA II Web interface.
However, each partition must have an x460 as the primary node. You cannot
2-way or 4-way
Up to 64 GB RAM
x460 single chassis
8-way
Up to 128 GB RAM
x460 + 1x MXE-460
Two chassis
16-way
Up to 256 GB RAM
x460 + 3x MXE-460
Four chassis
32-way
Up to 512 GB RAM
x460 + 7x MXE-460
Eight chassis
x460 x460
MXE-460
x460 x460
MXE-460
MXE-460
MXE-460
MXE-460
MXE-460
MXE-460
MXE-460
MXE-460
MXE-460
MXE-460

Chapter 1. Technical overview
7
partition a multi-node configuration so that any partition has only MXE-460
nodes. If you want maximum partitioning flexibility, all your nodes should be x460
systems.
Figure 1-2 shows some of the supported partitioning options. The key messages
from this chart are:
Partitioning is always at node boundaries. You cannot have, for example, two
microprocessors in a node in one partition and the other microprocessors in
that node in another partition.
The primary node for every partition must be an x460. The use of MXE-460s
limits the partitioning you can do.
An eight-node configuration is not shown in the figure, but the same rules apply.
Figure 1-2 Supported partitioning for 2-node and 4-node complexes
Important: If you plan to implement partitioning on your multi-node
configuration, then you must have one x460 for every partition and to be the
primary node for that partition. The secondary nodes can be MXE-460 or x460
units. MXE-460s are not supported as primary nodes.
x460
x460
Partition options:
1x 8-way
2x 4-way
x460
MXE-460
Partition options:
1x 8-way
x460
x460
Partition options:
1x 16-way
2x 8-way
1x 8-way & 2x 4-way
4x 4-way
x460
x460
x460
MXE-460
Partition options:
1x 16-way
MXE-460
MXE-460
x460
MXE-460
Partition options:
1x 16-way
2x 8-way
x460
MXE-460

8
Planning and Installing the IBM Eserver X3 Architecture Servers
1.4 IBM XA-64e third generation chipset
The x260, x366, and x460 use the third generation IBM XA-64e chipset. The
architecture consists of the following components:
One to four Xeon MP processors
One Hurricane Memory and I/O Controller (MIOC)
Two Calgary PCI Bridges
Figure 1-3 shows the block diagram of the X3 Architecture.
Figure 1-3 X3 Architecture system block diagram
66
Six PCI-X 2.0 slots:
64-bit 266 MHz
IBM XA-64e
core chipset
CPU 2
CPU 3
6
5
4
DDR2
SMI2
DDR2
SMI2
DDR2
SMI2
DDR2
SMI2
CPU 1
CPU 4
Memory
controller
("Hurricane")
3
2
1
HDD
backplane
Adaptec
SAS
Gigabit
Ethernet
ServeRAID
266 MHz266 MHz266
33
South
bridge
EIDE
RSA SL
Each:
667 MHz
5.33 GBps
667 MHz
5.33 GBps
6 GBps
667 MHz
5.33 GBps
Calgary
PCI-X bridge
PCI-X bridge
USB 2.0
Video
6 GBps
6 GBps
Serial
K/M
Scalability ports
Each: 6.4 GBps
(x460 and
MXE-460 only)

Chapter 1. Technical overview
9
Each memory port out of the memory controller has a peak throughput of 5.33
GBps. DIMMs are installed in matched pairs, two-way interleaving, to ensure the
memory port is fully utilized. Peak throughput for each PC2-3200 DDR2 DIMM is
2.67 GBps. The DIMMs are run at 333 MHz to remain in sync with the throughput
of the front-side bus.
Because there are four memory ports, spreading installed DIMMs across all four
memory ports can improve performance. The four independent memory ports, or
memory cards, provide simultaneous access to memory. With four memory
cards installed, and DIMMs in each card, peak memory bandwidth is 21.33
GBps.
The memory controller routes all traffic from the four memory ports, two
microprocessor ports, and the two PCI bridge ports. The memory controller also
has embedded DRAM, which holds a snoop filter lookup table. This filter ensures
that snoop requests for cache lines go to the appropriate microprocessor bus
and not both of them, thereby improving performance.
One PCI bridge supplies four of the six 64-bit 266 MHz PCI-X slots on four
independent PCI-X buses. The other PCI bridge supplies the other two PCI-X
slots (also 64-bit, 266 MHz), plus all the onboard PCI devices, including the
optional ServeRAID-8i and Remote Supervisor Adapter II SlimLine daughter
cards.
1.5 Scalable implementation
In the standard single chassis 4-way configuration, the X3 Architecture servers
act as an industry standard symmetric multiprocessor (SMP) system. Each
processor has equal access to all system resources.
However, with multi-node x460 configurations, a NUMA-like architecture
(non-uniform memory architecture) is implemented by connecting the scalability
ports of each node together (see Figure 1-3 on page 8). These ports are directly
connected to the Hurricane memory controller and allow high speed
communication between processors located in different nodes. The ports act like
hardware extensions to the CPU local buses. They direct read and write cycles
to the appropriate memory or I/O resources, as well as maintain cache
coherency between the processors.
In such multi-node configurations, the physical memory in each node is
combined to form a single coherent physical address space. The resulting
system has the property such that for any given region of physical memory,
some processors are closer to it than other processors. Conversely, for any
processor, some memory is considered local and other memory is remote.

10
Planning and Installing the IBM Eserver X3 Architecture Servers
Memory can be described in one of three ways, depending on the relation to a
given processor:

Local
: Memory is in the same node as the processor.

Remote
: Memory is installed in another node that is directly connected via
scalability cables.

Far
: In 8-node configurations, some nodes are not directly connected
together (nodes 1 and 7, for example, as shown in Figure 3-9 on page 109).
In such situations, memory that is two “hops” (two cables) away from a given
processor is “far”.
The term NUMA is not completely correct since not only can memory be
accessed in a non-uniform manner, but also I/O resources. PCI-X and USB
devices may be associated with nodes. The exceptions to this situation are
existing I/O devices, such as diskette and CD-ROM drives, which are disabled
because the classic PC architecture precludes multiple copies of these existing
items.
The key to this type of memory configuration is to limit the number of processors
that directly access a piece of memory, thereby improving performance because
of the much shorter queue of requests. The objective of the operating system is
to ensure that memory requests be fulfilled by local memory whenever possible.
However, an application running on CPUs in node 1 may still need to access
memory physically located in node 2 (a remote access). This access incurs
longer latency because the travel time to access remote memory on another
expansion module is clearly greater. Many people think this is the problem with
NUMA. But this focus on latency misses the actual problem NUMA is attempting
to solve: shorten memory request queues.
Another way to think about this is to ask yourself this question: You are checking
out in your favorite grocery store, with a shopping cart full of groceries. Directly in
front of you is a check-out lane with 10 customers standing in line but 20 meters
to your left is another check-out lane with only two customers standing in line.
Which would you go to? The check-out lane closest to your position has the
lowest latency because you don't have far to travel. But the check-out lane 20
meters away has much greater latency because you have to walk 50 feet.
Clearly most people would walk the 20 meters and thereby incurring a delay to
arrive at a check-out lane with only 2 customers instead of 10. We think this way
because our experience tells us that the time waiting to check-out with 10 people
ahead of us (the request queue) is far longer than the time needed to walk to the
“remote” check-out lane (latency) and wait for only two people ahead.

Chapter 1. Technical overview
11
This analogy clearly communicates the performance effects of queuing time
versus latency. In a computer server, with many concurrent outstanding memory
requests, we would gladly incur come additional latency (walking) to spread
memory transactions (check-out process) across multiple memory controllers
(check-out lanes), because this greatly improves performance by reducing the
queuing time.
The performance implications of such a configuration are significant. It is
essential that the operating system recognize which processors and ranges of
memory are local, which are remote, and which are far.
Clearly, we do not want to walk 20 meters to a check-out lane that has 20
customers checking out, when one is directly in front of us with only two
customers. So to reduce unnecessary remote access, the x460 maintain a table
of data in the firmware called the Static Resource Allocation Table (SRAT). The
data in this table is accessible by operating systems such as Windows Server
2003 (Windows 2000 Server does not support it) and current Linux kernels.
These modern operating systems attempt to allocate resources that are local to
the processors being used by each process. So when a process and its threads
start on node 1, all execution and memory access will be local to node 1. As
more processes are added to the system, the operating system will balance them
across the nodes. In this case, most memory accesses will be evenly distributed
across the multiple memory controllers, reducing remote access, greatly
reducing queuing delays, and improving performance.

12
Planning and Installing the IBM Eserver X3 Architecture Servers
1.6 Current models
The following models were announced at the time of publication.
1.6.1 x460 and MXE-460 models
The following are the currently available models of the x460 (Table 1-2) and the
MXE-460 modular expansion enclosure (Table 1-3). The MXE-460 is similar to
the x460, except the MXE-460 has no memory or processors standard and no
DVD-ROM drive. The MXE-460 is used to form x460 multi-node configurations.
Table 1-2 x460 models
The x460 supports two or four processors per node, up to eight nodes.
Table 1-3 lists the MXE-460. The type of microprocessors installed in the
MXE-460s must match the primary x460 in the multi-node configuration.
Table 1-3 MXE-460 models
Model Standard/max CPU Cores/
socket
L2
cache
L3
cache
Std/max memory
8872-1RY 2x 2.83 GHz Xeon MP / 4 Single-core 1 MB 4 MB 2 GB (2x 1 GB) / 64
8872-2RY 2x 3.00 GHz Xeon MP / 4 Single-core 1 MB 8 MB 2 GB (2x 1 GB) / 64
8872-3RY 2x 3.33 GHz Xeon MP / 4 Single-core 1 MB 8 MB 2 GB (2x 1 GB) / 64
8872-5RU 2x 2.67 GHz Xeon 7020 / 4 Dual-core 1+1 MB 0 MB 2 GB (2x 1 GB) / 64
8872-6RU 2x 3.00 GHz Xeon 7040 / 4 Dual-core 2+2 MB 0 MB 2 GB (2x 1 GB) / 64
Model Standard/max CPU Cores/socket L2 cache L3 cache Std/max memory
8874-1RY 0 / 4 (match the at-
tached x460)
Single-core 1 MB Varies 0 / 64 GB
8874-2RU 0 / 4 (match the at-
tached x460
Dual-core Varies 0 MB 0 / 64 GB
Note: The U or Y in the model numbers is for countries in North and South
America. For EMEA, substitute G, for example 1RG. For Asia-Pacific
countries, the letter varies from country to country. Consult the announcement
letter or the xSeries Configuration and Option Guide, found at:
http://www.pc.ibm.com/support?page=SCOD-3ZVQ5W

Chapter 1. Technical overview
13
In addition to these models, IBM has also announced models that include
Windows Server 2003, Datacenter Edition. These are described in 4.5,
“Datacenter offerings” on page 147.
Both the x460 and MXE-460 support a maximum of 64 GB using 4 GB DIMMs in
16 sockets. The MXE-460 has no standard memory installed, but requires at
least 2 GB of memory. The amount of memory installed does not have the match
the other nodes in a multi-node complex, but for performance reasons, this is
recommended.
Both the x460 and MXE-460 have six internal PCI-X 2.0 slots. Unlike the x445,
the x460 does not support the attachment of an RXE-100 Remote Expansion
Enclosure. The RSA II SlimLine and ServeRAID-8i options do not occupy any of
these six PCI-X slots.
1.6.2 x366 models
The models shown in Table 1-4 are available.
Table 1-4 x366 models announced in March 2005
The x366 supports one, two, three, or four processors.
Both models support a maximum of 64 GB using 4 GB DIMMs in 16 sockets. To
achieve the maximum, you will need to install three additional memory cards
(one four-socket card is standard), remove the standard pair of 1 GB DIMMs,
and insert 16 DIMMs.
The x366 has six internal PCI-X 2.0 slots. Unlike the x365, the x366 does not
support the attachment of an RXE-100 Remote Expansion Enclosure. The RSA
II SlimLine and ServeRAID-8i options do not use any of these six PCI-X slots.
Model CPU (std/max) Cores/
socket
L2 cache L3 cache Memory (std/max)
8863-1RY 1x Xeon MP 3.16 GHz / 4 Single-core 1 MB None 2 GB (2x 1 GB) / 64
8863-2RY 1x Xeon MP 3.66 GHz / 4 Single-core 1 MB None 2 GB (2x 1 GB) / 64
8863-3RU 1x 2.67 GHz Xeon 7020 / 4 Dual-core 1+1 MB None 2 GB (2x 1 GB) / 64
8863-4RU 1x 3.00 GHz Xeon 7040 / 4 Dual-core 2+2 MB None 2 GB (2x 1 GB) / 64
Express models: Certain countries also offer preconfigured systems called
Express models. These are preconfigured with more processors, memory,
disk, or a ServeRAID controller. The advantage of these models is fewer part
numbers to order.

14
Planning and Installing the IBM Eserver X3 Architecture Servers
1.6.3 x260 models
The following models of the x260 (Table 1-5) were announced in July 2005.
Table 1-5 x260 models announced in July 2005
The x260 supports one to four processors.
All of the x260 series support a maximum of 64 GB using 4 GB DIMMs in 16
sockets. x260s have four standard DIMMS installed, but have 12 expandable
sockets.
The x260 has six internal PCI-X 2.0 slots. The optional RSA II SlimLine and
optional ServeRAID-8i adapters do not occupy any of these six PCI-X slots.
For clients who purchase a tower model and later wish to install the server into a
suitable rack, a Tower To Rack Conversion Kit, part number 32R0719, is also
available and provides the hardware needed to convert the tower unit to a
7U-high, rack-mounted server. It includes slides, cable management arm,
latches, and miscellaneous mounting hardware.
Model Configuration Standard/max CPU L2 cache Standard/max memory
8865-12U Tower 1x 3.16 GHz Xeon MP / 4 1 MB 1 GB (2x 512 MB) / 64 GB
8865-1SU Rack 1x 3.16 GHz Xeon MP / 4 1 MB 1 GB (2x 512 MB) / 64 GB
8865-11U Tower 1x 3.16 GHz Xeon MP / 4 1 MB 2 GB (2x 1 GB) / 64 GB
8865-1RU Rack 1x 3.16 GHz Xeon MP / 4 1 MB 2 GB (2x 1 GB) / 64 GB
8865-21U Tower 1x 3.66 GHz Xeon MP / 4 1 MB 2 GB (2x 1 GB) / 64 GB
8865-2RU Rack 1x 3.66 GHz Xeon MP / 4 1 MB 2 GB (2x 1 GB) / 64 GB
Note: The U in the model numbers is for countries in North and South
America. For EMEA, substitute G for the U, for example, 1RG. For
Asia-Pacific countries, the letter varies from country to country. Consult the
announcement letter or the xSeries Configuration and Option Guide, found at:
http://www.pc.ibm.com/support?page=SCOD-3ZVQ5W
Express models: Certain countries also offer preconfigured systems called
Express models. These are preconfigured with various combinations of more
processors, memory, disk, and a ServeRAID controller. The advantage of
these is fewer part numbers to order.

Chapter 1. Technical overview
15
1.7 Front and rear layouts
The three X3 Architecture servers share many components. This section
describes the components accessible from the front and rear of each server.
1.7.1 x460 and MXE-460
The front panel of the x460 and MXE-460 is shown in Figure 1-4.
Figure 1-4 Front panel of the x460 and MXE-460
Processor tray
(behind bezel)
Six hot-swap
disk drive
bays
DVD-ROM
drive (x460)
USB port Operator
panel
Panel release
button to display
light path panel

16
Planning and Installing the IBM Eserver X3 Architecture Servers
The rear panel of the x460 and MXE-460 is shown in Figure 1-5.
Figure 1-5 Rear panel of the x460 and MXE-460
1.7.2 x366
Figure 1-6 shows the x366 and major components on the front of the unit.
Figure 1-6 Front panel of the x366
The rear panel of the x366 is shown in Figure 1-7 on page 17.
RSA II SlimLine
serial port
System serial port
Keyboard
Mouse
RSA II SlimLine
Ethernet port
2x Gigabit
Ethernet
2x USB
IXA RS-485
(not service processor)
Video
Scalability ports
1 2
3
Processor tray
(behind bezel)
Six hot-swap
disk drive
bays
DVD-ROM
drive
USB port Operator
panel
Panel release
button to display
light path panel

Chapter 1. Technical overview
17
Figure 1-7 Rear view of the x366
1.7.3 x260
Figure 1-8 shows the front panel of the x260.
Figure 1-8 Front panel of the x260
Service processor
serial port
System serial port
Keyboard
Mouse
Service processor
Ethernet port
2x Gigabit
Ethernet
2x USB
IXA RS-485
(not service processor)
Video
Operator information
panel
Standard
CD-ROM drive
Standard
diskette drive
USB port
Six optional drive bays
(requires backplane)
Six standard
drive bays
Two 5.25” SL drive
bays (or one full-height
drive bay
Processor tray
(behind bezel)

18
Planning and Installing the IBM Eserver X3 Architecture Servers
Figure 1-9 shows the rear panel of the x260.
Figure 1-9 Rear panel of the x260
1.8 Processors
The X3 Architecture servers all use Intel Xeon processors with EM64T
extensions.
The x460 models use either
Potomac
single-core processors or
Paxville

dual-core processors. Models of the x460 have two processors installed. Two or
four processors are supported. Installed processors must be identical in speed
and cache size. You can connect MXE-460s to the x460 to form larger
configurations. The x460 and each MXE-460 must have four identical processors
installed. Multi-node capabilities are discussed in detail in 3.3, “Multi-node
configurations” on page 104.
The x366 uses either
Cranford
single-core processors or
Paxville
dual-core
processors. The x260 uses
Cranford
processors. Models of the x260 and x366
have one processor installed. One, two, three, or four processors are supported.
Installed processors must be identical in speed and cache size.
RSA II serial
System serial
Keyboard
Mouse
RSA II Ethernet
Gigabit Ethernet 2
Gigabit Ethernet 1
Two USB ports
Unused port
(RS-485 for unsupported
IXA Adapter only)
Video

Chapter 1. Technical overview
19
The processors are accessible from the front of the server on a sliding tray. The
tray can be pulled out once the memory cards and fans are removed (see
Figure 1-10).
Figure 1-10 Processor tray (x460 shown)
The VRMs for processors 1 and 2 are integrated on the microprocessor board;
the VRMs for processors 3 and 4 come with the processor options.
See 3.1, “Processor subsystem” on page 90 for a further discussion about what
you should consider before implementing a solution.
The
Cranford
and
Paxville
processors have two levels of cache on the
processor die:
L2 cache is 1 MB in size. The L2 cache implements the Advanced Transfer
Cache technology, which means L2-to-processor transfers occur across a
256-bit bus in only one clock cycle.
L1 execution trace cache is used to store micro-operations (that is, decoded
executable machine instructions); it serves those to the processor at rated
speed. This additional level of cache saves decode time on cache hits.
Dual-core upgrade: The Dual Core X3 Upgrade Kit, part 39Y6580, can be
used to upgrade an existing single-core x366 or x460 system to dual-core
capability. The kit does not include the dual-core processors.
FRO
NT
FRONT
FRONT
1
2
4
3
The processor tray pulls out
from the front of the server
and houses the CPUs,
VRMs, and memory
controller.

20
Planning and Installing the IBM Eserver X3 Architecture Servers
The
Potomac
Xeon MP processor used in the x460 and MXE-460 has three
levels of cache on the processor die:
The L3 cache is 4 MB or 8 MB, depending on the processor.
The L2 cache is 1 MB in size. The L2 cache implements the Advanced
Transfer Cache technology, which means L2-to-processor transfers occur
across a 256-bit bus in only one clock cycle.
The L1 execution trace cache is used to store micro-operations and decoded
executable machine instructions. It serves them to the processor at rated
speed. This additional level of cache saves decode time on cache hits.
Key features of both processors are described in the following subsections.
1.8.1 Models with dual-core processors
The new Paxville dual-core processors are a concept similar to a two-way
system except that the two processors, or
cores
, are integrated into one silicon
die. This brings the benefits of two-way SMP with less power consumption and
faster data throughput between the two cores. To keep power consumption
down, the resulting core frequency is lower, but the additional processing
capacity means an overall gain in performance.
Figure 1-11 compares the basic building blocks of the Xeon MP single-core
processor (Potomac) and dual-core processor (Paxville).
Figure 1-11 Feature of single core and dual core processors
L1
Instruct
Cache
L1
Data
Cache
Processor
Core
L3
Cache
L2
Cache
L1
Instruct
Cache
L1
Data
Cache
Processor
Core
L2
Cache
Single-core Xeon MP
(Code name: Potomac)
Dual-core Xeon 7020/7040
(Code name: Paxville)
One processor core
L3 cache implemented
High frequency
Two processor cores
No L3 cache
Moderate high frequency
More parallelism
L1
Instruct
Cache
L1
Data
Cache
Processor
Core
L2
Cache

Chapter 1. Technical overview
21
In addition to the two cores, the dual-core processor has separate L1 instruction
and data caches for each core as well as separate execution units (integer,
floating point, and so on), registers, issue ports, and pipelines for each core. A
dual core processor achieves more parallelism than Hyper-Threading
Technology because these resource are not shared between the two cores.
Estimates are that there is a 1.2 to 1.5 times improvement comparing the
dual-core Xeon MP with current single-core Xeon MP.
With double the number of cores for the same number of sockets, it is even more
important that the memory subsystem is able to meet the demand for data
throughput. The 21 GBps peak throughput of the X3 Architecture of the x460 with
four memory cards is well suited to dual-core processors.
1.8.2 667 MHz front-side bus
The Pentium® III Xeon processor in older servers had a 100 MHz front-side bus
that equated a burst throughput of 800 MBps. With protocols such as TCP/IP,
this has been shown to be a bottleneck in high-throughput situations.
Both the Cranford and Potomac Xeon MP improves on this by using two 133
MHz clocks, out of phase with each other by 90°, and using both edges of each
clock to transmit data. This is shown in Figure 1-12.
Figure 1-12 Quad-pumped front-side bus
This increases the performance of the front-side bus without the difficulty of
high-speed clock signal integrity issues. Because the bus is eight bytes wide, the
end result is an effective burst throughput of 5.33 GBps, which can have a
substantial impact especially on TCP/IP-based LAN traffic.
1.8.3 Hyper-Threading
Hyper-Threading technology enables a single physical processor to execute two
separate code streams,
threads
, concurrently. To the operating system, a
processor with Hyper-Threading appears as two
logical
processors, each of
which has its own architectural state: data, segment and control registers, and
advanced programmable interrupt controller (APIC).
133 MHz clock A
133 MHz clock B

22
Planning and Installing the IBM Eserver X3 Architecture Servers
Each logical processor can be individually halted, interrupted, or directed to
execute a specified thread, independently of the other logical processor on the
chip. Unlike a traditional two-way SMP configuration that uses two separate
physical processors, the logical processors share the execution resources of the
processor core, which include the execution engine, the caches, the system bus
interface, and the firmware.
The basic layout of a Hyper-Threading-enabled microprocessor is outlined in
Figure 1-13, where you can clearly see that only the components for the
architectural state of the microprocessor have doubled.
Figure 1-13 Architectural differences associated with Hyper-Threading
Hyper-Threading Technology is designed to improve server performance by
exploiting the multithreading capability of operating systems, such as Microsoft
Windows 2003, Linux, and server applications, in such a way as to increase the
use of the on-chip execution resources available on these processors.
Fewer or slower processors usually achieve the best gains from
Hyper-Threading, because there is a greater likelihood that the software can
spawn sufficient numbers of threads to keep both paths busy. The following
performance gains are likely:
Two physical processors: up to about 25% performance gain
Four physical processors: up to about 15% performance gain
Eight physical processors: up to about 10% performance gain
Architectural state
Cache
Processing
resources
Cache
Processing
resources
Architectural
state
Architectural
state
Physical processor
Logical processor
Processor
without
Hyper-Threading
Processor with
Hyper-Threading

Chapter 1. Technical overview
23
Tests have shown that software often limits SMP scalability, but clients should
expect improved results as software matures. Best-case applications today are:
Databases
Java™
Web servers
E-mail
Note that Microsoft licensing of the Windows 2000 Server operating systems is
by number of processors: four-way for Server, eight-way for Advanced Server,
and 32-way for Datacenter Server. Therefore, the appearance of twice as many
logical processors can potentially affect the installation of the operating system.
See 4.1, “Supported operating systems” on page 136 for details.
Windows Server 2003 understands the concept of physical processors versus
logical processors.
1.8.4 Intel 64-bit Extension Memory Technology (EM64T)
First introduced in the Xeon DP
Nocona
processor, EM64T is a 64-bit extension
to the industry standard IA32 32-bit architecture. EM64T adds:
A set of new 64-bit general purpose registers (GPR)
64-bit instruction pointers
The ability to process data in 64-bit chunks
Even though the names of these extensions suggest that the improvements are
simply in memory addressability, Intel EM64T is, in fact, a fully functional 64-bit
processor.
A 64-bit processor is defined as one that is able to address 64-bits of virtual
address space. A 64-bit processor can store data in 64-bit format and perform
arithmetic operations on 64-bit operands. In addition, a 64-bit processor has
general purpose registers (GPRs) and arithmetic logical units (ALUs) that are
64-bit wide.
There are now three 64-bit implementations in the “Intel compatible processor”
marketplace:
Intel IA64, as implemented on the Itanium® 2 processor
Intel EM64T, as implemented on the Xeon DP Nocona and Irwindale
processors and Xeon MP Cranford and Potomac processors.
AMD AMD64, as implemented on the Opteron processor

24
Planning and Installing the IBM Eserver X3 Architecture Servers
The discussion whether or not EM64T really are 64-bit processor can be
confusing to some because Intel calls its architecture “Extended Memory 64
Technology”. We know that it extends the IA32 instruction set. Therefore, are
EM64T processors “real” 64-bit chips? The answer is yes. When these
processors operate in 64-bit mode, the addresses are 64-bit, the GPRs are 64
bits wide, and the ALUs are able to process data in 64-bit chunks. Therefore,
these processors are full-fledged 64-bit processors in this mode.
The Xeon MP with EM64T extends the previous Xeon MP in three ways:
64-bit addressing and 64-bit registers
Additional 128-bit Streaming SIMD Extensions (SSE) registers and 64-bit
general purpose registers
Double precision (64-bit) integer support
There are three distinct operation modes available in EM64T:
32-bit mode
The first and, in the near future, probably most widely used mode will be
32-bit mode. In this mode, any EM64T processors will act just like any other
IA32 compatible processor. You can install your 32-bit OS on such a system
and run 32-bit applications; however, you will not be able to make use of the
new features, such as the flat memory addressing above 4 GB or the
additional 64-bit General Purpose Registers (GPRs). 32-bit applications will
run just as fast as they would on any current 32-bit processor.
Most of the time, IA32 applications will run even faster, since there are
numerous other improvements that boost performance regardless of the
maximum address size.
Compatibility mode
The second mode supported by the EM64T is compatibility mode, which is an
intermediate mode of the full 64-bit mode described below. In order to run in
compatibility mode, you will need to install a 64-bit operating system and
64-bit drivers. If a 64-bit OS and drivers are installed, the Xeon processor will
be enabled to support a 64-bit operating system with both 32-bit applications
or 64-bit applications.
Compatibility mode gives you the ability to run a 64-bit operating system while
still being able to run unmodified 32-bit applications. Each 32-bit application
will still be limited to a maximum of 4 GB of physical memory. However, the 4
GB limit is now imposed on a per-process level, not at a system-wide level.
This means that every 32-bit process on this system gets its very own 4 GB of
physical memory space (assuming sufficient physical memory is installed).
This is a big improvement compared to IA32, where the operating system
kernel and the application have to share 4 GB of physical memory and

Chapter 1. Technical overview
25
applications usually get no more then 3 GB of memory. 32-bit applications
modification does not require, but BIOS updates will needed
Additionally, compatibility mode does not support the virtual 8086 mode, so
real-mode existing applications are not supported. 16-bit protected mode
applications are, however, supported.
Full 64-bit mode
The final mode is the full 64-bit mode. Intel refer to it as IA-32e mode. This
mode is when a 64-bit operating system and 64-bit application are used. In
the full 64-bit operating mode, an application can have a virtual address
space of up to 40-bits (that equates to 1 TB of addressable memory). The
amount of physical memory will be determined by the server chipset, how
many DIMM slots the server has, and the maximum DIMM capacity
supported and available at the time.
Applications that run in full 64-bit mode will get access to the full physical
memory range (depending on the operating system), and will also get access
to the new GPRs as well as to the expanded GPRs. However it is important to
understand that this mode of operation requires not only a 64-bit operating
system (and of course 64-bit drivers), but also requires a 64-bit application
that has been recompiled to take full advantage of the various enhancements
of the 64-bit addressing architecture
Figure 1-14 gives an overview of the supported modes.
Figure 1-14 Intel 64-bit Extension Technology Supported Modes
EM64T xSeries
Hardware
EM64T xSeries
Hardware
EM64T xSeries
Hardware
Migrating to 64-bit computing
32-bit
device drivers
64-bit
device drivers
Operating
System
Device Drivers
Hardware
Applications
32-bit apps.
32-bit & 64-bit apps.
64-bit apps.
32-bit Mode
Compatibility Mode
64-bit Mode
Windows Server
2003 (32-bit)
Windows Server
2003 x64 Edition
64-bit
device drivers
Windows Server
2003 x64 Edition

26
Planning and Installing the IBM Eserver X3 Architecture Servers
For more information about EM64T, see:
http://www.intel.com/technology/64bitextensions/
1.8.5 The benefit of 64-bit computing
In the same way that 16-bit processors and 16-bit applications are no longer
used in this space, it is likely that at some point in the future 64-bit processors
and applications will replace their 32-bit counterparts.
Processors using the EM64T architectures are making this transition very
smooth by offering 32-bit and 64-bit modes. This means that the hardware
support for 64-bit may likely be in place before you upgrade or replace your
applications with 64-bit versions. The X3 Architecture servers all use
EM64T-enabled processors.
The question you should be asking is whether the benefit of 64-bit processing is
worth the effort of upgrading or replacing your 32-bit applications. The answer is
that it depends on the application. The following are examples of applications
that will benefit from 64-bit computing:
Encryption applications: Most encryption algorithms are based on very large
integers and would benefit greatly with the use of 64-bit GPRs and ALUs.
Scientific applications: Integer-based scientific applications will benefit.
Floating-point operations do not benefit from the larger integer size since
floating-point registers are already 80 or 128 bits wide even in 32-bit
processors.
Applications requiring more than 4 GB of memory: This is the biggest
advantage of 64-bit computing for commercial applications: the flat,
potentially massive, address space.
Enterprise applications, such as databases, are currently implementing Page
Addressing Extensions (PAE) and Addressing Windows Extensions (AWE)
addressing schemes to access memory above the 4 GB limit imposed by 32-bit
address limited processors. These addressing extension schemes support
access to memory up to 64 GB in size. One constraint with PAE and AWE,
however, is that memory above 4 GB can only be used to store data; it cannot be
used to store or execute code. So these addressing schemes only make sense
for applications such as databases, where large data caches are needed.
In contrast, a 64-bit virtual address space provides for direct access of up to 2
Exabytes (EB). And even though we call these processors 64-bit, none of the
current 64-bit processors actually supports the full 64 bits of physical memory
addressing, simply because this is such an enormous amount of memory.

Chapter 1. Technical overview
27
Table 1-6 lists the memory addressability by processor. These values are the
upper limits imposed by the processors. Memory addressing is usually limited
further by the chipset implemented in the server and by the DIMM technology
available.
Table 1-6 Memory supported by processors
1.9 XceL4v cache
The XceL4v dynamic server cache serves two purposes in the X3 Architecture
servers:
As a single, 4-way server (x260, x366 or x460), the XceL4v and its embedded
DRAM (eDRAM) is used as a snoop filter to reduce traffic on the front side
bus. It stores a directory of all processor cache lines to minimize snoop traffic
on the dual front side buses and minimize cache misses.
When the x460 is configured as a multi-node server, this technology
dynamically allocates 256 MB of main memory in each node for use as an L4
cache directory and scalability directory. In a 32-way configuration, this
means there will be 2 GB of XceL4v cache.
With advances in chip design, IBM has now reduced the latency of main memory
to below that of the XceL4 cache in the x445. In other words, the time it takes to
access data directly from memory is almost as fast as accessing it from L3. As a
result, on a four-way system such as the x260 and x366, there is little/no need for
either a L3 cache or L4 cache.
Processor Flat addressing Addressing with PAE-36
Intel Xeon MP “Gallatin” (32-bit) 4 GB (32-bit) 64 GB
Intel EM64T “Nocona” (64-bit) 64 GB (36-bit) 64 GB in compatibility mode
Intel EM64T “Cranford” (64-bit) 64 GB (36-bit) 64 GB in compatibility mode
Intel EM64T “Potomac” (64-bit) 1 TB (40-bit) 64 GB in compatibility mode
Intel EM64T “Paxville” (64-bit) 1 TB (40-bit) 64 GB in compatibility mode
Intel Itanium (64-bit) 18 TB (44-bit) Not applicable
AMD Opteron (64-bit) 256 TB (48-bit) 64 GB in compatibility mode
Intel Itanium 2 (64-bit) 1024 TB (50-bit) Not applicable

28
Planning and Installing the IBM Eserver X3 Architecture Servers
Since the L3 cache is inline, when cache misses do occur, it adds significant
overhead to memory access. The L3 cache rate has to be very high for it to keep
up with the 3.66 GHz processor. In most server applications with multiple users,
the threads competing for L3 cache generate a lower hit rate, and the latency of
the L3 drops performance. The same applies to any L4 cache.
As a result, there is no performance benefit in implementing either an L3 or L4
cache on the four-way x260 or x366. For these reasons, there is 0 MB of XceL4v
cache on the x260 and x366 servers.
1.10 Memory
Memory in the X3 Architecture servers is PC2-3200 ECC DDR2 DIMMs.
Standard memory is as follows:
x260: 1 GB or 2 GB of RAM standard (using 512 MB DIMMs)
x366: 2 GB standard (using 512 MB DIMMs)
x460: 2 GB standard (using 512 MB DIMMs)
MXE-460: None standard
Memory is implemented in the servers using memory cards. Each server
supports up to four memory cards, and each card has four DIMM sockets.
Memory cards are standard as follows:
x260: One card standard
x366: One card standard
x460: Two cards standard
MXE-460: Two cards standard
Using 4 GB DIMMs in every socket, a total of 16 DIMMs, the servers can hold 64
GB of RAM.
The memory is two-way interleaved, meaning that memory DIMMs are installed
in pairs. There are four ports to memory, with each supporting up to 5.33 GBps
data transfers.
The DIMMs operate at 333 MHz instead of 400 MHz, as per the PC2-3200 spec,
so that throughput is 2.67 GBps, or 333 MHz x 8 bytes. At 2.67 GBps and
two-way interleaving, the throughput matches that of the front-side bus at 5.3
GBps so that bus transfers remain in sync.

Chapter 1. Technical overview
29
There are a number of advanced features implemented in the X3 Architecture
memory subsystem, collectively known as
Active Memory
:
Memory ProteXion
The Memory ProteXion feature (also known as
redundant bit steering
)
provides the equivalent of a hot-spare drive in a RAID array. It is based in the
memory controller, and it enables the server to sense when a chip on a DIMM
has failed and to route the data around the failed chip.
Normally, 128 bits out of every 144 are used for data and the remaining 16
bits are used for ECC functions. However, X3 Architecture needs only 12 bits
to perform the same ECC functions, thus leaving four bits free. These four
bits are equivalent to an x4 memory chip on the DIMM that Memory ProteXion
uses. In the event that a chip failure on the DIMM is detected by memory
scrubbing, the memory controller can reroute data around that failed chip
through these spare bits.
It can do this automatically without issuing a Predictive Failure Analysis®
(PFA) or light path diagnostics alert to the administrator, although an event is
logged to the service processor log. After the second DIMM failure, PFA and
light path diagnostics alerts would occur on that DIMM as normal.
Memory scrubbing
Memory scrubbing is an automatic daily test of all the system memory that
detects and reports memory errors that might be developing before they
cause a server outage.
Memory scrubbing and Memory ProteXion work in conjunction with each
other and do not require memory mirroring to be enabled to work properly.
When a bit error is detected, memory scrubbing determines if the error is
recoverable or not. If it is recoverable, Memory ProteXion is enabled and the
data that was stored in the damaged locations is rewritten to a new location.
The error is then reported so that preventative maintenance can be
performed. As long as there are enough good locations to allow the proper
operation of the server, no further action is taken other than recording the
error in the error logs.
Note: In BIOS, there is a Memory Array setting in Advanced Settings.
When you set it to High Performance Memory Array, BIOS reconfigures
the server for maximum performance at the expense of some fault
tolerance features, including Memory ProteXion. For a production
environment, we recommend that you not select this setting, thereby
keeping your system protected from memory failures with Memory
ProteXion.

30
Planning and Installing the IBM Eserver X3 Architecture Servers
If the error is not recoverable, then memory scrubbing sends an error
message to the light path diagnostics, which then turns on the proper lights
and LEDs to guide you to the damaged DIMM. If memory mirroring is
enabled, then the mirrored copy of the data from the damaged DIMM is used
until the system is powered down and the DIMM replaced.
Memory mirroring
Memory mirroring is roughly equivalent to RAID 1 in disk arrays, in that
usable memory is halved and a second copy of data is written to the other
half. If eight GB is installed, then the operating system sees four GB once
memory mirroring is enabled. It is disabled in the BIOS by default. Because all
mirroring activities are handled by the hardware, memory mirroring is
operating system independent.
When memory mirroring is enabled, certain restrictions exist with respect to
placement and size of memory DIMMs and the placement and removal of
memory cards.
Chipkill™ memory
Chipkill is integrated into the XA-64e chipset, so it does not require special
Chipkill DIMMs and is transparent to the operating system. When combining
Chipkill with Memory ProteXion and Active Memory, X3 Architecture provides
very high reliability in the memory subsystem.
When a memory chip failure occurs, Memory ProteXion transparently handles
the rerouting of data around the failed component, as described above.
However, if a further failure occurs, the Chipkill component in the memory
controller reroutes data. The memory controller provides memory protection
similar in concept to disk array striping with parity, writing the memory bits
across multiple memory chips on the DIMM. The controller is able to
reconstruct the missing bit from the failed chip and continue working as usual.
One of these additional failures can be handled for each memory port for a
total of four Chipkill recoveries.
Hot-add and hot-swap memory
The X3 Architecture servers support the replacing of failed DIMMs while the
server is still running. This hot-swap support works in conjunction with
memory mirroring. The server also supports adding additional memory while
the server is running. Adding memory requires operating system support.
These two features are mutually exclusive. Hot-add requires that memory
mirroring be disabled and hot-swap requires that memory mirroring be
enabled.
In addition, to maintain the highest levels of system availability, if a memory error
is detected during POST or memory configuration, the server can automatically
disable the failing memory bank and continue operating with reduced memory

Chapter 1. Technical overview
31
capacity. You can manually re-enable the memory bank after the problem is
corrected by using the Setup menu in the BIOS.
Memory mirroring, Chipkill, and Memory ProteXion provide multiple levels of
redundancy to the memory subsystem. Combining Chipkill with Memory
ProteXion allows up to two memory chip failures for each memory port on the
system, for a total of eight failures sustained. For example:
1.The first failure detected by the Chipkill algorithm on each port does not
generate a light path diagnostics error because Memory ProteXion recovers
from the problem automatically.
2.Each memory port could then sustain a second chip failure without shutting
down.
3.Provided that memory mirroring is enabled, the third chip failure on that port
would send the alert and take the DIMM offline, but keep the system running
out of the redundant memory bank.
1.11 Serial Attached SCSI
Serial Attached SCSI (SAS) is the logical evolution of SCSI. SAS uses much
smaller interconnects than SCSI, while offering SCSI compatibility, reliability,
performance, and manageability. In addition, SAS offers longer cabling
distances, smaller form factors, and greater addressability.
SAS 1.0 technology is replacing Ultra320 SCSI in SCSI and RAID controllers.
The X3 Architecture family of servers have a disk subsystem that is based on an
Adaptec AIC-9410 SAS controller. The servers do not support RAID as standard,
but these servers also support the addition of the ServeRAID-8i daughter card,
which in turn supports a variety of RAID levels and stripe sizes up to 512 KB.

32
Planning and Installing the IBM Eserver X3 Architecture Servers
Beyond the upgrades in I/O processor and memory speeds, SAS-based
products will differ from SCSI-based products in the following ways:
Higher bandwidth
Ultra320 SCSI supports 320 MBps of bandwidth per channel. SAS 1.0
supports three Gbps, approximately 300 MBps, of bandwidth for each port.
So while the two Ultra320 SCSI channels of the ServeRAID-6M can
potentially support 640 MBps of bandwidth, the onboard SAS controller with
its eight ports could support up to 24 Gbps, approximately 2.4 GBps, of
bandwidth. Therefore, bandwidth will be limited by PCI-X or PCI Express bus
speeds.
Greater drive support
SCSI-based products support 14 drives for each channel. By cascading drive
enclosures, SAS-based products will support up to 72 drives for four ports.
The AIC-9410 SAS controller is central to the SAS implementation on the X3
Architecture servers. It has eight SAS ports, each supporting a maximum
transfer rate of 3 Gbps.
Depending on the server, the drives are either directly attached to the SAS ports
(x366 and x460) or they are connected via SAS Expanders (x260). This is shown
in the following two figures.
Figure 1-15 SAS block diagram (x460 and x366)
With the x336 and x460 (Figure 1-15), only six of the eight ports are used from
the SAS controller and these are connected directly to the six drive bays in the
server via the backplane.
64-bit path to
66 MHz PCI-X
bridge
Optional
ServeRAID-8i
6-slot
backplane
AIC-9410 SAS
controller
(SAS initiator)
Ports 1-4
Ports 5-8
Ports 4 & 8 not
connected
SAS drive
Six drive bays
SAS drive

Chapter 1. Technical overview
33
Figure 1-16 SAS block diagram (x260)
For the x260 (Figure 1-16), four ports from the SAS controller chip are fed into
the expander and the signals are expanded out to six devices, similar in concept
to a router. The x260 comes standard with only one of the two drive backplanes.
In such a configuration, only ports 1-4 are used. Ports 5-8 are used only when
the second backplane is installed.
In terms of performance, there is no benefit in inserting drives in any particular
order (for example, to spread the drive load across ports). We recommend you
simply install the drives sequentially. The peak transfer rate is 3 Gbps per port or
12 Gbps per SAS expander; however, it is more likely that the peak drive transfer
rate will be less than this value.
All three servers support hot-swap drives. The number and type of drive bays in
each server are as follows:
x260: Six 3.5” hot-swap bays standard with another six optional
x366: Six 2.5” hot-swap bays
x460 and MXE-460: Six 2.5” hot-swap bays
For details on supported drives, see 3.5.3, “Internal storage” on page 124.
When the optional ServeRAID-8i is installed, you can see from the above
diagrams that the flow of data still goes through the SAS controller. With the
ServeRAID-8i installed, the adapter controls the RAID data layout and instructs
the controller to write the data to the requested drives. For details about the
ServeRAID-8i, see 3.5.2, “ServeRAID-8i” on page 123.
64-bit path to
66 MHz PCI-X
bridge
Optional
ServeRAID-8i
Ports 1-4
Ports 5-8
Six bays
SAS drive
SAS
expander
Six bays
AIC-9410 SAS
controller
(SAS initiator)
SAS
expander
6-slot
backplane
(standard)
6-slot
backplane
(optional)
SAS drive
SAS drive
SAS drive

34
Planning and Installing the IBM Eserver X3 Architecture Servers
For drivers, if you do not have the ServeRAID-8i installed, then use the SAS
driver. If you have a ServeRAID-8i installed, then use these ServeRAID-8i driver.
Both are available from the xSeries driver matrix:
http://www.pc.ibm.com/support?page=MIGR-4JTS2T
1.12 PCI-X slots
As shown in Figure 1-17, there are six full-length hot-swap PCI-X 2.0 slots
internal to the x260 and all are vacant in the standard models. The six slots all
support hot-plug PCI-X 3.3 V, 32-bit or 64-bit PCI and PCI-X 2.0 adapters.
Figure 1-17 PCI-X slots
All six slots have the following characteristics:
Separate bus from the other slots and devices. This means the speed of the
adapter does not affect the other adapters.
PCI-X 2.0
266 MHz, each supporting lower speed adapters
64-bit, each supporting 32-bit adapters as well
3.3 V
PCI-X slot 6
266 MHz 64-bit
PCI-X slot 5
266 MHz 64-bit
PCI-X slot 4
266 MHz 64-bit
PCI-X slot 3
266 MHz 64-bit
PCI-X slot 2
266 MHz 64-bit
PCI slot 1
266 MHz 64-bit

Chapter 1. Technical overview
35
One additional dedicated PCI slot is reserved for the ServeRAID-8i adapter.
The PCI subsystem also supplies these I/O devices:
Adaptec AIC-9410 Serial-attached SCSI (SAS) controller
Broadcom dual port 5704 10/100/1000 Ethernet
ATI 7000-M video controller with 16 MB video memory
Three USB ports, one on the front panel, two on the rear
Remote Supervisor Adapter II SlimLine adapter (optional in a dedicated
socket on the I/O board)
EIDE interface for the CD-ROM drive
Serial port
Here is some further configuration information:
Video adapters are not supported.
The PCI slots support 3.3V adapters only. 5V adapters, such as the Adaptec
2944UW, are not supported.
The system scans PCI-X slots to assign system resources. The system
attempts to start the first device found, with the search order as follows:
a.DVD-ROM.
b.Integrated dual Gigabit Ethernet controller.
c.Integrated SAS devices.
d.Internal PCI and PCI-X slots (in the order 1, 2, 3, 4, 5, and 6).
e.If the x460 is attached to a MXE-460 Modular Expansion Enclosure, the
ordering continues as 7, 8, 9, 10, 11, 12, and so forth.
Note: There is no parallel port on the x260. For parallel port connections, use
the NetVista USB Parallel Printer Cable, part number 19K4164.
Restriction: The x460 does not support connectivity to an external PCI slot
enclosure such as the RXE-100.

36
Planning and Installing the IBM Eserver X3 Architecture Servers
1.13 Remote Supervisor Adapter II SlimLine
The x460 and MXE-460 both have the Remote Supervisor Adapter II SlimLine
service processor (shown in Figure 1-18 on page 36) as a standard component,
while it is an optional upgrade in the x260 and x366. This adapter is installed in a
dedicated slot, and provides the similar functionality as the Remote Supervisor
Adapter II PCI option available for other xSeries servers. The processor can be
configured using the service processor Web interface.
Figure 1-18 Remote Supervisor Adapter II SlimLine
Key features of the Remote Supervisor Adapter II SlimLine include:
IBM ASIC with an integrated PowerPC® 405 core executing at 200 MHz
16 MB SDRAM and 4 MB flash ROM
System-independent graphical console redirection
– Built-in video compression hardware eliminates drivers.
– Graphics response, up to five times faster than with the original RSA,
makes monitoring and control more efficient.
– System-independent installation eliminates the need to install service
processor drivers, helps save IT staff time, and reduces installation
complexity.
Remote control with remote diskette and CD-ROM drive support
– Enables remote booting and software loading of the server for application
and operating system installation and updates.
– Performs configuration remotely, helping to save IT time and money by
reducing on-site presence and server downtime.
Scriptable command-line interface and text-based serial console redirect
– Command-line interface supports program control of server management
functions using scripts.

Chapter 1. Technical overview
37
– Serial text redirect provides access to text-mode BIOS and text-based
system consoles, such as Linux, NetWare, and Windows EMS
(Emergency Management Services).
– Program control of text-based console uses scripts.
– Uses point-to-point protocol support.
User authentication and authority features
– User IDs, passwords, and login permission attributes can be stored in an
LDAP server.
– Enhanced user authority levels set the access rights for users to match job
responsibilities for managing your xSeries servers.
– Secure Sockets Layer (SSL) encrypts the data transmitted between the
LDAP server and the Remote Supervisor Adapter II.
Investment protection
– Integrates with IBM Director and Director Agent.
1.14 Baseboard Management Controller
The Baseboard Management Controller (BMC) is a small, independent
micro-controller used to perform low-level system monitoring and control
functions, as well as remote IPMI interface functions. It uses multiple I2C bus
connections to communicate out-of-band with other onboard devices. The BMC
provides environmental monitoring for the server. If environmental conditions
exceed thresholds or if system components fail, the BMC lights the light path
diagnostic LEDs to help you diagnose the problem and also records the error in
the BMC system event log.
The BMC functions are as follows:
Initial system check at A/C on
The BMC monitors critical I2C devices in standby power mode to determine if
the system configuration is safe for power on.
BMC Event log maintenance
The BMC maintains and updates an IPMI-specified event log in non-volatile
storage. Critical system information is recorded and made available for
external viewing.
System power state tracking
The BMC monitors the system power state and logs transitions into the
system event log.

38
Planning and Installing the IBM Eserver X3 Architecture Servers
System initialization
The BMC has I2C access to certain system components that might require
initialization before power-up.
System software state tracking
The BMC monitors the system and reports when the BIOS and POST phases
are complete and the operating system has booted.
System event monitoring
During runtime, the BMC continually monitors critical system items, such as
fans, power supplies, temperatures, and voltages. The system status is
logged and reported to the service processor, if present.
System fan speed control
The BMC monitors system temperatures and adjusts the fan speed
accordingly.
The BMC also provides the following remote server management capabilities
through the OSA SMBridge management utility program (see 5.5, “OSA
SMBridge utility” on page 194 for details).
1.15 Light path diagnostics and PFA
To limit the client’s need to slide the server out of the rack to diagnose problems,
a light path diagnostics panel is located at the front of the x260, x366, and x460.
This panel slides out from the front of the server so the client can view all light
path diagnostics-monitored server subsystems. In the event that maintenance is
required, the client can slide the server out of the rack and, using the LEDs, find
the failed or failing component.

Chapter 1. Technical overview
39
Light path diagnostics can monitor and report on the health of microprocessors,
main memory, hard disk drives, PCI-X and PCI adapters, fans, power supplies,
VRMs, and the internal system temperature (see Figure 1-19).
Figure 1-19 Light path diagnostic panel
PFA is a mechanism developed by IBM that periodically measures critical
components, such as hard drives, power supplies, fans, processors, and
memory to reduce unscheduled system downtime.
If a predefined threshold of one of these components is exceeded, for example,
the number of single bit errors on a memory module, a PFA alert will be issued.
This enables you to replace the component before a failure actually occurs.
In case a PFA alert is raised, the Error LED on the light path diagnostic panel will
be lit and the RSA logs an event. The RSA can also be configured to forward
alerts via e-mail or to an IBM Director server.
Power-control
button
Power-on LED
USB connector
Hard disk
drive activity
Information
LED
System
error
Locator LED
Release latch - slide to the left
and pull out to display the light
path diagnostics panel
DASD
NMI
PCI
BRD
I/O
BRD
PS
NONRED
SP
LINK
RAID
MEM
CPU
CPU
BRD
FAN
PCI
VRM
OVER SPEC
TEMP
REMIND
LOG
Light Path
Diagnostics

40
Planning and Installing the IBM Eserver X3 Architecture Servers

© Copyright IBM Corp. 2006. All rights reserved.
41
Chapter 2.
Positioning
IBM Eserver X3 Architecture, the third generation of the IBM Enterprise
X-Architecture technology, delivers high performance with 64-bit memory
addressability, outstanding availability, and manageability required for the next
generation of industry standard servers.
This chapter positions the servers in the marketplace and describes the target
applications for them. Topics covered are:
2.1, “Positioning the servers” on page 42
2.2, “Focus market segments and target applications” on page 52
2.3, “Server consolidation” on page 56
2.4, “XpandOnDemand” on page 64
2.5, “Positioning the x460 and the MXE-460” on page 66
2.6, “Application scalability” on page 68
2.7, “Scale-up versus scale-out” on page 82
2

42
Planning and Installing the IBM Eserver X3 Architecture Servers
2.1 Positioning the servers
X3 Architecture delivers a formidable combination of 64-bit performance,
mission-critical availability, unmatched modular scalability, and investment
protection not previously available in the industry-standard x86 server market.
With extensive high-performance server chipset development experience, IBM is
uniquely positioned to offer a robust and powerful server, offering innovation that
delivers real business IT results.
IBM has invested over $100M bringing the new chip set and a completely
refreshed set of products to market (the xSeries 260, xSeries 366, and the
flagship xSeries 460), reinforcing the future of 64-bit x86 computing and the need
for single system image scalability (known as
scale-up
).
X3 architecture servers were specifically designed for the demanding
commercial transaction processing marketplace where clients value fast
performance and rock-solid reliability. IBM engineers expanded on the XA-32
chipset designs of the x440 (first generation) and x445 (second generation) to
produce a highly scalable design, as shown in Table 2-1.
Table 2-1
Enterprise X-Architecture generations
Item XA-32
First generation
XA-32
Second generation
XA-64e
Third generation
First available 2002 2003 2005
Servers x360, x440 x365, x445 x260, x366, x460
Processors Xeon MP, Xeon DP
Clock speeds to 2.0 GHz
Xeon MP, Xeon DP
Clock speeds to 3.0 GHz
EM64T 64-bit
Clock speeds to 3.66 GHz
Dual core processors
Cache L2 512 KB 512 KB 1 MB or 2 MB per core
Cache L3 512 KB - 2 MB 1 - 4 MB 0 MB (x366, x260, dual core)
4 or 8 MB (single-core x460)
Cache L4 32 MB XceL4 64 MB XceL4
12% lower latency
256 MB XceL4v cache for
multi-node only (x460)
Front side
bus
Single-bus 32-bit
Architecture, 400 MHz
Single-bus 32-bit
Architecture, 400 MHz
Dual-bus x86-64 Architecture,
667 MHz

Chapter 2. Positioning
43
These servers are designed for the demands of the application and database
serving tiers offering leadership performance and reliability to power enterprise
workloads such as ERP, CRM, and in-house developed applications, including
Web services implementations.
The X3 Architecture offers investment protection for both software and hardware
to meet the demands of this changing marketplace. With 32-bit compatibility on a
64-bit platform, the X3 Architecture allows you to migrate according to your
business needs. The servers supports both 32-bit and 64-bit operating systems
and applications.
And now with the addition of dual-core processors and an upgrade path for
existing x366 and x460 clients, these servers are an excellent platform for
business growth.
With the xSeries 460, IBM XpandOnDemand scalability combined with XceL4v
Dynamic Server Cache and the latest 64-bit Intel Xeon MP processor, the X3
servers offer unmatched levels of flexibility and expandability, up to 32
processors and 512 GB of RAM.
As the third-generation of Enterprise X-Architecture, the IBM ^ X3
Architecture powers servers with the latest performance technologies to reduce
latency and increase bandwidth for your commercial x86 applications. First to
market with each of these high-performance server advancements, X3
Architecture servers feature Active PCI-X 2.0 up to 266 MHz, DDR2-based
Active Memory, and highly reliable Serial Attached SCSI (SAS) hard drives.
Memory 6.4 GBps peak
bandwidth
Memory Mirroring
ChipKill
Memory ProteXion
6.4 GBps peak
bandwidth
Memory Mirroring
ChipKill
Memory ProteXion
Hot-swap memory
40% more aggregate
I/O
21.3 GBps peak
bandwidth
Memory Mirroring
ChipKill
Memory ProteXion
Hot-swap memory
Hot-add memory
Support for EM64T
x86 64-bit Extensions
PCI bridge
I/O controller
PCI-X 66 -133 MHz
slots
Remote I/O Support
PCI-X 66 -133 MHz
slots
Remote I/O Support
PCI-X 2.0 support
All slots 266 MHz
Item XA-32
First generation
XA-32
Second generation
XA-64e
Third generation

44
Planning and Installing the IBM Eserver X3 Architecture Servers
Figure 2-1 shows the architectural differenences between the x445 and the x460
servers and the improvements in bandwidth
Figure 2-1 EXA2 versus EXA3 chipset comparison
The features and benefits of the IBM ^ X3 Architecture include the
following:
Scalability
– Highly modular: incremental CPU, I/O, and Memory.
– Building block architecture.
– Optimized for virtualization.
– Pay-as-you-grow.
Performance
– 4-way x86 Performance.
– 8-way to 32-way scalable performance with the x460 and MXE-460.
– 32-bit and 64-bit,
Intel EM64T.
– Dual core models and upgrade options.
– Active PCI-X 2.0.
IBM XA-64e third
generation chipset
CPU 2
CPU 3
DDR2
SMI2
DDR2
SMI2
DDR2
SMI2
DDR2
SMI2
CPU 1
CPU 4
Memory
controller
("Hurricane")
Calgary
PCI-X bridge
Scalability
ports
PCI-X bridge
xSeries 460
with IBM XA-64e third generation chipset
Scalability
ports
IBM XA-32 second
generation chipset
64 MB
L4 cache
CPU 1
CPU 2
CPU 3
CPU 4
Processor &
cache controller
"Twister"
Memory
controller
8x DDR
8x DDR
PCI bridge
PCI bridge
xSeries 445
with IBM XA-32 second generation chipset
3X
5X
2X
3X

Chapter 2. Positioning
45
Availability
– Active Memory (mirroring, hot-swap + Memory ProteXion).
– Light path diagnostics.
– PFA with IBM Director.
– Redundant hardware components.
Manageability
– Remote deployment.
– Multi-chassis partitioning.
– Integrated hardware monitoring.
– Integrated security.
Competitive advantages
– The x366 and x260 offer higher performance than competitive
Potomac-based servers using lower priced Cranford processors.
– The industry’s only greater than 4-socket EM64T server.
– The industry’s highest performing x86-64 4-way, 8-way, 16-way, and
32-way.
– Three years and over 80 number 1 Benchmarks! Enterprise
X-Architecture has delivered more number 1 benchmarks than any other
industry-standard server architecture:
• x360: 15 number one benchmarks
• x440: 35 number one benchmarks
• x365: 11 number one benchmarks
• x445: 21 number one benchmarks
• x366: Six number one benchmarks
The following list will help you decide which of the three X3 Architecture servers
to choose:
If you plan to implement a system with eight or more processors, the only
choice is the x460. Similarly, if your application can benefit from more than 64
GB of RAM, then choose a multi-node x460 configuration.
If you plan to implement a 4-way system, and do not plan to scale-up further
than 8-way, then the x366 offers better price/performance than the x460 in
the 4-way space
If you are planning 4-way or fewer processors, and a priority is rack density,
then choose the x366 over the x260. Conversely, if you require more internal
disk drivers or an internal tape drive, then choose the x260.

46
Planning and Installing the IBM Eserver X3 Architecture Servers
2.1.1 x366 versus x360
The x366 is capable of expanding to four CPUs and offers many new features
over its predecessor, the x360. It is based on the IBM XA-64e chipset and utilizes
new Xeon MP processors from Intel, which incorporate Intel EM64T 64-bit
extensions, making this both an excellent 32-bit or 64-bit platform as well as a
great environment to migrate from 32-bit to 64-bit.
The Active memory subsystem of the x366 is the most highly available memory
subsystem in this level of server in the market today. In addition to incorporating
both ECC and Chipkill memory, the x366 also incorporates memory mirroring,
hot-swap memory, and Memory ProteXion, an IBM exclusive technology.
Memory ProteXion provides a highly available memory configuration without
having to resort to the significant expense associated with a mirrored memory
configuration.
Memory on the x366 scales all the way from 2 GB to 64 GB, which is double that
of the previous generation. All six I/O slots have been enhanced to support
adapters up to 266 MHz. Neither the optional ServeRAID 8i or RSA II cards
occupy PCI-X slots, since they connect via dedicated sockets on the system
planar.
The x366 also incorporates the new serial attach SCSI (SAS) technology, with
six 2.5” hot-swap disk bays available to accommodate disk expansion all the way
up to 440 GB of internal disk storage. Both integrated dual Gigabit Ethernet and
a easily serviced 3U mechanical package carry over from the previous
generation of the 4-way rack product.
Table 2-2 compares the x366 with the x360.
Table 2-2 Major differences between x366 and x365:
Component x365 x366
Chipset XA-32 second generation chipset XA-64e third generation chipset
SMP 1-way to 4-way SMP 1-way to 4-way SMP
single-core or dual-core processors
Processors Xeon MP up to 3.0/4 MB Xeon MP single-core to 3.66 GHz/1 MB
Xeon dual-core to 3.00 GHz/4 MB
Front side bus Single-bus 32-bit Architecture, 400
MHz FSB
Dual-bus x86-64 Architecture
667 MHz FSB
Cache L2: 512 MB integrated on the CPU
L3: 1-4 MB integrated on the CPU
L4: None
L2: Integrated on the CPU
L3: None
L4: None (XceL4v acts as snoop filter)

Chapter 2. Positioning
47
Memory 32 GB max memory
2-way interleaving
16 max DIMM slots (8 standard)
DDR SDRAM PC2100
6.4 GBps
64GB max memory
2-way interleaving
16 max DIMM slots (4 standard)
DDR2 SDRAM PC2-3200
21.3 GBps memory bandwidth
Active Memory
features
ECC
Memory ProteXion
Chipkill
Memory mirroring
Hot-swap
ECC
Memory ProteXion
Chipkill
Memory mirroring
Hot-swap
Hot-add
SCSI LSI Ultra320 SCSI
Integrated RAID-1
Adaptec Serial Attached SCSI (SAS)
Optional RAID with ServeRAID-8i
Storage Six hot-swap 3.5” bays
Max internal storage = 876 GB
Six hot-swap 2.5” bays
Max internal storage = 440 GB
PCI-X slots Active PCI-X: six internal
4 @ 133 MHz
1 @ 100 MHz
1 @ 33 MHz (RSA adapter)
RXE-100 for 6-12 additional slots
Active PCI-X 2.0: six internal
All @ 266 MHz
No support for the RXE-100
Video ATI Radeon ATI Radeon 16 MB
Optical drive 24X CD-ROM 8X DVD-ROM
Power supplies 2 x 950W hot-swap standard
(only one standard in some models)
One 1300W hot-swap
One additional PSU optional
Ethernet Broadcom 5704 dual port Gigabit Broadcom 5704 dual port Gigabit
Management Remote Supervisor Adapter II standard BMC controller standard
RSA II SlimLine optional
Warranty 3-year next business day 9x5 warranty 3-year next business day 9x5 warranty
Rack height 3U 3U
Component x365 x366

48
Planning and Installing the IBM Eserver X3 Architecture Servers
2.1.2 x460 versus x445
The x460 is second of the new X3 products announced and it too adds many
new features over its predecessor, the x445. Like the x366, it is based on the
IBM XA-64e chipset and also utilizes the new Xeon MP processors from Intel,
which incorporate Intel’s EM64T 64-bit extensions. The x460 also supports
dual-core processors.
Each x460 supports up to four processors compared with the eight processors of
the x445. The x460 supports up to eight nodes connected together to form a
single 32-way complex.
The x460 uses the same Active memory subsystem as the x366, which are the
most highly available memory subsystem in this level of server in the market
today. It also incorporates both ECC and Chipkill memory, as well as memory
mirroring, hot-add memory, and Memory ProteXion.
The memory on the x460 scales to 64 GB in a single chassis. This scales up to
512 GB in a 32-way complex (eight nodes). More and more application providers
and clients alike are looking for larger amounts of memory so they can load the
entire application into memory to gain the highest possible performance. The
peak memory bandwidth of the x460 is over 21 GBps, up from 6.4 GBps in the
x445.
All six PCI-X 2.0 slots have been enhanced to 266 MHz. The standard RSA II
SlimLine card and the optional ServeRAID-8i card have dedicated connectors
and do not consume one of these PCI-X slots. As you add additional chassis to a
base x460, you can take advantage of the additional I/O slots that each chassis
provides up to a maximum of 48 slots across the maximum of eight chassis.
Because of this significant I/O support, the RXE-100 that was used in the
previous generation of the scalable product is no longer be supported.
The x460 also incorporates the new serial attach SCSI (SAS) technology, with
six 2.5” hot-swap disk bays available to accommodate disk capacity up to 440
GB per node. As with the I/O slots, the disk bays of each additional chassis can
be used as you scale-up a base x460 configuration by adding more chassis.
Dual-port Gigabit Ethernet carries over from the predecessor product, but the
older 4U mechanical package has been replaced with a denser and more
modular 3U mechanical package that has excellent serviceability features.
Table 2-3 on page 49 shows the major differences between the x460 and the
x445.

Chapter 2. Positioning
49
Table 2-3 Major differences between x460 and x445
Component x445 x460
Chipset XA-32 second generation chipset XA-64e third generation chipset
SMP Xeon MP: 2-way to 16-way
Xeon DP: 2-way to 4-way
2-way to 32-way
Processors Intel Xeon MP to 3.0/4M
Maximum eight per chassis
Intel Xeon DP 3.0GHz up to 4-way;
maximum of four per chassis
Xeon MP to 3.33 GHz 8 MB
Dual-core processors to 3.00 GHz
Maximum of four per chassis
Front side bus 400 MHz front side bus 667 MHz front side bus
Cache L3: 1-4 MB
L4: 64 MB XceL4 per 4-way (dedicated
cache memory)
L3: 4-8 MB on single-core processors
L4: 256 MB XceL4v per 4-way (uses
main memory)
Memory 64 GB maximum (32 sockets)
16-way maximum
DDR SDRAM PC2100
2-way Interleaving
6.4 GBps memory bandwidth
64 GB max per chassis (16 sockets)
32-way maximum: 512 GB
DDR2 SDRAM PC2-3200
2-way Interleaving
21.3 GBps memory bandwidth
Active Memory
features
ECC
Memory ProteXion
Chipkill
Memory mirroring
Hot-swap (top CEC only)
Hot-add (top CEC only)
ECC
Memory ProteXion
Chipkill
Memory mirroring
Hot-swap
Hot-add
SCSI LSI Ultra320 SCSI; supports RAID-1
Optional ServeRAID
Adaptec Serial Attached SCSI (SAS)
Optional RAID with ServeRAID-8i
Storage Two hot-swap 3.5” bays
Maximum internal storage = 292 GB
Six hot-swap 2.5” bays
Maximum internal storage = 440 GB
PCI-X slots Active PCI-X: six internal
2 @ 133 MHz
2 @ 100 MHz
2@ 66 MHz
RXE-100 for 6-12 additional slots
Active PCI-X 2.0: six internal
All @ 266 MHz
No support for the RXE-100
Video ATI Rage XL 8 MB ATI Radeon 16 MB
Serial port One, using the supplied cable and
PCI-slot bracket assembly (consumes a
PCI slot)
One serial port

50
Planning and Installing the IBM Eserver X3 Architecture Servers
2.1.3 x260 versus x255
The x260 is the third of the new X3 products and offers many new features over
its predecessor, the x255. The x260 is based on the IBM XA-64e chipset and like
the x366, uses the Intel Xeon MP “Cranford” processors, which incorporate the
EM64T 64-bit extensions. These extensions allow the addressing of much higher
amount of physical memory and the execution of 64-bit instructions and 32-bit
instructions simultaneously, making this both an excellent 32 or 64-bit platform
as well as a great environment to migrate from 32-bit to 64-bit.
The x260 uses the same Active Memory subsystem as the x366, which is the
most highly available memory subsystem in this level of server in the market
today. It also incorporates both ECC and Chipkill memory as well as memory
mirroring and Memory ProteXion.
Memory on the x260 scales all the way from 2 GB to 64 GB, which is more than
double that of the previous generation. All six I/O slots have been enhanced to
266 MHz and remain available, even if you install a ServeRAID 8i or RSA II card;
you do not need to utilize card slots to add these capabilities.
The x260 also implements Serial Attach SCSI (SAS) technology, and the x260
can have up to 12 internal 3.5” disk drives attached. The ability to insert six
drives is standard and a further six can be installed with the optional 6-bay
backplane. The x260 also supports either one or two internal tape backup
devices through its 5.25” drive bays.
Both integrated dual Gigabit Ethernet and an easily serviced 7U mechanical
package that is available in either tower or rack-mount versions carries over from
the previous generation of the 4-way tower product. A tower-to-rack conversion
kit is also available to convert an existing x260 tower model to be installable in a
rack.
Removable
media
24X-8X DVD-ROM
Diskette drive
8X DVD-ROM
No diskette drive
Power supplies Two 1200W hot-swap power supplies Two 1300W hot-swap power supplies
Ethernet Broadcom 5704 dual port GbE Broadcom 5704 dual port GbE
Service
processor
RSA II for EXA standard BMC integrated
RSA II SlimLine standard
Warranty 3-year next business day 9x5 warranty 3-year next business day 9x5 warranty
Rack height 4U 3U
Component x445 x460

Chapter 2. Positioning
51
The remote Supervisor Adapter II card has been re-designed as a SlimLine
option and occupies a slot on the Super I/O card, as opposed to requiring one of
the PCI-X slots on the I/O planar board.
Table 2-4 shows the major differences between the x260 and the x255.
Table 2-4 Major differences between x260 and x255
Component x255 x260
Chipset ServerWorks GC-HE chipset XA-64e third generation chipset
SMP 1-way to 4-way 1-way to 4-way
Processors Xeon MP: 2.0/1M, 2.2/2M, 2.7/2M,
3.0/4M
Xeon MP: 3.16/1M or 3.6/1M L2
Front side bus 400 MHz 667 MHz front side bus
Memory 24 GB max per chassis
12 DIMM sockets (all standard)
DDR PC1600 SDRAM
2-way or 4-way interleaved
64 GB max per chassis
16 DIMM sockets max, 4 standard
DDR2 SDRAM PC2-3200
2-way Interleaving
21.3 GBps memory bandwidth
Active Memory
features
ECC
Chipkill
Memory mirroring
Hot-spare memory
ECC
Memory ProteXion
Chipkill
Memory mirroring
Hot-swap
Hot-add
Cache L2: 512 KB integrated on the CPU
L3: 1-2 MB integrated on the CPU
L4: None
L2: 1 MB integrated on the CPU
L3: None
L4: None (XceL4v acts as a snoop filter
only)
SCSI Dual-channel Adaptec Ultra160 SCSI
Optional RAID with ServeRAID
Adaptec Serial Attached SCSI (SAS)
Optional RAID with ServeRAID-8i
Storage Six hot-swap 3.5” bays standard
Six additional bays with optional
backplane
Maximum internal storage = 1.7 TB
Six hot-swap 3.5” bays standard
Six additional bays with optional
backplane
Maximum internal storage = 876 GB
PCI-X slots Active PCI-X: Seven internal
6 @ 100 MHz
1 @ 33 MHz for RSA II card
Active PCI-X 2.0: six internal
All @ 266 MHz
No support for the RXE-100
Video ATI Rage XL 8 MB ATI Radeon 16 MB

52
Planning and Installing the IBM Eserver X3 Architecture Servers
2.2 Focus market segments and target applications
The X3 Architecture servers from IBM are designed for the demands of the
application and database serving tiers offering leadership performance and the
proven reliability of the Intel Xeon MP processor architecture to power
mission-critical stateful workloads, such as:
Enterprise Resource Planning (ERP)
ERP is an industry term for the broad set of activities supported by
multi-module application software that helps a manufacturer or other
companies to manage the important parts of its business, including product
planning, parts purchasing, maintaining inventories, interacting with suppliers,
providing customer service, and tracking orders. ERP can also include
application modules for the finance and human resources aspects of a
business. Typically, an ERP system uses or is integrated with a relational
database system.
These applications today use a Web-based infrastructure with interfaces to
suppliers, clients, and internal company employees. There are three general
architectures used by enterprise solutions:
– Four-tier architecture (often referred to as an Internet architecture) with
client systems, Web servers, application servers, and database servers
– Three-tier architecture, which includes client systems, Web/application
servers and database servers
Ethernet Single Port Broadcom 5703 Gigabit
Ethernet
Dual Port Broadcom 5704 Gigabit
Ethernet
Serial port One serial port One serial port
Removable media 48X CD-ROM
Diskette drive
Tape drive optional
40x CD-ROM
Diskette drive
Tape drive optional
Power supplies Two hot-swap 370W standard, four
maximum
Two hot-swap 775W standard, 4
maximum
Service processor Optional RSA II BMC integrated
Optional RSA II SlimLine
Warranty 3-year next business day 9x5 warranty 3-year next business day 9x5 warranty
Mechanical 7U rack or tower
available tower-to-rack conversion kit
7U rack or tower
available tower-to-rack conversion kit
Component x255 x260

Chapter 2. Positioning
53
– Two-tier architecture, which includes client systems and database servers
Key ERP software vendors are SAP (SAP Business Suite and My SAP
All-in-one), Oracle (PeopleSoft and JD Edwards), Microsoft (Axapta®) and
Baan
Customer Relationship Management (CRM)
CRM is an IT-industry term for methodologies, software, and usually Internet
capabilities that help an enterprise manage client relationships in an
organized way. The application can use a four-tier, three-tier, or two-tier
architecture similar to ERP applications.
Key CRM software vendors are Siebel, Oracle (PeopleSoft and JD Edwards),
SAP (SAP Business Suite and My SAP All-in-one), Baan, and Onyx
Supply Chain Management (SCM)
SCM is the oversight of materials, information, and finances as they move,
through a process, from supplier to manufacturer to wholesaler to retailer to
consumer. SCM involves coordinating and integrating these flows both within
and among companies. The application also can use a four-tier, three-tier, or
two-tier architecture.
Key SCM software vendors are I2, SAP (SAP Business Suite and My SAP
All-in-one), Oracle (JD Edwards and PeopleSoft) and International Business
System (IBS)
Business Intelligence (BI)
BI is a broad category of applications and technologies for gathering, storing,
analyzing, and providing access to data to help enterprise users make better
business decisions. BI applications include the activities of decision-support
systems, query and reporting, online analytical processing (OLAP), statistical
analysis, forecasting, and data mining.
Key BI software vendors are SAS, Hyperion, Cognos, Business Objects and
Crystal Decisions
Database
The X3 Architecture servers are ideal as database servers or application
servers, with their fast 4-way and above processors and their large and very
fast memory subsystems. The x460 in particular provides an extremely
scalable platform with room to scale to additional nodes. These configurations
use an external storage enclosure or SAN, depending on the size of the
database, which is driven by the number of users.
The 16-way and 32-way configuration can deliver a highly reliable and
capable platform for clients who need to run multiple instances of databases
that can scale beyond eight processors.

54
Planning and Installing the IBM Eserver X3 Architecture Servers
Key database software vendors are IBM (DB2), Microsoft (SQL server), and
Oracle
Server Consolidation
Server Consolidation or SCON is a process of centralizing business
computing workloads to reduce cost, complexity, network traffic,
management overhead and, in general, to simplify the existing IT
infrastructure and provide a foundation for new solution investment and
implementation.
Server consolidation is discussed in detail in 2.3, “Server consolidation” on
page 56
Key server consolidation software vendors are VMware (ESX Server and
GSX Server) and Microsoft (Virtual Server)
eCommerce
eCommerce is the use of Internet technologies to improve and transform key
business processes. This includes Web-enabling core processes to
strengthen customer service operations, streamlining supply chains, and
reaching existing and new clients. In order to achieve these goals, e-business
requires a highly scalable, reliable, and secure server platform.
Key software vendors are IBM (WebSphere®) and BEA.
2.2.1 x260 target applications
The x260 is targeted at the following
applications:
Database: SQL Server, DB2, and
Oracle
ERP/CRM/SCM: SAP, Siebel, and
i2
E-mail and collaboration:
Exchange and Notes
Web services: WebSphere
The x260 is the departmental workhorse of the new generation of X3 servers.
With its outstanding 64-bit performance and high availability features, the x260
server is well suited to a wide variety of different computing environments.
The x260 should be targeted to departmental and server consolidation
applications as well as ERP/CRM/SCM, e-mail, and Collaboration and Web
Services. It is also an excellent server for infrastructure applications, such as

Chapter 2. Positioning
55
departmental databases, where the application is not one that needs the
additional scalability for which the 4-way scalable offering is better suited.
2.2.2 x366 target applications
The x366 is targeted at the following applications:
Database: SQL Server, DB2, and
Oracle
ERP/CRM/SCM: SAP, Siebel, and
i2
E-mail and collaboration:
Exchange and Notes
Web services: WebSphere
Server consolidation: VMware ESX Server
With its outstanding 64-bit performance and high availability features, the x366
server is targeted at the application tier in a multi-tier computing environment.
The x366 should be targeted to applications such as ERP/CRM/SCM, e-mail and
Collaboration, Web Services, and Server Consolidation.
With the elimination of any premium for buying a 4-way over two a 2-way, the
x366 should be thought of as an upgradeable 2-way with more memory address
slots (16 versus 8). The extra DIMM capacity of the x366 is perfect for 64-bit
applications, and with twice the slots, clients can buy lower density DIMM sticks,
saving even more money. The x366 is the better mid-tier commercial application
server and offers application performance headroom with no price penalty.
2.2.3 x460 target applications
The x460 is targeted at the following
applications:
Database: SQL Server, DB2, and
Oracle
ERP/CRM/SCM: SAP, Siebel, and
i2
Server consolidation: VMware ESX
Server
With its outstanding 64-bit performance, high availability features, and scalability,
the x460 server is targeted primarily at the database tier or the high-end of the
application tier in a multi-tier computing environment.

56
Planning and Installing the IBM Eserver X3 Architecture Servers
The x460 is targeted to very demanding applications, such as ERP/CRM/SCM
and Server Consolidation. It is also an excellent server for database applications,
providing the additional scalability of up to 32 processors and 512 GB of RAM
that the x366 offering cannot.
2.3 Server consolidation
Server consolidation is the first step towards designing a more rational, efficient,
and flexible IT environment that delivers consistently high levels of services to
the entire organization. It is a lot more than simply replacing smaller boxes with
fewer bigger boxes. It encompasses not just hardware, but software, services,
and, most importantly, the systems management procedures that tie it all
together.
Server consolidation is the discipline of simplifying end-to-end IT infrastructures,
including servers, databases, applications, networks, and systems management
processes, with the goal of reducing complexities and costs and establishing a
stable foundation for growth and new solution deployment.
Over the last decade, Intel-based servers were purchased as a cost-effective
alternative to mainframes and proprietary UNIX® servers. As demand grew, both
servers and IT administrators were added to distributed departments. The
frequent addition of a small number of inexpensive servers was not viewed as a
major budget item when companies were flourishing. However, as server
numbers grew, it triggered an enormous amount of hidden costs in software
licensing, maintenance, management, and data center floor space, power, and
cooling.
An unpredictable growth of the number of Intel base servers and applications in
organization or departments happens due to number of reasons:
It is a practice to run one application on one server.
Application incompatibility.
Reliable workload management techniques for splitting physical resources of
one server by applications is missing.
Lack of applications scalability.
2.3.1 Why consolidate servers?
The main objectives of server consolidation are:
Need to make your company more responsive, more flexible, and more
resilient in the face of changing market conditions and client demands

Chapter 2. Positioning
57
Reduced IT budget for hardware and IT staff
Consolidating enterprise resources is a critical task for every organization, since
by consolidating resources, many systems can be utilized more efficiently,
maximizing resource utilization, increasing the efficiency of system management,
security, and reliability, and reducing server count and complexity.
The goals of server consolidation often include:
Reduce computing costs and complexities
Improve productivity through better management of information, data, and
resources
Improve service levels, and establish a flexible, stable foundation for growth
and new solution deployment
Reduce hardware, software costs
Better use existing facility
Reduce the maintenance, upgrade, and licensing costs
Optimize server performance
Faster deployment of applications in a more secure environment
Standardize procedures and operations
Reduce total cost of ownership
One of the most important things to remember is that there are no “off-the-shelf”
solutions for server consolidation. Every organization requires a unique solution
that will match its unique infrastructure and business model.

58
Planning and Installing the IBM Eserver X3 Architecture Servers
2.3.2 Types of server consolidation
There are four general types of server consolidation projects that clients are
implementing, offering a wide range of business value through varying degrees
of solution complexity and investment. The most effective approach for a specific
enterprise could fall into any one of these four categories, or it may require work
across all four areas:
Centralization
Physical consolidation
Data integration
Application integration
These areas are shown in Figure 2-2.
Figure 2-2 Types of server consolidation
The four types are implemented as follows:
Location centralization
Centralization is commonly called datacenter consolidation. By simply
relocating existing servers to fewer numbers of IT sites, economies of scale of
operation can provide simplified management and cost improvement.
Centralization is typically the initial step most companies take towards doing
something broader. And this is almost always where organizations begin to
Application
integration
Data
integration
Physical
consolidation
Location
centralization
Complexity and Return on Investment
Total Cost of Ownership
Integrate key applications
and data
Manage distributed IT
resources and control
spiraling IT costs
Integrate, manage,
control and secure data

Chapter 2. Positioning
59
support mergers and acquisitions, by co-locating data centers before merging
operations and business processes.
Physical consolidation
Physical consolidation is what most people think of first when they consider
server consolidation. It is the process of replacing or reducing the actual
number of servers by replacing many small servers with fewer, more powerful
servers or clustered systems.
This can take place within the same architecture or across architectural
boundaries. For example, replacing several two-way Windows servers with
one 16-way VMware ESX Server-based system with several virtual
machines, or with a few BladeCenter chassis.
This approach is typically appropriate for implementations of key packaged
applications such as SAP, PeopleSoft, and Siebel, where minimal integration
with other applications and data is required. Additional benefits can be gained
through data integration and application integration. While these are often
more complex projects that require extensive analysis, planning, and
implementation, they can provide significant return-on-investment.
Data integration
Data integration involves physically combining data from different sources
across the enterprise into a single repository or format. The result is that the
merged data can reside on fewer servers and more centralized and
consistent storage devices, greatly lowering the total costs associated with
managing the data. The data can, on file servers across disparate operating
systems, be consolidated to a single system. Also, multiple types of
databases, such as DB2, SQL Server, Oracle, and Sybase can be converged
to a single database architectures.
Application Integration
This approach involves consolidating and combining multiple applications and
data to fewer server architectures for not only reduced complexity, but also for
business process integration and automation. It allows for co-locating of
mixed workloads within an unified infrastructure so that applications can
communicate with each other and work together seamlessly.
For example, some business processes can be achieved through integration
of order entry, inventory, fulfillment, and accounting systems (Web serving,
ERP, business intelligence, and other applications) that represent the
backbone of the core business.

60
Planning and Installing the IBM Eserver X3 Architecture Servers
2.3.3 Benefits of server consolidation
There are numerous benefits to server consolidation, depending on the type of
consolidation you plan to implement:
Centralization:
– Reduction in administration costs
– Increased reliability and availability
– Lower operation costs
– Improved security and management
Physical server consolidation:
– Reduced hardware, operating system, and database types
– Reduced number of servers per application (mail and file server)
– Reduced hardware and software costs
– Improved processor utilization
– Reduced facilities costs (space, power, and A/C)
– Lower operations costs
– Improved manageability
– Improved reliability
Data integration:
– Improved resource utilization
– Reduction in administration costs
– Reduced duplicated data
– Enhanced data access and integrity
– Reduced storage management costs
– Improved backup/recovery capabilities
– Improved reliability
Application integration
– Reduction in administration costs
– Increased reliability and availability
– Reduced facilities costs (space, power, and A/C)
– Lower operation costs
– Scalability
The main benefits of server consolidation, generally, are:
Single point of control.
Rapidly growing firms, especially those growing through mergers and
acquisitions, frequently felt that disparate distributed systems were so
unwieldy to manage that they were losing control, which could constrain
further corporate growth.
A single point of control allows enterprises to:
– Reduce or eliminate department operational costs.
– Reduce some software licenses.

Chapter 2. Positioning
61
– Reduce the number of the systems and disk storage costs.
– Reduce maintenance charges.
– Avoid multiple copies of the same application on distributed systems.
– Reduce owner operational costs.
– Offer better availability of service.
– Improve systems management.
– Have better version control management.
– Have better software distribution.
– Reduce risk and increase security.
Giving users better services.
With a consolidated infrastructure, it is easier to offer 24x7 support to end
users. The response time is much better than with an overly distributed
environment, and the data is more easily accessible while being highly
protected. The control procedures are simpler, while security becomes even
higher. And information sharing is improved, giving end users increased data
consistency. The availability of service is improved mainly due to a reduction
in the time needed to communicate between clients and servers in a single
location.
Regaining flexibility.
The standardization of procedures, releases, and servers also makes it easier
to install new application software. Computing resource consolidation enables
a trouble-free upgrade of the information system and less costly adaptation to
organization or environment changes. Enterprises can react more quickly to
market changes, since storage is readily available and can easily be
reallocated.
Avoid floor space constraints.
While a small server may be easily fit into a closet, as compute demands
increase, enterprises find that suitable floor space is hard to find for
proliferating small servers. The solution is a central site outfitted with
appropriate power, cooling, access to communications links, and so on, and
populated with more powerful systems, each giving more performance in the
same footprint.
Reduction of the Total Cost of Ownership (TCO).
There are several costs associated with server consolidation, including:
– Hardware costs: New servers and infrastructure, and upgrades.
– Software costs: Fewer software licenses are required with fewer servers.
– Disruption costs: Migration and change management.

62
Planning and Installing the IBM Eserver X3 Architecture Servers
Manageability and availability
Server consolidation can help you improve manageability and availability of IT
systems in the following ways:
– Enterprise management: Integrated operations allows for consistent
management of all facilities and IT services.
– Consistent performance: Providing consistent response time at peak load
periods is very important.
– Dependability: Commonly cited problems of distributed environments
include frequency of outages and excessive requirements for manual
intervention by the IT staff. In addition, it provides the following benefits:
• It is easier to enforce consistent user policies in a consolidated
environment.
• Fewer servers lead to a simpler network structure that is easier to
manage.
• Reorganization following mergers or acquisitions is easier in a
well-controlled environment.
• Consolidation encourages standardization of tools, processes, and
technologies to provide a stable and consistent application platform.
Data access and protection.
Server consolidation can help you improve data access and protection in the
following ways:
– Network technology: The growth of networking and network speeds is
enabling the centralization of IT networks today and will continue and
expand into the future.
– Fragmentation and duplication of data: This is a core issue in most
organizations with large numbers of distributed servers.
– Physical security: Consolidation of servers in a central data center can
restrict unwanted access and ensure a more secure environment.
– Integrity, local backup, and recovery: Enterprises are concerned about the
dangers of business disruption, client lawsuits, and regulatory action in the
event of severe data loss, and they need to implement effective disaster
recovery procedures.
Use your existing investment in IT.
Server consolidation can help you leverage existing investments in the
following ways:
– Expand existing servers: Add new capabilities to the existing installation
rather than deploy new dedicated servers.

Chapter 2. Positioning
63
– Optimization of capacity utilization: In order to manage performance and
have a level of acceptable, consistent response times, enterprises
typically run servers at 50-60% utilization. Excess or underutilized capacity
on one server cannot be shared with workloads of other servers in a
distributed environment. Consolidation can help the client to use much of
the excess capacity for more efficient utilization of resources.
– Optimization of skilled resources: Under the distributed alternative,
systems management responsibilities are often only part-time, extra-duty
assignments such that a critical skill level is rarely achieved. Furthermore,
since other departments may employ disparate architectures and
applications, there is little opportunity to benefit from the experiences of
others.
Scalability and workload growth.
Server consolidation can help you handle scalability and workload growth
issues in the following ways:
– True scalability: Server consolidation provides the ability to deal with peak
usage without crashing or seriously degrading performance. It also
provides an upgrade path without degradation in response, excessively
complex forms of database partitioning, or other problems.
– Granular upgrades: Server consolidation provides the ability to quickly
grow the number of users, the number of applications, or the size of an
application when needed, without major disruptions to the current
production environment.
Service level.
Most companies spend much of their IT budget for services. They need
services for hardware, software, and infrastructure maintenance. Server
consolidation can help you to reduce the increasing service costs in the
following ways:
– Delivery of a specified service level is costly if servers are uncontrolled.
– Management of servers as corporate assets is easier when they are
centralized.
– Application deployment is quicker and easier when not spread over a
large number of servers.
– Staff time freed from server maintenance tasks can be used for activities
more directly related to business needs.

64
Planning and Installing the IBM Eserver X3 Architecture Servers
Business continuity.
Almost all enterprises need to run their business without interruption.
Business interruption can be very costly and it influences the productivity of
your business. Server consolidation can help you to run your business
without interruption in the following ways:
– Consolidating IT resources can help you ensure that critical business
information and processes are accessible and shared across the
enterprise.
– Implementing critical new solutions that may enable a competitive edge is
easier.
In most cases, developing a sound and visionary IT realignment and
consolidation strategy will not only provide IT benefits, but can also serve to
improve the economic and operational conditions of an enterprise for strategic
business achievements.
Table 2-5 shows how benefits in IT can translate into real business benefits.
Table 2-5 Business benefits
2.4 XpandOnDemand
XpandOnDemand is the term given to the ability of the x460 to scale-up as
required by adding nodes to an x460 complex. This ability ensures that clients
only pay for the computing resources they need at the present time, without
sacrificing the investment they have made in hardware and software should they
choose to upgrade.
IT benefits to consolidation Business benefits to consolidation
Reduce complexity
Lower costs and dramatically improve
TCO
Simplify systems management
Enhance resources utilization
Drive open adoption of standards end
to end
Improve uptime/availability and
increase recoverability
Improve performance and optimize
scalability
Accelerate e-business integration
Attract and retain clients more
efficiently
Maximize revenue per client
Bring new services online quickly
Respond to new business challenges
quickly by establishing a foundation
for future growth
Consolidate operations and overall
systems control
Reduce or eliminate redundancy in
infrastructure and personnel
Better service client, employees,
suppliers, and partners

Chapter 2. Positioning
65
The investment protection features of the x460 include the following:
Processors: Clients can upgrade from 2-way to 4-way, 8-way, 16-way, and
ultimately 32-way processing should they require it. Models with single-core
processors can be upgraded to dual-core processors.
Memory: The standard 2 GB can be expanded to 512 GB
PCI-X slots: The x460 has six PCI-X 2.0 slots standard, but an x460 complex
can be expanded to 48 slots if required.
Drives and USB devices. The number of internal disk drives and other internal
resources is also increased as a result of such expansions.
These “pay-as-you-grow” design options are achieved by connecting x460
servers (or x460s with MXE-460 expansion units) together to form multi-chassis
(or multi-node) complexes. Each node contains processors, memory, and PCI-X
slots and other components in the x460. By joining the nodes together, the single
operating system running on the entire complex has full access to all resources
in all attached compute nodes.
In addition, once the nodes are connected together to form a larger complex,
should the business need arise, you can divide the complex back into partitions.
These partitions are formed on node boundaries. For example, an 8-node x460
complex could be divided into a 4-node partition and two 2-node partitions, each
running its own operating system. This flexibility ensures that the computing
resources match the need of the business.
All the X3 Architecture servers support 64-bit operating systems using the Intel
Xeon MP processors with EM64T extensions. This allows you to grow more
easily as the needs of your business change over time, transitioning to 64-bit
applications or adding incremental performance capacity when you need it
without the penalty of paying for costly up-front infrastructure.
The X3 Architecture servers are also ready for the next wave of processor
technology, with the support of the Intel dual-core CPUs.
The main advantages of XpandOnDemand capability are:
Investment protection: You pay only for performance that is necessary today.
You do not need to pay more then you need now. Supports 32-bit and 64-bit
applications on the same platform. You can migrate to 64-bit as needed or
when required 64-bit versions of commercial applications will be available.
“Pay-as-you-grow”: Server performance is growing with your company. You
do not need to buy a new, more powerful server; just expand the server you
already have.

66
Planning and Installing the IBM Eserver X3 Architecture Servers
Configuration and performance flexibility: Server configuration and
performance can be simply modified to smooth peak loading or to periodically
perform some more processor intensive tasks, such as weekly accounting
and billing.
Near linear performance increase: By adding additional nodes to the xSeries
460 based on the Enterprise X-Architecture, you will get not only additional
processors and power, you will also get more memory capabilities, more
PCI-X slots, more internal storage, more front side buses, more memory
controllers with additional memory buses, and more chipsets on the
motherboard, which, in some cases, gives you practically linear performance
increase.
2.5 Positioning the x460 and the MXE-460
The MXE-460 is almost identical to the x460. The purpose of the MXE-460 is to
act as an expansion node for multi-node configurations. The MXE-460 is less
expensive than the x460 and can be used as a secondary node in partitions.
When building a multi-node configuration, you can use a combination of x460
and MXE-460 systems as the nodes. It is the number of x460s you have that
determines how you can partition the complex. For example, if you require a
8-node complex, you can do the following:
If you configure one x460 and seven MXE-460, then you can only create one
partition (a 32-way partition).
If you configure eight x460s and no MXE-460s, then you will have maximum
flexibility as to what partitions you can make, from a single 32-way partition, to
eight 4-way partitions, and combinations in between.
The technical differences between the x460 and the MXE-460 are as follows:
The x460 comes standard with processors and memory. The MXE-460 ships
standard with neither installed. Clients will need to install matching CPUs and
the appropriate amount of memory (we recommend you also match the
amount of memory installed).
The x460 has a DVD-ROM as standard. The MXE-460 has no optical drive
standard.
Tip: The key is that for every partition you wish to create, the primary node in
that partition must be an x460. Having an MXE-460 as a primary partition is
not supported.

Chapter 2. Positioning
67
Some components of the country kit are different. For example, the MXE-460
country kit does not include a ServerGuide™ CD.
The MXE-460 is not supported as a primary node in any partition.
Once you have cabled and configured a multi-node complex, you need to
partition the complex. Most clients will not create more than one single partition
in the complex. However, there are some advantages to creating multiple
partitions:
You can run different operating systems or versions on different partitions
without the need for products such as VMware ESX Server.
You can easily reconfigure the partitions if you need to perform some periodic
processor-intensive or memory-intensive tasks, such as a weekly accounting
or a Business Intelligence analysis task.
Figure 2-3 shows the configuration alternatives for 2-node and 4-node
complexes, and the partitioning options available to you as a result. Note that the
8-node configuration is not shown in the diagram, but the same rules apply.
If you want maximum partitioning flexibility, all your nodes should be x460
systems.
Figure 2-3 Supported partitioning for 2-node and 4-node complexes
x460
x460
Partition options:
1x 8-way
2x 4-way
x460
MXE-460
Partition options:
1x 8-way
x460
x460
Partition options:
1x 16-way
2x 8-way
1x 8-way & 2x 4-way
4x 4-way
x460
x460
x460
MXE-460
Partition options:
1x 16-way
MXE-460
MXE-460
x460
MXE-460
Partition options:
1x 16-way
2x 8-way
x460
MXE-460

68
Planning and Installing the IBM Eserver X3 Architecture Servers
The key rules about partitioning are as follows:
Partitioning is always at node boundaries. You cannot form, for example, a
partition using two processors in a node and two processors in another node.
The primary node for every partition must be an x460. The use of MXE-460s
limits the partitioning you can do.
All nodes in a multi-node partition must have four processors installed.
2.6 Application scalability
If you plan to increase performance of your system, there several issues that you
should consider:
Application scalability
Operating system scalability
Server scalability
Storage scalability
Adding processors can improve server performance under certain circumstances
because software instruction execution can be shared among the additional
processors. However, both the operating system and, more importantly, the
applications must be designed to take advantage of the extra processors. Merely
adding processors does not guarantee a performance benefit.
For example, not all applications can use the full power of four processors in the
one server. File and print servers often only take advantage of one or two
processors and popular mail systems typically only scale well to four processors.
Table 2-6 shows the suitability of multi-processor systems to application types.
Table 2-6 Processor scalability by application type
Processors File &
print
Web
server
E-mail
collaboration
Business
logic
Database Server
consolidation
1 way Suitable Suitable Suitable Suitable
2 way Suitable Suitable Suitable Suitable Suitable Suitable
4 way Suitable Suitable Suitable Suitable
8 way Suitable Suitable
16 way Suitable Suitable
32 way Suitable Suitable

Chapter 2. Positioning
69
Processors are only one part of the scalability story. It is typically very important
that the memory, disk, and networking subsystems are examined for the effect
on scalability. It is typical that the performance gains from adding processors are
only realized when memory is added in parallel. For disk intensive applications,
such as OLTP-type applications, it is essential to have a large disk array to
stream data to the CPU and memory subsystems so that any disk-related delays
are kept to a minimum.
It is important to plan your system in advance according to your business
requirements so that you will not have to replace your server, operating system,
or storage subsystem, because your server no longer meets your processing
needs, for example, the operating system does not support more then four
processors, or your server is not able to hold more than six PCI-X adapters.
Table 2-7 shows how application types scale and what is required to achieve
peak performance. This table lists the server configurations used to produce
some recent benchmark results. As you can see, the amount of memory and
disks varies widely depending on the application.
Table 2-7 Benchmarks configurations of processors, memory, and disk
Benchmark Processors Memory (GB) Disk drives
TPC-C (Database) 8 128 500
4 64 351
TPC-H (Decision support) 8 32 138
4 16 160
SAP SD (ERP) 16 128 14
8 16 15
Oracle OASB 11.5.6 8 32 42
PeopleSoft GL 8.4 8 8 92
Exchange MMB3 4 8 303
SPECjbb2000 (Java) 8 32 1
SPEC CPU2000 (CPU-intensive) 8 4 2

70
Planning and Installing the IBM Eserver X3 Architecture Servers
The different server configurations reflect the different workloads of the these
benchmarks. The workload that the benchmark generates causes the server to
bottleneck in a particular subsystem. To alleviate the problem, resources are
added to that subsystem until the bottleneck is moved to another subsystem.
The last row in Table 2-7 on page 69 also highlights the component-focused
nature of the SPEC benchmarks and the CPU-intensive applications they serve.
This 8-way server required only 4 GB of memory and two disks. Clearly, the
workload isolates the CPU with very little dependency on other subsystems. This
means that the benchmark may be very good for comparing raw CPU
performance, but it provides limited information regarding the performance of the
entire system. The CPUs in a system may be very fast, but performance remains
poor if the memory or I/O subsystems cannot supply data to them quickly
enough.
2.6.1 Enterprise application scaling
Enterprise applications enable you to run your business more effectively and are
often referred to as back-office applications. They bring together four major
application groups to create integrated end-to-end solutions:
Business Intelligence
Customer Relationship Management
Enterprise Resource Planning
Supply Chain Management
Enterprise applications work with your most critical business data so it is
important that these applications are highly available and secure. There are
three general architectures used by these applications, as shown in Figure 2-4
on page 71:
A three-tier architecture (often referred to as an Internet architecture) includes
client systems, Web servers, application servers, and database servers.
A two-tier architecture includes client systems, Web servers, and
database/application servers.
A three-in-one tier architecture includes client systems and database servers.

Chapter 2. Positioning
71
Figure 2-4 Enterprise solution architectures
While three-tier architecture has far greater complexity, it also allows for greater
scalability. The architecture selected for a solution will depend on your business
needs, the type of application deployed, and the number of planned users. In
most cases, if you need to scale your applications, use a two-tier or three-tier
architecture. Smaller clients may prefer to implement a “three-in-one”
implementation, simply because it is easier to manage and the number of users
supported can be handled by the “three-in-one” solution.
See also 2.2, “Focus market segments and target applications” on page 52.
2.6.2 SAP product offering
SAP is the largest supplier of integrated business solutions in the world. SAP is
traditionally seen as an ERP vendor. Today, through its suite of modular
mySAP.com solution offerings, SAP can provide integrated end-to-end solution
applications that span ERP, CRM, SCM, and BI. In addition, SAP offers
industry-specific solutions based on mySAP.com.
Three-tier
Two-tier
Three-in-one
Web servers
Web servers
Application servers Database servers
Application and
Database servers
Web, Application and
Database servers
Clients
Clients
Clients

72
Planning and Installing the IBM Eserver X3 Architecture Servers
SAP offerings are:
SAP Business One. For clients with 10-250 employees. For wholesale
distribution and retail industries.
mySAP All-in-One. For clients with 10-100 employees. For these industries:
– Fabrication and Assembly
– Electronics and Electrical
– Automotive
• Food and Beverage
– Wholesale Distribution
mySAP Business Suite. For high-end SMB and ISU clients in all industries.
Typical SAP solutions are usually implemented as two-tier or three-tier
configurations, as shown in Figure 2-4 on page 71.
Two-tier
– Simple network
– Simple to upgrade
– Simple and inexpensive to manage
– Easy transition to Three-Tier configuration
Three-tier architecture
– Nearly unlimited scalability
– Improved availability and multiple networks
You can use the xSeries product selection criteria shown in Table 2-8 for SAP
Solutions.
Table 2-8 xSeries product positioning for SAP solutions
xSeries Suggested fit within SAP solution
x460 DB Server in large 3-tier installations with external storage
Large 2-tier installations with external storage
Powerful application server in front of a pSeries® or zSeries®
database in a 3-tier deployment
x366 Medium to large 3-tier installation as back-end database or
application server
Medium 2-tier installation with external storage
x260 Small to medium 2-tier installation
Database server within a medium 3-tier DB server
Test and development platform

Chapter 2. Positioning
73
For information about SAP solutions including sizing, refer to the following Web
Pages:
IBM and SAP Home Page
http://www.ibm.com/solutions/sap/us/en/
IBM hardware sizing portal
http://www.ibm.com/servers/solutions/finder/portal/hw-sizing.html
IBM Sizing and Planning Questionnaire for SAP Solutions
http://www.ibm.com/support/techdocs/atsmastr.nsf/PubAllNum/PRS261
SAP Benchmark results
http://www.sap.com/solutions/benchmark
For IBM employees, the following links are relevant resources:
IBM internal SAP portal on the Sales Support Advisor
http://w3.ncs.ibm.com/solutions/SAP/
IBM SAP International Competence Center
http://w3.ncs.ibm.com/solution.nsf/SOL/BTIY-5PBJ73
2.6.3 Oracle/J.D. Edwards product offering
The J.D. Edwards product on the Intel platform is EnterpriseOne, which is a suite
of integrated industry-specific business application modules:
Asset life cycle management
x346/x236 Application server within a medium 3-tier implementation
Database server within small to medium 3-tier implementation
Citrix or Terminal Server
CRM Comm Station or J2EE™ application Server
Test and development platform
Small 2-tier installation (75 users)
x336 Application server within a medium 3-tier SAP implementation
Web server
Citrix or Terminal Server
ITS or J2EE application server
CRM Comm Stations or CRM repository server
xSeries Suggested fit within SAP solution

74
Planning and Installing the IBM Eserver X3 Architecture Servers
Customer Relationship Management (CRM)
Financial management
Human Capital Management
Manufacturing and supply chain management (SCM)
Project management
Supply management
EnterpriseOne tools and technology
Typical EnterpriseOne solutions are based on the following architectures, as
shown in Figure 2-4 on page 71:
All-in-one, where the application server, application logic, batch workload,
and database server all run on one server.
– This is a perfect solution for the majority of J.D. Edwards clients, from
small to mid-market.
– Consolidated and simplified IT architecture.
– The x460 is ideally suited to run all of J.D. Edwards’ server functions on
one server.
Three-tier configuration, when DB and application servers are separate
servers.
EnterpriseOne supports the following
Databases:
– IBM DB2 UDB
– Microsoft SQL Server
– Oracle
Operating systems:
– Microsoft Windows
– Red Hat Linux
A typical J.D. Edwards solution includes:
Citrix for Windows client over WAN
DB2 UDB
WebSphere Application Server
WebSphere Portal Server
You can use the xSeries product selection criteria shown in Table 2-9 on page 75
for J.D. Edwards EnterpriseOne solutions.

Chapter 2. Positioning
75
Table 2-9 xSeries Product Positioning for J.D. Edwards Solutions
IBM has built a strong foundation and alliance with J.D. Edwards based on a
25-year relationship serving mid-market clients. Approximately 85% of all J.D.
Edwards clients are running on IBM ^ hardware.
2.6.4 Siebel product offering
Siebel offers solutions in the CRM arena: an integrated suite of sales, marketing
and client service applications for field sales and service, customer service,
telesales, telemarketing, third-party resellers, and internet-based e-commerce.
The core design of Siebel CRM solutions is modularity to meet the needs of all
businesses.
Siebel offers a broad range of product sets for enterprise clients and for
mid-market clients:
Business Analytics is a suite of advanced Business Intelligence applications
suitable for sales, services, marketing, finance, human resources, and supply
chain requirements.
Siebel Professional CRM Edition: The product line formerly known as Siebel
eBusiness Applications provides modular CRM functionally, allowing
xSeries Suggested Fit within J.D. Edwards Solution
x460 Powerful all-in-one solution
Excellent database server
Powerful application server in front of a pSeries or zSeries
database in a 3-tier deployment
x366 Database node or medium-sized implementation
Application server in front of x460 back-end database
Citrix or terminal server
x260 All-in-one solutions for small to medium implementations
Excellent choice for clients wanting to maximize internal
storage
x346/x236 Front-end Web server
Application server for small to medium Implementations
Citrix, terminal, deployment, or Crystal Reports server
x336 Front-end Web server
Citrix or terminal server
Crystal Reports server

76
Planning and Installing the IBM Eserver X3 Architecture Servers
organizations to choose the specific CRM capabilities needed to run their
business. This product offers an easy path to Siebel enterprise applications
and can be deployed with Siebel CRM OnDemand.
Siebel CRM OnDemand: This hosted CRM software offering is delivered via
the Web and accessible from a Web browser at a fixed price per user, per
month. Clients can deploy Siebel CRM OnDemand quickly, easily, and
affordably, without any upfront IT investments. Siebel CRM OnDemand
delivers complete sales force automation, marketing automation and
customer service functionality, built-in client analytics, virtual call center
technology, embedded CRM best practices, and world class hosting services
and support.
Typical Siebel solution based on two following architectures, as shown in
Figure 2-4 on page 71:
All-in-one, where all the application servers, Web servers, and DB servers run
on one server
Three-tier configuration, where the DB, application server, and Web server
are separate servers
You can use the xSeries product selection criteria shown in Table 2-10 for Siebel
solutions.
Table 2-10 xSeries Product Positioning for Siebel Solutions
IBM and Siebel Systems have created an International Competency Center
(ICC) in San Mateo, California, dedicated to joint solution development. We have
extended our collaborative efforts by opening and jointly staffing a support center
in Montpellier/La Gaude, France, where we validate Siebel architecture on IBM
hardware and software and showcase advanced technologies. Clients, business
xSeries Suggested Fit within Siebel Solution
x460 Excellent Siebel database server node
Powerful application server in front of a large pSeries,
zSeries, or even a 16-way x460 back-end database
x366/x260 Small/medium Siebel database server
Excellent remote Siebel server or application server
x346/x236 Remote Siebel server
Small application server
Web server
x336 Excellent fit as Web server
Gateway server

Chapter 2. Positioning
77
partners, consultants, and integrators alike are invited to observe these efforts
firsthand. Solution centers also demonstrate and provide a wide array of sales
and technical support options.
Some useful links are listed below:
IBM-Siebel portal
http://www.ibm.com/solutions/businesssolutions/siebel
IBM Sizing and Planning Questionnaire for Siebel
http://www.ibm.com/support/techdocs/atsmastr.nsf/PubAllNum/PRS593
For IBM employees, the internal IBM-Siebel portal is at:
http://w3.ncs.ibm.com/solutions/Siebel
2.6.5 Oracle/PeopleSoft product offering
PeopleSoft was originally a human resources solution. The current solution now
includes components that help link clients, suppliers, partners, and employees to
provide a competitive advantage.
PeopleSoft uses IBM ^ servers to run their own business and has
chosen DB2 as their strategic database platform. In addition, PeopleSoft has
embedded WebSphere Application Server into PeopleSoft applications. IBM can
offer the most comprehensive and integrated solution stack for PeopleSoft
ranging from hardware, middleware, financing, and services.
Together, IBM and PeopleSoft share over 6,000 clients worldwide using
PeopleSoft products. These clients are using a variety of solutions from classic
financial management tools to advanced supply chain, procurement, customer
relationship management, and fully Web-enabled offerings.
There are three product lines to address the varying needs of the market:
PeopleSoft Enterprise is targeted at large enterprises in industries, such as
financial services and health care.
PeopleSoft EnterpriseOne is focused on product-related industries, such as
construction, manufacturing and distribution, and asset intensive industries
PeopleSoft World is the leading platform for iSeries clients, especially those
who want a low risk solution, a single database, and Web enablement

78
Planning and Installing the IBM Eserver X3 Architecture Servers
Typical PeopeSoft solutions are based on an Internet application server
architecture with a “no code on the client” philosophy that allows rapid
deployment with PeopleSoft systems using standard protocols.
PeopeSoft solutions implement one of three architectures, as shown in
Figure 2-4 on page 71:
All-in-one, where all the application servers, Web servers, and database
servers run on one server
Three-tier configuration, where the database, application server, and Web
server are separate servers
Two-tier configuration, where the DB server runs on one server and the
application server and Web server are on a separate server
People Soft supports:
Databases
– DB2 zOS
– DB2 UDB
– Informix®
– SQL Server
– Oracle
– Sybase
Operating systems for the application servers
– Linux
– Windows
– AIX®
You can use the xSeries product selection criteria shown in Table 2-11 for
PeopleSoft solutions.
Table 2-11 xSeries product positioning for PeopleSoft solutions
xSeries Suggested fit for PeopleSoft solutions
x460 Excellent scalable database server node
Powerful server for an all-in-one PeopleSoft solution
Database server for 500+ users for PeopleSoft SCM solutions
or 1000 users for PeopleSoft CRM solutions
Powerful application server in front of a large-scale pSeries or
zSeries back-end database

Chapter 2. Positioning
79
2.6.6 Microsoft SQL product offering
Table 2-12 on page 80 shows the Microsoft SQL Server product offering, which
ranges from the Enterprise Edition down to the Personal edition.
The Enterprise Edition is the most scalable from both a CPU and memory
perspective. In addition, it offers some unique features, such as failover
clustering, distributed partition views, log shipping, SAN support, ROLAP, and
VL dimensions, and other advanced OLAP and analysis functions.
The Standard edition is one that is often used and targeted towards small to
medium sized businesses. It includes all of the core SQL server features
except those mentioned above and it does include analysis services.
Developer is the same as the Enterprise, but it is only for use for development
and test purposes only. In addition, you will have to download SQL Server CE
and get distribution rights for it as well.
The Personal is a subset of the Standard, which again is ideal for mobile
computers and is not available for purchase, but it ships with the Enterprise
and Standard editions.
x366/x260 x366: Small/medium PeopleSoft database server
Good as an application server in-front of an x460, pSeries, or
zSeries back-end database
x260: Excels as a test and deployment platform when a large
server is required by the client
All-in-one server when internal storage is required
x346 Powerful Web application server
Small application server
Reporting server for PeopleSoft applications
BladeCenter Can host multiple application servers in an extremely rack
dense form factor
Opportunity to have Web server, application server, and other
point solutions servers in a single chassis
x336/x236 x236: Affordable test and deployment server
x336: Excellent Web server
xSeries Suggested fit for PeopleSoft solutions

80
Planning and Installing the IBM Eserver X3 Architecture Servers
Table 2-12 Microsoft SQL Server product offering
2.6.7 DB2 Universal Database
DB2® UDB for Windows is available in several editions that can scale from
hand-held devices running Windows CE to server editions running on Windows
Server 2003, Datacenter Edition.
DB2 UDB V8.1 is certified for Microsoft Windows Server 2003 for the Standard,
Enterprise, and Datacenter editions of the operating system, both the 32-bit and
64-bit versions.
There are several server editions of DB2 UDB V8.1 that can be deployed on the
Windows Server 2003 operating system. These include:
Express Edition
Workgroup Server Edition
Workgroup Server Unlimited Edition
Enterprise Server Edition
Table 2-13 on page 81 shows the various DB2 UDB server editions.
Edition Max CPU Max RAM Unique features
Enterprise 32 64 GB Failover clustering
Distributed partitioned views
Log shipping
Parallel operation for SMP systems
SAN support
ROLAP and VL dimensions
Distributed partitioned cubes
Other advanced OLAP/analysis
Standard 4 2 GB All Core SQL Servers features
except those listed above
Includes Analysis Services
Developer 32 64 GB For development and test use
Includes SQL Server CE
download/distribution rights
Personal Subset of Standard, ideal for mobile computers; not available for
purchase, ships with Enterprise Edition and Standard Edition

Chapter 2. Positioning
81
Table 2-13 DB2 UDB server editions
Express Edition
Express Edition is the newest member of the product family and is a specially
tailored full-feature relational database for small and medium businesses
designed to be embedded as a transparent part of a business solution. This
edition can be deployed on servers with up to two processors and is only
available for 32-bit platforms.
Workgroup Server Edition
Workgroup Server Edition is designed for small organizations and
departmental use that supports a small number of connected users.
Workgroup Server Edition uses a licensing model designed to provide an
attractive price point for smaller installations while still providing a full-function
database server.
This edition is ideal for deployment on servers with up to four 32-bit
processors and when the number of concurrent or registered users is small
(25–30 users), as a license is required for each concurrent or registered user
in addition to the base server license.
Workgroup Server Unlimited Edition
The Workgroup Server Unlimited Edition offers the same features provided
with Workgroup Server Edition, but with a simpler per-processor licensing
model. This edition is ideal for deployment on 32-bit servers with up to four
processors when the number of concurrent or registered users is large (more
than 25–30 users) or unknown (Web servers), since no additional license is
required for each concurrent or registered user.
Edition Max processor Description
Express 2-way (32-bit) A full-function database server designed to be
embedded as a transparent part of a business
solution with a named user licensing model.
Workgroup
Server
4-way (32-bit) A full-function database server designed for
small organizations or departmental use with a
per-user licensing model.
Workgroup
Server
Unlimited
4-way (32-bit) A full-function database server designed for
small organizations or departmental use with a
per-processor licensing model.
Enterprise
Server
Unlimited
(32-bit and 64-bit)
A full-function database server designed for
the enterprise with flexible scale-up and
scale-out capability.

82
Planning and Installing the IBM Eserver X3 Architecture Servers
DB2 UDB Enterprise Server Edition
DB2 UDB Enterprise Server Edition is designed for large enterprises that
require a database management system that can scale to support
multi-terabyte databases. It is available in both 32-bit and 64-bit versions.
There is a migration path that allows you to migrate databases from 32-bit to
64-bit versions without having to rebuild the database from scratch.
The Enterprise Server Edition is ideal for deployment on servers of all sizes,
as it provides a number of features that are not available in the Workgroup
Server Edition of the product.
In general, this edition is required when you need support for more than four
processors, have a large number of concurrent or registered users, or you
require the scalability provided by 64-bit platforms and the Database
Partitioning Feature (DPF).
DPF allows you to partition a database within a single system or across a
cluster of systems. This feature provides support for very large databases
(VLDBs) by partitioning the database into smaller parts that can more easily
be managed by individual partitioned database servers (engines). This
approach not only supports very large databases with complex workloads, it
also provides great scalability for database administration tasks, such as
loads, backups, and recovery.
2.7 Scale-up versus scale-out
The goal of system scalability is to increase performance at a rate that is
proportional to increases in system resources for a given workload.
There are two methods to achieving system scalability. They are:
Scale-up: Increasing the capacity of the single system image by adding (in
particular) processors, memory, and disk.
Scale-out: Adding systems that can be managed and run together.

Chapter 2. Positioning
83
Figure 2-5 Scale-up versus scale-out
Scale-up
Scaling up is achieved by adding resources, such as memory, processors, and
storage, to an existing system that runs on a single server. It is also referred to as
vertical scaling. The benefit to scaling up is that it is relatively easy, as in general
it only requires hardware or software that is designed to take advantage of
additional memory, processor, and storage resources.
For example, your database server may start out on a 2-way SMP system with
four GB of memory and six hard drives. As the database grows in size or the
number of users increase, you can easily scale-up by adding additional
processors, memory, and disk resources to maintain the same level of
performance. You may eventually need to replace the server with one that is
capable of supporting more resources, but today on xSeries servers, you can
scale-up to systems that support 32 processors and 64 GB of memory on 32-bit
versions of operation systems and 512 GB of memory on operation systems with
64-bit extension EM64T.
NUMA architecture, when compared to other architectures, provides near liner
scalability and minimum overhead in resource management that limits the
scalability of a single large systems when you add processors, memory, and
storage.
1 Node
4 Node 8 Node
+16 Node
Scale Out
Scale Up
IBM eServer BladeCenter
(Oracle 9i RAC or DB2)
IBM eServer 326, x336, x346
(Oracle 9i RAC or DB2)
IBM eServer x366
(Oracle 9i RAC or DB2)
Uni
2-Way
4-Way
+8-Way
IBM eServer x460
(Oracle 9i, DB2, or SQL Server)
IBM eServer x260
(SQL Server or DB2)
IBM eServer x236
(SQL Server or DB2)
IBM eServer Cluster 1350
(Oracle 9i RAC or DB2)

84
Planning and Installing the IBM Eserver X3 Architecture Servers
The xSeries x460 server is a good example of an SMP system based on
Enterprise X-Architecture and NUMA technologies. The server starts with a base
2-way configuration and you can add incremental capacity as your needs grow to
a maximum of 32 processors. Likewise, memory can be expanded from 2 GB to
512 GB. This modular server architecture delivers investment protection without
the upfront costs of expensive switch-based alternatives.
Advantages of scale-up:
Easier to configure and administer
Good when most queries access small blocks of data
Best for applications that maintain state (OLTP)
Add CPU and memory as required (scheduled downtime especially for CPUs)
All tools and queries work as expected
Can be maintained by lesser skilled DBAs
Disadvantages:
Requires higher cost hardware
The database has finite capacity limits tied to hardware
Must balance CPU, memory, and I/O to achieve peak performance
Fail-over cluster server usually configured equal to primary
Databases such as Microsoft SQL Server are well suited to a scale-up approach
using a multi-node x460 configuration.
Scale-out
Scale-out means add discrete servers to your server “farm” to gain more
processing power. Although many options exist for implementing a farm
comprised of small low-end servers, we consider the use of the IBM^
BladeCenter or 1U rack servers such as the xSeries 336 as the most viable
alternative when discussing this requirement.
Scale-out is sometimes called horizontal scaling, as well as in general being
referred to as clustering. However, clustering can sometimes be ambiguous in
that there are distinct types of clusters, which include high availability, load
balancing, and high-performance computing.
Load balancing is the goal of scaling out. That is to say, we scale-out by adding
one or more servers to an existing system to balance the system load as we add
additional demands on the system. For example, your database server may start
out on a 2-way system with 4 GB of memory and six hard drives. As the
database grows in size or the number of users increase you scale-out by adding
another server with two processors, 4 GB of memory, and six disk drives to
maintain the same level of performance. Although you do not necessarily have to

Chapter 2. Positioning
85
add an additional server with the exact specifications, it does reduce the
complexity of scaling out.
The benefit to scaling out is that you can achieve near linear scalability. That is,
as you add each additional server to the system, you effectively increase your
system capacity proportionally. Thus, scaling out provides much better returns in
terms of the additional costs associated with adding additional servers to the
system. Another benefit inherent with scaling out is that a cluster of smaller
servers generally costs less than a single large system.
The drawback to scaling out is that it requires system and database
administrators that understand the technology well enough so that it can be
implemented effectively. Another drawback is that clustering requires software
that is specifically designed for the task.
Advantages of scale-out:
Uses lower cost hardware.
Scaling is near linear.
The database size is not gated by hardware.
Preferred when queries access large blocks of data.
Best for serving stateless applications (Web).
Disadvantages:
Requires more skilled DBA to maintain cluster.
Management and scheduling more complex.
Depends upon intelligent data partitioning.
Introduces query overhead.
Maintenance activities require downtime.
Cluster aware applications can be much more expensive then stand-alone
version.
Architectures for scaling out
There are two distinct approaches to scaling out database management
systems. These are generally referred to as a
shared architecture
and a
shared-nothing architecture
. Both architectures attempt to achieve the same
goal. That is, to implement a database management system that consists of a
cluster of servers, provides linear scalability, and appears as single database to
the end users.

86
Planning and Installing the IBM Eserver X3 Architecture Servers
A
shared
architecture attempts to accomplish this goal while sharing the
database. As additional servers are added to the system, they all share or
attempt to share the same database, which resides on shared storage, hence the
name shared architecture.
Oracle is an example of a database application that implements a shared-disk
approach.
Figure 2-6 shows an example of a scale-out shared architecture.
Figure 2-6 Scale-out shared architecture
On the other hand, a
shared-nothing
architecture accomplishes the same goal by
dividing a large database into smaller and more manageable “parts,” called
partitions. The term shared-nothing simply refers to the fact that as additional
servers are added to the system, each server manages a clearly defined portion
of the database.
Figure 2-7 shows an example of a scale-out shared nothing architecture.
Figure 2-7 Scale-out shared nothing architecture
The fact that the database is partitioned should not imply that the system cannot
be implemented on shared storage.
IBM DB2 and Microsoft SQL Server both implement a shared-nothing approach.
FAStT700
SAN
Switch
Scale-out,
shared
Scale-out,
shared nothing

Chapter 2. Positioning
87
Choosing scale-up or scale-out
Microsoft SQL Server is well-suited for scale-up configurations, such as a
multi-node x460 configuration. It follows a single server, shared nothing
approach and it is a high performance solution for Windows environments
Oracle uses a shared-disk approach and is suited to scale-up or scale-out. It is a
leading solution for middle market UNIX, Windows, and Linux environments.
Scale-out capabilities can be extended with Oracle 9i RAC.
DB2 is suited to scale-up or scale-out. It is developed following a multi-server,
shared nothing approach, and is the highest performing database environment
for mainframe, UNIX, and Linux environments. Scale-up is preferred for smaller
databases (150-200 GB). For larger databases, large block I/O, data
warehousing and decision support applications, use a scale-out deployment.

88
Planning and Installing the IBM Eserver X3 Architecture Servers

© Copyright IBM Corp. 2006. All rights reserved.
89
Chapter 3.
Hardware planning
In this chapter, we discuss the topics you need to understand before you finalize
the configuration of your x460, x366, or x260 solution. The topics covered are:
3.1, “Processor subsystem” on page 90
3.2, “Memory subsystem” on page 95
3.3, “Multi-node configurations” on page 104
3.4, “RSA II SL and BMC initial setup” on page 119
3.5, “Storage options” on page 122
3.6, “Power considerations” on page 126
3.7, “Performance tuning and optimization” on page 129
3.8, “ServerProven” on page 130
3.9, “Solution Assurance” on page 131
3

90
Planning and Installing the IBM Eserver X3 Architecture Servers
3.1 Processor subsystem
All three X3 Architecture servers use Intel Xeon MP processors, featuring the
64-bit EM64T technology and driven by a 667 MHz dual front side bus.
The x260 and the x366 are available with up to four Intel “Cranford” processors.
The x460 support up to four Intel “Potomac” processors, but can scale-up further
by connecting up to seven MXE-460 expansion enclosures to form a single
system image of up to 32 processors. These multi-node complexes are
discussed further in 3.3, “Multi-node configurations” on page 104.
Table 3-1 lists the processors supported by each server.
Table 3-1 Xeon MP processors used in the X3 Architecture servers
Table 3-2 lists the processors standard in each server.
Table 3-2 Processors standard in each model
Speed L2 / L3 cache x260 x366 x460 MXE-460 Part number
Intel “Cranford” processors
3.16 GHz 1 / 0 MB Supported Supported No No 13N0694
3.66 GHz 1 / 0 MB Supported Supported No No 13N0695
Intel “Potomac” processors
2.83 GHz 1 / 4 MB No No Supported Supported 13N0715
3.00 GHz 1 / 8 MB No No Supported Supported 13N0714
3.33 GHz 1 / 8 MB No No Supported Supported 13N0713
Intel “Paxville” dual-core processors
2.67 GHz 1+1 / 0 MB No Supported Supported Supported 25R8941
3.00 GHz 2+2 / 0 MB No Supported Supported Supported 25R8942
Server Server Core speed Cores/socket Installed Maximum Part number
x260 8863-1RY 3.16 GHz Single-core 1 4 13N0694
x260 8863-2RY 3.66 GHz Single-core 1 4 13N0695
x366
8863-1RY
3.16 GHz
Single-core
1
4
13N0694
x366
8863-2RY
3.66 GHz
Single-core
1
4
13N0695
x366
8863-3RU
2.67 GHz
Dual-core
1
4
25R8941

Chapter 3. Hardware planning
91
The following guidelines apply to each of these servers:
All processors must be the same type, speed, and L2/L3 cache size.
The Voltage Regulator Modules (VRM) for processors 1 and 2 are integrated
on the microprocessor board. The VRMs for processor 3 and 4 are shipped
with the microprocessor option and must be installed in the appropriate
sockets. If you use this option to install a CPU in sockets 1 or 2 (in the
MXE-460 for example), the VRM in the package is not used.
The installation order for the microprocessors and the position of the VRM
sockets are depicted in Figure 1-10 on page 19.
To ensure proper airflow and cooling, each processor socket must be
populated with either an processor air baffle, or with a heat sink when a
processor is installed.
3.1.1 Dual-core upgrades
The Dual Core X3 Upgrade Kit, part 39Y6580, is available to upgrade any
single-core x460 and MXE-460 servers to be dual-core capable. This kit contains
a new processor planar.
x366
8863-4RU
3.00 GHz
Dual-core
1
4
25R8942
x460 8872-1RY 2.83 GHz Single-core 2 4 13N0715
x460 8872-2RY 3.00 GHz Single-core 2 4 13N0714
x460 8872-3RY 3.33 GHz Single-core 2 4 13N0713
x460 8872-5RU 2.67 GHz Dual-core 2 4 25R8941
x460 8872-6RU 3.00 GHz Dual-core 2 4 25R8942
MXE-460
8874-1RY
Match the x460
Single-core
0
4
Match the x460
MXE-460
8874-2RU
Match the x460
Dual-core
0
4
Match the x460
Server Server Core speed Cores/socket Installed Maximum Part number
Note: The Y and U in the model numbers is for countries in North and South
America. For EMEA, substitute G (for example, 1RG). For Asia-Pacific
countries, the letter varies from country to country. Consult the announcement
letter or the xSeries Configuration and Option Guide, found at:
http://www.pc.ibm.com/support?page=SCOD-3ZVQ5W

92
Planning and Installing the IBM Eserver X3 Architecture Servers
Processors are not part of the kit and must be purchased separately. All
previously installed processors will need to be replaced with dual-core
processors, as a mix of processors is not supported.
3.1.2 Processor configuration options
BIOS includes a number of settings that clients can set related to processor
configuration and performance. To access these settings, do the following:
1.Press F1 during system startup when prompted.
2.From the Main Menu, choose Advanced Setup.
3.Select CPU Options. Figure 3-1 appears.
Figure 3-1 CPU options in BIOS
Hyper-Threading Technology
As described in 1.8, “Processors” on page 18, the processors used in these
servers include a technology called Hyper-Threading, which makes a single CPU
appear to the operating system as two logical processors, so that they can
receive and process two data/instruction streams simultaneously.
Hyper-Threading is enabled by default. Fewer or slower processors usually yield
the best gains from Hyper-Threading, because there is a greater likelihood that
the software can spawn sufficient numbers of threads to keep both paths busy.
Note: If a client has implemented a four-node or eight-node configuration
using servers with single-core processors, and plans to upgrade their complex
to dual-core processors, then they will need to comply with export regulations.
In addition, they will need to replace one memory card in each node with part
number 40K2450.
This is described in more detail in 3.3.4, “Export regulations for 32-core com-
plexes” on page 118.
Note: In x460 multi-node configurations, all nodes in a partition should have
the same Hyper-Threading setting. See 3.3.2, “Scalable system setup” on
page 110 for more information.
Hyperthreading Technology
Clustering Technology
Processor Adjacent Sector Prefetch
Processor Hardware Prefetcher
Processor Execute Disable Bit
[ Enabled ]
[ Logical Mode ]
[ Enabled ]
[ Enabled ]
[ Disabled ]
CPU Options

Chapter 3. Hardware planning
93
Clustering Technology
For certain operating systems, it is necessary to configure how the routing of
processor interrupts in a multi-processor system is handled. It is a low level
setting that sets a multi-processor interrupt communication protocol (XAPIC).
The settings are functional only, and do not affect performance.
In the Clustering Technology menu, choose the appropriate mode for your
operating system, as advised in the operating system installation instructions in
Chapter 4, “Operating system installation” on page 135.
Clustering here refers to the ability of the x260, x366, and x460 to have CPUs
across multiple processor buses: the processors are “clustered” into pairs of
processors, one pair per bus. Each server has two processor buses, and each
additional node in an x460 multi-node complex has two extra processor buses.
The choices are Logical Mode, Physical Mode, and Special Mode.
The Logical Mode is the default mode for the system. It can be used by
Windows and Linux.
Physical Mode is the required setting for Linux to boot systems when there
are 16 or more processor cores in the complex (that is 16-way or larger with
the current processors, or 8-way or more when dual-core processors become
available).
The Special Mode is required for the 64-bit edition of Red Hat Enterprise
Linux 3.0 and only when single-core processors are installed. This is the only
operating system that needs it and Windows Server 2003 with Service Pack 1
will not even boot in this mode.
When dual-core
Paxville
processors are used, the necessary BIOS update
will remove Special Mode from the choices and, if it was previously set, will
force it to Logical Mode.
Processor Adjacent Sector Prefetch
When this setting is enabled, which is the default, the processor retrieves both
sectors of a cache line when it requires data that is not currently in its cache.
When it is disabled, the processor will only fetch the sector of the cache line that
contains the data requested.
Note: The term
clustering
here does not refer to the cluster technology
provided by such services as Microsoft Cluster Service.

94
Planning and Installing the IBM Eserver X3 Architecture Servers
This setting may affect performance, depending on the application running on the
server and memory bandwidth utilization. Typically, it will affect certain
benchmarks by a few percent, although in most real applications it will be
negligible. This control is provided for benchmark users that wish to fine-tune
configurations and settings.
Processor Hardware Prefetcher
By default, hardware prefetching is enabled, which enables the processors to
prefetch extra cache lines for every memory request. Recent tests in the
performance lab have shown that you will get the best performance for most
commercial application types if you disable this feature. The performance gain
can be as much as 20% depending on the application.
To disable prefetch, go to BIOS Setup (press F1 when prompted at boot) and
select Advanced Settings → CPU and set HW Prefetch to Disabled. For
high-performance computing (HPC) applications, we recommend you leave HW
Prefetch enabled. Future releases of BIOS that ship to enable dual-core will have
HW Prefetch disabled by default.
Processor Execute Disable Bit
Processor Execute Disable Bit is a function of new Intel processors with lets you
prevent the execution of data that is in memory as though it were code. When
enabled (the default is disable), this will prevent viruses or the like from gaining
unauthorized access to applications by exploiting buffer overruns in those
applications.
If this feature is enabled, and provided the operating system has marked the
memory segment as containing data, then the processor will not execute any
code in the segment.
This parameter in BIOS is disabled in BIOS by default just in case the
applications you run on your server are affected by it (at least one Java
application is known to fail if this setting is enabled). For added protection, you
may wish to enable it, but you should first test your applications to ensure they
will still run as expected before enabling this in a production environment.
Note: This function is only used for 32-bit operating environments where the
processor is in one of the following modes:
Legacy protected mode, if Physical Address Extension (PAE) is enabled
on a 32-bit operating system.
IA-32e mode, when EM64T is enabled on a 64-bit operating system.
The operating system must also implement this function.

Chapter 3. Hardware planning
95
For more details of the Execute Disable Bit function, see:
http://www.intel.com/cd/ids/developer/asmo-na/eng/149308.htm
3.2 Memory subsystem
This section discusses the memory subsystem: available options, installation
rules, and configuration recommendations.
The X3 Architecture servers implement memory using 1-4 memory cards, each
of which holds four DIMMs, as shown in Figure 3-2. The servers have one or two
memory cards installed as standard (model dependant); additional ones are part
number 13M7409.
Figure 3-2 Memory card locations (x460 and x366 chassis shown)
AC
DC
Each memory card has
four DIMM sockets.
DIMMs must be installed
in matched pairs.
The servers supports
up to four cards.
Models have one or
two cards standard.
Memory cards can be
hot-swapped or
hot-added (specific
restrictions apply).

96
Planning and Installing the IBM Eserver X3 Architecture Servers
The memory cards connect directly into the system planar and are powered by
two separate memory power buses (two memory cards on each bus). As shown
in Figure 3-3, memory cards 1 and 2 are on power bus 1, and memory cards 3
and 4 are on power bus 2. The power arrangement is of particular importance
with the hot-swap and hot-add functions, as discussed in 3.2.2, “Hot-swap
memory” on page 99 and 3.2.3, “Hot-add memory” on page 100 respectively.
The system memory uses ECC DDR2 DIMMs meeting the PC2-3200 standard.
The DIMM layout of the memory cards is also shown in Figure 3-3.
Figure 3-3 Memory implementation
The standard memory cards and DIMMs are shown in Table 3-3.
Table 3-3 Standard memory cards and DIMMs
Supported DIMM options are:
1 GB upgrade kit, part number 73P2865 (contains two 512 MB DIMMs)
2 GB upgrade kit, part number 73P2866 (contains two 1 GB DIMMs)
4 GB upgrade kit, part number 73P2867 (contains two 2 GB DIMMs)
8 GB upgrade kit, part number 30R5145 (contains two 4 GB DIMMs)
Four memory cards
(1 or 2 standard)
Each card has
four DIMM sockets
If memory is installed,
card 1 has sockets
1 and 3 filled with
1 GB DIMMs
1
2
3
4
DIMM
Socket 1
DIMM
Socket 4
Hot-swap
enabled LED
Memory port
power LED
Power bus 2
(cards 3 & 4)
Power bus 1
(cards 1 & 2)
Server Memory cards Standard DIMMs
x260 One (card 1) Two 1 GB DIMMs or two 512 MB DIMMs (model
dependant), installed in memory card 1.
x366 One (card 1) Two 1 GB DIMMs, installed in memory card 1.
x460 One (cards 1 and 2) Two 1 GB DIMMs, installed in memory card 1.
Memory card 2 contains no DIMMs.
MXE-460 Two (cards 1 and 2) None (all nodes should match).

Chapter 3. Hardware planning
97
With 4 GB DIMMs, a total amount of 64 GB RAM can be installed. In an x460
eight-node system, the maximum installable memory size is 512 GB (8 x 64 GB).
The configuration rules are as follows:
A minimum of one memory card containing two DIMMs is required for the
server to operate. This also applies to the MXE-460.
Memory is two-way interleaved to ensure maximum data throughput. As a
result, the DIMMs are always installed in pairs of the same size and type to
populate a memory bank. Banks are DIMM sockets 1 and 3 and 2 and 4, as
shown in Figure 3-3 on page 96.
The installation sequence for memory cards is 1-2-3-4. There are two ways to
fill the DIMMs sockets, depending on whether cost or performance is the
more important consideration:
– Cost-effective configuration: To minimize cost, you can install the memory
DIMMs by filling each memory card before adding DIMMs to the next
memory card.
– Performance-optimized configuration: As described in 1.4, “IBM XA-64e
third generation chipset” on page 8, there are four independent memory
ports. Therefore, to optimize performance, you can spread the DIMMs (still
installed in matched pairs) across all four memory cards before filling each
card with two more DIMMs.
For more performance tuning options, see also 3.7, “Performance tuning and
optimization” on page 129.
A more detailed description and the exact sequence for installation can be
found in the User’s Guide of each server.
If you are planning to implement a four-node or eight-node complex using
models with dual-core processors, you will need to comply with export
regulations and install one export-controlled memory card, part number
40K2450, in each node. This is described in more detail in 3.3.4, “Export
regulations for 32-core complexes” on page 118.
Note: At the time of writing, the 1 GB kit (two 512 MB DIMMs) was only
supported on the x260. Check ServerProven® for the latest information.
x460 Multi-node configurations: As discussed in 3.3, “Multi-node
configurations” on page 104, with x460 multi-node configurations, 256 MB of
memory in each node is allocated to the XceL4v cache. As a result, the
memory seen by the operating system is reduced by 256 MB for each node (2
GB for an eight-node complex, for example).

98
Planning and Installing the IBM Eserver X3 Architecture Servers
3.2.1 Memory mirroring
Memory mirroring is available on all X3 Architecture servers for increased fault
tolerance. Memory mirroring is operating system independent, since all mirroring
activities are handled by the hardware.
To understand what memory mirroring and hot-swap capabilities exist with the
server, you must first understand how the memory cards are powered. The
system planar has two separate memory power buses that are split between the
four memory cards. As shown in Figure 3-3 on page 96, memory cards 1 and 2
are on power bus 1, and memory cards 3 and 4 are on power bus 2.
Mirroring takes place across the two power buses. In other words, the memory
DIMMs in cards 1 and 2 are mirrored to the memory DIMMs in cards 3 and 4.
With memory mirroring enabled in BIOS, you can hot-swap one memory card at
a time on each memory power bus. Once memory mirroring is enabled, the data
that is written to memory will be stored in two locations. For read operations, data
is read from the DIMMs with the least amount of reported memory errors through
memory scrubbing.
Memory scrubbing is an automatic and regular test of all the system memory that
detects and reports memory errors before they cause a server outage. If memory
scrubbing determines that the DIMM is damaged beyond use, read and write
operations are redirected to the remaining good DIMM. Memory scrubbing then
reports the damaged DIMM and the light path diagnostic displays the error. If
memory mirroring is enabled, then the mirrored copy of the data from the
damaged DIMM is used until the system is powered down and the DIMM
replaced. After the damaged DIMM is replaced, memory mirroring will copy the
mirrored data back onto the new DIMM.
Key configuration rules relating to memory mirroring are as follows:
Memory mirroring must be enabled in the BIOS (it is disabled by default).
Both memory cards must have the same total amount of memory, and must
have identical DIMMs. In other words, DIMMs must be installed in matched
quads to support memory mirroring. Partial mirroring is not supported. See
the server’s Installation Guide for information about the exact installation
order required.
Important: Because of memory mirroring, you will only have half of the total
amount of memory available. If 8 GB is installed, for example, then the
operating system sees 4 GB once memory mirroring is enabled (it is disabled
in the BIOS by default).

Chapter 3. Hardware planning
99
3.2.2 Hot-swap memory
These servers supports hot-swap memory, which means that if a DIMM fails, it
can be replaced with a new DIMM without powering down the server. This
advanced feature allows for maximum system availability. Hot-swap memory
requires that memory mirroring be enabled.
To easily identify whether hot-swap is enabled and the status of power to the
memory card, each memory card has a green
memory hot-swap enable
LED,
and a green
memory port power
LED on the top panel of the memory card, as
shown in Figure 3-3 on page 96. The memory card has eject levers with sensors,
so that the system can recognize when a memory card is being removed and
power down that card’s slot accordingly.
The overall process to hot-swap a failed DIMM is as follows:
1.Verify that memory mirroring and hot-swap are enabled by checking the
memory hot-swap enabled LED on the memory cards.
2.When a DIMM fails, you will be alerted via the memory LED on the light path
diagnostics panel (and by other means via the service processor if this has
been configured).
3.Locate the memory card with the failed DIMM by using the Error LED on the
memory card.
4.Remove the memory card containing the failed DIMM.
5.Press the button on the memory card to identify which DIMM has failed. The
LED next to the failed DIMMs lights up.
6.Replace the failed DIMM and reinsert the memory card.
For a more detailed description of how to correctly hot-swap memory and which
sequence to follow, see the server’s User’s Guide.

100
Planning and Installing the IBM Eserver X3 Architecture Servers
3.2.3 Hot-add memory
The hot-add memory feature enables you to add DIMMs without turning off the
server. This section shows the requirements for enabling the hot-add memory
feature on the server.
The requirements are as follows:
Operating system support: Adding usable system memory to a running
operating system requires operating system support. This is done via an
ACPI sequence. Currently, the only operating systems that have this
capability and are supported on the X3 Architecture servers are the
Enterprise and Datacenter editions of Windows Server 2003.
Memory hot-add must be specifically enabled in the BIOS setup. When this is
done, the system allocates blank windows of memory space for future
memory additions. By enabling hot-add, memory mirroring will automatically
be disabled.
Memory cards 3 and 4 must not be installed yet because these are the
only
ones that can be hot-added.
If only one memory card is installed (memory card 1) prior to the hot-add
operation, then
only
one additional memory card may be added in slot 3.
If two memory cards (1 and 2) are already in the system, then two additional
memory cards
must
be added in slots 3 and 4. The card in slot 4 must be
inserted first. This is because power to the second bus will be applied once
card 3 is inserted and closed.
The DIMMs must be added two at a time (matched pairs) and they must also
match the equivalent pair of DIMMs on the matching memory card on the
other power bus.
A minimum of 4 GB of memory must be installed in the server in order for
hot-add memory to be available. Additionally, for 32-bit operating systems,
the Physical Address Extension (PAE) mode has to be enabled to take
advantage of the additional memory.
For information About how to perform a hot-add operation, and more information
about the restrictions, see the server’s User’s Guide.
Note: Hot-add and hot-swap are mutually exclusive. You can only enable one
of these features.

Chapter 3. Hardware planning
101
3.2.4 Memory ProteXion: redundant bit steering
Redundant bit steering is the technical term for Memory ProteXion.
When a single bit in a memory DIMM fails, the function known as redundant bit
steering (RBS) automatically moves the affected bit to an unused bit in the
memory array, removing the need to perform the ECC correction and thereby
returning the memory subsystem to peak performance. The number of RBS
actions that can be performed depends on the type of DIMMs installed in the
server:
A pair of single-sided DIMMs can perform one RBS action. These are the 512
MB and 1 GB DIMMs. A pair of single-sided DIMMs is also known as a single
Chip Select Group (CSG).
A pair of double-sided DIMMs (also known as stacked DIMMs) can perform
two RBS actions. These are the 2 GB and 4 GB DIMMs. A pair of
double-sided DIMMs is also known as two Chip Select Groups.
RBS is supported in both non-mirrored and mirrored configurations.
In the X3 Architecture servers, DIMMs are installed in matched pairs in banks.
Each memory card installed in the server is comprised of two banks:
DIMM sockets 1 and 3 form bank 1.
DIMM sockets 2 and 4 form bank 2.
Memory errors are handled as follows:
If a single-bit error occurs in a CSG, RBS is used to correct the error.
If a second single-bit error occurs in the same CSG, the ECC circuitry is used
to correct the error.
So, for example, if an x366 is configured with 16x 1 GB DIMMs (which are
single-sided), then each pair corresponds to a single CSG, so the server has a
total of eight CSGs. This means that the server can survive up to 16 single-bit
memory failures, two in each pair of DIMMs (RBS recovery, and then ECC).
Notes:
Once you have added a memory card with two DIMMs, you cannot add
more memory to that same memory card without powering off the server.
Enabling hot-add reserves a portion of the memory map for the memory
that may be hot-added in the future. If you do not plan to use hot-add, we
recommend that you not enable this feature in BIOS.

102
Planning and Installing the IBM Eserver X3 Architecture Servers
As a second and more complex example, if the same server is installed with
eight 1 GB DIMMs (single-sided) and eight 2 GB DIMMs (double-sided), then this
means there are a total of 4 + 8 = 12 CSGs (four from the four pairs of 1 GB
DIMMs and eight from the eight pairs of 2 GB DIMMs). This means the server
can survive up to 24 single-bit memory errors.
The first single bit correctable error on a CSG will result in an RBS event and
RBS log entry from the SMI Handler. The second will result in automatic
hardware correction and a threshold log entry from the SMI Handler.
3.2.5 Memory configuration in BIOS
Depending on your needs, the system memory can be configured in four different
ways:
Redundant Bit Steering (RBS) (default)
Full Array Memory Mirroring (FAMM)
Hot Add Memory (HAM)
High Performance Memory Array (HPMA)
You configure the memory subsystem in the server’s BIOS Setup menu by
selecting Advanced Settings → Memory → Memory Array Setting. The
window in Figure 3-4 should appear.
Figure 3-4 Memory options in BIOS
The choices are as shown in Table 3-4.
Table 3-4 Memory configuration modes in BIOS
Memory Card 1
Memory Card 2
Memory Card 3
Memory Card 4
Memory Array Setting
[ RBS (Redundant Bit Steering) ]
Memory Settings
Mode Memory
ProteXion
Memory-
mirroring
Hot-swap
memory
Hot-add
memory
HPMA (high performance memory array) Disabled Disabled Disabled Disabled
RBS (redundant bit steering) (default)
Yes Disabled Disabled Disabled
FAMM (full array memory mirroring)
Yes
Yes
Yes Disabled
HAM (hot-add memory)
Yes Disabled Disabled
Yes

Chapter 3. Hardware planning
103
The memory configuration mode you select depends on what memory features
you want to use:
Redundant Bit Steering (RBS):
This option enables Memory ProteXion and is the default/standard setting.
Select RBS if you are not using mirroring, hot-swap, or hot-add. See 3.2.4,
“Memory ProteXion: redundant bit steering” on page 101 for details of how
RBS works.
Full Array Memory Mirroring (FAMM)
Select FAMM to enable memory mirroring (and to enable hot-swap).
Memory mirroring reduces the amount of addressable memory by half on
each chassis in the partition, but provides complete redundancy of all
addressable memory. RBS is available in this mode. See 3.2.1, “Memory
mirroring” on page 98 for more information.
Hot-Add Memory (HAM)
Select HAM to enable the use of the hot-add in the future.
HAM provides an array layout that supports runtime hot memory add within
an OS that supports that feature. This setting has lower performance and may
also restrict the amount of memory that can be installed in each chassis, as
addressable ranges must be reserved on each chassis for the hot add
function. RBS is available in this mode. See 3.2.3, “Hot-add memory” on
page 100 for more information.
High Performance Memory Array (HPMA)
HPMA optimizes the installed memory array on each chassis in the partition
for maximum memory performance. Hardware correction (ECC) of a single
correctable error per chip select group (CSG) is provided, but RBS is not
available. See 3.2.4, “Memory ProteXion: redundant bit steering” on page 101
for a discussion of chip select groups.
We recommend you do not select the HPMA setting in a production environment,
as this disables Memory ProteXion.

104
Planning and Installing the IBM Eserver X3 Architecture Servers
3.3 Multi-node configurations
The IBM
XpandOnDemand
scalability features give the flexibility to expand the
x460 server’s capacity in terms of number of CPUs, memory, and I/O slots, as
the demand grows.
Clients can expand the server in the following ways:
CPUs from two-way up to 32-way
Memory from 2 GB to 512 GB
PCI-X slots from six to 48
This scalability is achieved by connecting multiple x460s or MXE-460s to the
base x460. These
nodes
then form a single complex. The supported expansion
steps are listed in Table 3-5.
An x460 server can be configured together with one, three, or seven MXE-460s
to form a single 8-way, 16-way, or 32-way complex.
Table 3-5 x460 scalability options
You can also form multi-node complexes using multiple x460s or combinations of
x460s and MXE-460s. With these combinations, you can partition the complex,
as described in 3.3.3, “Partitioning” on page 116.
Note: This section only applies to the x460 and MXE-460. The x260 and x366
do not support multi-node configurations.
Nodes CPUs Maximum RAM PCI slots Number of MXE-460s*
1 2-way 64 GB Six None
1 4-way 64 GB Six None
2 8-way 128 GB 12 One
4 16-way 256 GB 24 Three
8 32-way 512 GB 48 Seven
* Additional nodes can be either MXE-460 or x460 servers
Important: As we discuss in 3.3.4, “Export regulations for 32-core complexes”
on page 118, all 32-core or 64-core configurations will require US government
approval regardless of where the complex will be installed.

Chapter 3. Hardware planning
105
As shown in Figure 3-5, a scalable system comprises of an x460 server and one,
three, or seven MXE-460 systems.
A fully configured, eight-node, and scalable system would have 32 processors,
512 GB of memory (using 4 GB DIMMs), 48 PCI-X 2.0 adapters, 3.5 TB of disk
space (non-RAID), and 16 Gigabit Ethernet connections.
Figure 3-5 The four multi-node configurations supported
The following configuration rules apply:
Multi-node configurations
The multi-node configurations can have more than one x460. Only one is
shown in Figure 3-5 for simplicity. In fact, if you wish to use partitioning as
described in 3.3.3, “Partitioning” on page 116, you will need one x460 for
each partition to be the primary node.
The first x460 is known as the primary node; all other nodes in a complex are
called secondary nodes. The primary node must always be an x460
(MXE-460s cannot be primary nodes).
Only one, two, four, or eight nodes are supported. Other combinations cannot
be selected in the configuration utility.
2-way or 4-way
Up to 64 GB RAM
x460 single chassis
8-way
Up to 128 GB RAM
x460 + 1x MXE-460
Two chassis
16-way
Up to 256 GB RAM
x460 + 3x MXE-460
Four chassis
32-way
Up to 512 GB RAM
x460 + 7x MXE-460
Eight chassis
x460 x460
MXE-460
x460 x460
MXE-460
MXE-460
MXE-460
MXE-460
MXE-460
MXE-460
MXE-460
MXE-460
MXE-460
MXE-460

106
Planning and Installing the IBM Eserver X3 Architecture Servers
Processors
The x460 and all MXE-460s must each have four processors installed and all
processors must be the same speed and cache size. The x460 and MXE-460
are technically capable of having less than four CPUs installed in a multi-node
configuration, but this type of configuration will require SPORE approval
before it can be supported.
Memory
For performance reasons, you should have the same amount of memory in
each node. A minimum of 2 GB of RAM is required in each node. The x460
and MXE-460 are technically capable of having less than 2 GB of RAM
installed in a multi-node configuration, but this type of configuration will
require SPORE approval before it can be supported.
In a multi-node configuration, 256 MB of RAM per node is allocated to the
XceL4v cache. In an 8-node, 32-way complex, this means that 2 GB of RAM
is allocated to XceL4v and is unavailable to the operating system.
32-core and 64-core configurations
As described in 3.3.4, “Export regulations for 32-core complexes” on
page 118, partitions with 32 single-core processors or 16 or 32 dual-core
processors are prevented from being configured unless a special memory
card is installed.
Firmware
All system firmware, including the system BIOS, diagnostics, BMC firmware,
and RSA II SlimLine firmware, must be at the same level across all systems.
Updating the system BIOS in every node in a scalable system can be
performed from the primary node. The BIOS code in all secondary nodes will
automatically be updated as well. The server diagnostics, as well as the BMC
and RSA II SlimLine firmware, must be individually updated on each node,
but this can be performed remotely: The RSA II firmware can be updated
using the RSA II Web interface or IBM Director. The BMC firmware can be
updated with an RSA II remote console session using the remote diskette
function.
Disk drives installed in any of the MXE-460s are seen by the operating
system as normal disk drives.
If you install the optional ServeRAID 8i RAID adapter into each node, you will
be able to form RAID arrays, but these arrays cannot span nodes. All drives in
each array must be local to the RAID adapter.
All PCI-X slots and onboard Gigabit Ethernet ports in the MXE-460 are visible
to the operating system as well.

Chapter 3. Hardware planning
107
The concept of forming a single system image from multiple nodes is called
merging
the nodes. Once a multi-node system is started and the nodes are
merged, all existing devices on the secondary nodes (COM ports, keyboard,
video, mouse, and DVD-ROM drives) are disabled. PCI devices are on all
nodes (including the onboard Gigabit Ethernet and USB controllers) available
to the operating system.
Power control is as follows:
– Pressing the power button of any node in a complex will power on or off
the entire complex.
– Pressing the reset button on any node in the complex will restart the entire
complex.
3.3.1 Cabling
In order to create a multi-node complex, the servers and expansion modules are
inter-connected using a high-speed data bus known as the
scalability bus
. The
connection uses copper colored
scalability cables
, which are available in two
different lengths:
2.3 m (7.5 ft) scalability cable, part number 13M7414
2.9 m (9.5 ft) scalability cable, part number 13M7416
The 2.3 m cable is usually used to connect nodes together except when the
longer 2.9 m cable is preferred, such as in the 8-node complex, where it is used
to connect nodes that are too far away for the 2.3 m cable (for example,
connecting node 1 and node 5).
To build your multi-node complex, you will need to order the appropriate number
of each type of scalability cable for your specific configuration, as in Table 3-6.
The x460 and MXE-460 do not come with scalability cables.
Table 3-6 Scalability cable configuration table
Note: These cables are not compatible with the x440 and x445 equivalent
cables.
Chassis configuration 2.3 m (short) cables
needed
2.9 m (long) cables
needed
4-way: 1-chassis None None
8-way: 2 chassis Two None
16-way: 4 chassis Six None

108
Planning and Installing the IBM Eserver X3 Architecture Servers
There are three different cabling schemes, depending on the number of nodes
used in the complex. These are shown in the following diagrams. In a 2-node
configuration, two scalability cables are used to connect both chassis. The
second cable provides redundancy for the chassis interconnect as well as a
slight performance benefit.
In any multi-node configuration, any
one
scalability cable may fail without impact
on the server’s operation. In this situation, a warning LED on the light path
diagnostic panel will be lit and an event will be logged into the RSA event log, as
shown in Figure 3-6
Figure 3-6 Messages in the RSA II event log when a scalability cable fails
Figure 3-7 depicts the scalability cabling plan for a 2-node / 8-way configuration.
Figure 3-7 Cabling for a two-node configuration
Figure 3-8 on page 109 depicts the scalability cabling plan for a 4-node / 16-way
configuration. This uses just the short (2.3 m) cables.
32-way: 8 chassis Eight Four
Chassis configuration 2.3 m (short) cables
needed
2.9 m (long) cables
needed
SMI reporting Scalability Event:Double Wide Link Down.Chassis Number = 1. Port
Number = 0.
SMI reporting Scalability Event:Link Down.Chassis Number = 1. Port Number = 1.
SMI reporting Scalability Event:Double Wide Link Down.Chassis Number = 2. Port
Number = 0.
SMI reporting Scalability Event:Link Down.Chassis Number = 2. Port Number = 1.
8-way configuration
Port 3
x460
(primary)
Port 3
MXE-460
RSAII
RSAII
Ethernet
network
Port 1
Port 2
Port 1
Port 2
2.3m cable
(13M7414)

Chapter 3. Hardware planning
109
Figure 3-8 Cabling for a four-node configuration
Figure 3-9 depicts the scalability cabling plan for an 8-node / 32-way
configuration. This one uses a combination of short (2.3 m) and long (2.9 m)
cables.
Figure 3-9 Cabling for a eight-node configuration
16-way configuration
x460
(primary)
MXE-460
(node 2)
Ethernet
network
MXE-460
(node 3)
MXE-460
(node 4)
2.3m cable
(13M7414)
Port 3
RSAII
Port 2
Port 1
Port 3
RSAII
Port 2
RSAII
Port 2
RSAII
Port 2
Port 1
Port 1
Port 3
Port 1
Port 3
32-way configuration
x460
(primary)
MXE-460
(node 2)
Ethernet
network
MXE-460
(node 3)
MXE-460
(node 4)
2.3m cable
(13M7414)
2.9m cable
(13M7416)
Port 3
RSAII
Port 2
Port 1
Port 3
RSAII
Port 2
RSAII
RSAII
Port 1
Port 1
Port 3
Port 1
Port 3
MXE-460
(node 6)
MXE-460
(node 7)
MXE-460
(node 8)
Port 3
RSAII
Port 2
Port 1
Port 3
RSAII
RSAII
Port 2
RSAII
Port 2
Port 1
Port 1
Port 3
Port 1
Port 3
MXE-460
(node 5)
Port 2
Port 2
Port 2

110
Planning and Installing the IBM Eserver X3 Architecture Servers
All Port 2 connectors use the longer 2.9 m cable, since they are further apart.
One exception to this is the Port 2 cable between nodes 3 and 6. The
recommendation is to use the longer cable so you can insert a KVM or other
device in the middle of the complex. The choice of cable is purely a matter of
physical connectivity; it does not affect signaling or performance.
3.3.2 Scalable system setup
The setup of a multi-node complex will be configured through the Web interface
of the RSA II SlimLine adapter, which is standard with every x460 and MXE-460.
The multi-node configuration data is stored on the RSA in all chassis that are
members of the scalable partition. Figure 3-10 on page 111 depicts the RSA
Web interface for the scalable partition configuration.

Chapter 3. Hardware planning
111
Figure 3-10 Create Scalable Partition Web interface
The communication between the RSA adapters is handled through the systems
management Ethernet connections. Therefore, it is very important to ensure
secure and reliable network connections. We recommend you connect all RSAs
to a separate management network, which is exclusively used for management
and not shared with the production or campus LAN.
For proper operation, it is necessary to maintain continuous connectivity between
the RSAs. As a result, we recommend you assign static IP addresses to the RSA
II cards. If you are using DHCP, you should configure address reservations to
ensure the addresses never change.

112
Planning and Installing the IBM Eserver X3 Architecture Servers
3.4, “RSA II SL and BMC initial setup” on page 119 describes how to prepare the
RSA II adapter.
To set up a scalable system, the following prerequisites must be fulfilled:
The firmware on all nodes must be at the same level. This includes system
BIOS, diagnostics, BMC firmware, and RSA firmware.
The settings made in the Configuration/Setup utility for memory and CPUs
must be identical.
During the configuration via the RSA II Web interface, all participating nodes
must be powered off, but connected to AC power.
The operating system must reside on the primary node. This means either
internal SAS hard disk drives or host bus adapters (HBAs) for SAN boot must
be installed in the primary node.
You can only create partitions out of set combinations of nodes. If you select
any other combination, you will get an error message when you try to create
the configuration. Valid node combinations are:
– Nodes 0, 1
– Nodes 2, 3
– Nodes 4, 5
– Nodes 6, 7
– Nodes 0, 1, 2, 3
– Nodes 4, 5, 6, 7
To create a scalable partition and then activate it, do the following:
1.Ensure all nodes are powered off but that they are connected to AC power.
2.Log in to the RSA of the server that will be the primary node in the complex.
Open a Web browser and enter the Ethernet address of the primary node’s
RSA II adapter.
3.In the RSA main menu, select Scalable Partitioning → Create Partition, as
shown in Figure 3-10 on page 111.
4.Choose the number of chassis in the drop-down menu, as shown in
Figure 3-11 on page 113.
Note: Scalable partitions can only be created using the primary node’s
RSA interface. It is not possible to create, delete, start, or stop a scalable
partition from a secondary node’s RSA interface.

Chapter 3. Hardware planning
113
Figure 3-11 Number of chassis
5.Enter a Scalable Partition Number (ID). This can be any numeric string from 1
up to 4294967295 (which is FFFFFFFF in hex). The scalable partition ID must
be unique for each partition in a complex.
Figure 3-12 Scalable Partition ID
6.Select an Chassis ID for the Primary Scalable Node in the drop-down menu,
as shown in Figure 3-13. The Chassis ID uniquely identifies each RSA within
a partition, although the actual value you choose does not represent any
functional value. Within a scalable partition, only the following sets of Chassis
IDs are valid: (0,1), (2,3), (4,5), (6,7), or (0,1,2,3) (4,5,6,7).
Figure 3-13 Primary node Chassis ID
Tip: Since the partition ID will be converted into a string in hex format,
which is displayed in the Scalable Partition Status windows, we
recommend you just use the values 1-8.

114
Planning and Installing the IBM Eserver X3 Architecture Servers
7.In Figure 3-14, choose all the nodes that will join the partition by selecting a
valid Chassis ID in the “In Scalable Partition?” drop-down menu. Next to each
where you select Yes, enter the IP address or host name of the RSA adapter
in that node.
If you intend to use host names instead of IP addresses for the RSA adapters,
note the following:
– Only fully qualified domain names (FQDN) may be used, for example,
mxe460rsa.itso.ral.ibm.com.
– The RSA must have a DNS server specified in ASM Control → Network
Protocols → Domain Name System (DNS).
Figure 3-14 Secondary nodes settings
8.Select a desired value in the Merge timeout minutes drop-down menu,
which is shown in Figure 3-15. The default timeout is five minutes, which is
suitable in most cases.
Figure 3-15 Merge timeout
The merge timeout determines the time the primary node in the complex will
wait for secondary nodes to merge. If for any reason one or more secondary
nodes fail to merge, the primary node will boot either stand-alone or with the
remaining nodes in a reduced configuration, after the merge timeout has
expired.
If you wish to shorten this value to reduce the possible downtime of a
multi-node complex for business critical applications, we recommend you wait

Chapter 3. Hardware planning
115
until you have finalized the configuration of the multi-node complex and
measure the time it takes until the Merge Complete message appears on all
nodes in the complex. You can then use a value larger than this time.
9.The option On merge failure, attempt partial merge?, shown in Figure 3-16,
gives you the choice of whether or not a partial merge attempt shall be made
for the case when one or more secondary nodes failed to merge in a 4-node
or an 8-node complex.
Figure 3-16 Partial merge
10.Click the Save button on the bottom right to save the new, scalable partition
configuration.
If any value should be invalid, the communication between the RSAs fails or
any other problem occurs, saving the partition configuration will fail and a
pop-up error message is shown. This ensures that no invalid configuration
can be saved or even become active.
After saving the partition configuration, the status page will be displayed (also
available by selecting Scalable Partitioning → Status) (see Figure 3-17).
Figure 3-17 Status after creating a new scalable partition

116
Planning and Installing the IBM Eserver X3 Architecture Servers
11.Now that the partition has been created, it must be made active by “moving it”
to be a current scalable partition. To do so, select Scalable Partitioning →
Control Partition(s) → Move New Scalable Partition to Current Scalable
Partition.
After moving the configuration information, the status will be displayed, as
shown in Figure 3-18.
Figure 3-18 Status after moving new scalable partition to current scalable partition
12.As soon as the partition configuration is finished successfully and looks as
shown in Figure 3-18, it can be started by selecting Scalable Partitioning →
Control Partition(s).
3.3.3 Partitioning
As discussed in 1.3, “Multi-node capabilities and partitioning” on page 4, the
complex can be configured as one scalable partition with two, four, or eight
nodes. Alternatively, it is also possible to split this complex into multiple
independent partitions. For example, an eight-node configuration can be split into
two 4-node systems by changing the configuration without changes to the
cabling.
Note: Moving the configuration data from New Scalable Partition to
Current Scalable Partition overwrites any existing partition configuration!

Chapter 3. Hardware planning
117
The decision whether partitioning is required or not must be made during the
planning stage of a multi-node system, since the primary node in a multi-node
complex always must be an x460. Configuring multiple partitions on a complex
that consists of one x460 with one, three, or seven MXE 460s attached is not
supported. You must have one x460 as the primary node for every partition you
create.
The actual steps to configure a server partition are exactly the same as
described in 3.3.2, “Scalable system setup” on page 110.
Figure 3-19 Four-node complex split into two partitions
As an example, to configure two 2-node partitions in a four-node complex, do the
following:
1.Log in to the RSA Web GUI of chassis 1.
2.Create a new partition with chassis 2 as the secondary node, as described in
3.3.2, “Scalable system setup” on page 110
3.Select Scalable Partitioning → Control Partition(s) → Move the new
Scalable partition configuration to Current Partition configuration.
Partition #
2
2
1
1
Chassis ID #
4 (MXE-460)
3 (x460)
2 (MXE-460)
1 (x460)

118
Planning and Installing the IBM Eserver X3 Architecture Servers
4.Select Scalable Partitioning → Control Partition(s) → Move New
Scalable Partition to Current Scalable Partition to make the new
configuration active.
5.Log in to the RSA web GUI of chassis 3.
6.Create a new partition with chassis four as the secondary node, as described
in 3.3.2, “Scalable system setup” on page 110.
7.Select Scalable Partitioning → Control Partition(s) → Move the new
Scalable partition configuration to Current Partition configuration.
8.Select Scalable Partitioning → Control Partition(s) → Move New
Scalable Partition to Current Scalable Partition to make the new
configuration active.
The two partitions now behave like two independent 8-way servers. Each
partition can be controlled independently via the RSA in each primary node. This
means they can be powered on and off without any impact on each other.
Though they are still wired as a 4-node complex, there will be no data transmitted
between the partitions.
3.3.4 Export regulations for 32-core complexes
Due to the power of the xSeries 460 scalable server, x460/MXE-460
configurations with 32 or more processor cores exceed the US Government
mandated composite theoretical performance (CTP) threshold of 190,000
millions composite theoretical operations per second (MTOPS) and require a
U.S. export license before you can ship a 32-core or 64-core system into certain
countries.
The scalable systems setup menu prevents such configurations from being
created unless a special export-controlled memory card is installed in every node
Note: As soon as the New Scalable Partition configuration is moved to the
Current Scalable Partition configuration, all previously active settings are
lost.
Note: Configurations with 32 or more processor cores are the following:
An eight-node (32-way) x460 complex with servers that use single-core
processors
A four-node (16-way) or eight-node (32-way) x460 complex with servers
that use dual-core processors

Chapter 3. Hardware planning
119
in the complex. This memory card is called the xSeries 460 Scalability
Enablement/Memory Adapter, part number 40K2450.
At least one 40K2450 memory card must be installed in every node. The check
for this card is performed at boot time by the RSA II SL service processor in each
node (the RSA II SL card needs firmware 26A or later). If a 40K2450 memory
card is not present in every node, then the complex will not merge and the
systems will boot in stand-alone mode instead.
Note: The enablement CD-ROM, part 31R1558, that was previously used to
unlock 32-way configurations, is no longer being used to fulfill this control.
If a 32-core or greater complex is desired, the client and the IBM salesperson or
IBM Business Partner will need to work closely with the IBM Export Regulation
Office (ERO) and the in-country Export Regulations Coordinator to initiate the
licensing process and acquire the documentation and approvals necessary for
IBM to be able to ship the system to the intended client. IBM employees can
access the ERO at this internal IBM Web site:
http://w3.ibm.com/chq/ero
Be aware that this process can be quite lengthy, lasting up to six months.
Therefore, the process should be initiated very early in the procurement cycle.
Once the required approvals have been obtained, the client will be eligible to
receive the memory card that will permit 32-core configurations. You can order
the memory card before approvals have been received, but it will ship only when
approvals are in place.
The export-controlled memory card simply replaces the regular memory card and
otherwise provides identical function.
3.4 RSA II SL and BMC initial setup
The Baseboard Management Controller (BMC) is standard on all three X3
Architecture servers. The RSA II SlimLine adapter is standard on the x460 and
MXE-460, and optional on the x260 and x366.
Both the BMC and the RSA II SlimLine can be initially configured using the BIOS
Configuration/Setup utility. Use it to configure the IP addresses of both service
processors.

120
Planning and Installing the IBM Eserver X3 Architecture Servers
3.4.1 RSA II SlimLine setup
By default, the RSA II SlimLine is configured to obtain an IP address using DHCP
and, if that is not available, then to configure a static IP address of
192.168.70.125.
To configure the RSA II SlimLine to use another static address, you can use one
of two methods:
Access the adapter from a client computer, for example using a mobile
computer computer connected to the RSA with a crossover Ethernet cable.
Open a Web browser and point it to the adapter’s IP address
(192.168.70.125).
Configure the adapter from the server’s BIOS Setup/Configuration:
a.Press F1 during system startup when prompted to enter the
Configuration/Setup utility.
b.Select Advanced Setup → RSA II Settings.
c.Select Use Static IP under DHCP Control and enter a static IP address,
subnet mask, and gateway in the appropriate fields as required, as shown
in Figure 3-20 on page 121.
d.Select Save Values and Reboot RSA II to make the configured network
settings active.
Tip: Since the RSA II SlimLine is optional on the x260 and the x366, when you
do install it, you must flash the adapter with the latest firmware for that
particular system. See 3.4.1, “RSA II SlimLine setup” on page 120 for more
information.
Tip: The default logon credentials are USERID and PASSW0RD. All
characters are uppercase, and the 0 in PASSW0RD is a zero and not the
letter O).

Chapter 3. Hardware planning
121
Figure 3-20 RSA setup within Configuration/Setup utility
Alternatively, an IP address also can be assigned using DHCP by selecting
DHCP Enabled in the DHCP Control field. If DHCP is used, we recommend you
use an IP reservation. The IP address assigned by the DHCP server is also
displayed in this window in the DHCP IP Address field.
The RSA’s MAC address is displayed as RSA II MAC Address, which can be
helpful for network switch configuration.
In the OS USB Selection field, select the appropriate value for your operating
system. This setting determines how the RSA presents an emulated keyboard
and mouse in a remote control session to the operating system.
For Microsoft Windows and Novell NetWare, select Other OS.
For Linux, select Linux OS.
The Periodic SMI Generation setting is set to Disabled by default and should not
be changed on the X3 Architecture servers. This feature was intended for
support of older operating systems that did not include adequate checking of
CPU states. Modern operating systems poll for CPU machine checks without this
feature. No function is lost by disabling it.
RSA II Settings
RSA II MAC Address
DHCP IP Address
DHCP Control
Static IP Settings
Static IP Address
Subnet Mask
Gateway
OS USB Selection
Periodic SMI Generation
Save Values and Reboot RSA II
<<<RESTORE RSA II DEFAULTS>>>
[ Linux OS ]
[ Disabled ]
[ 009.042.171.230 ]
[ 255.255.255.000 ]
[ 009.042.171.003 ]
00-0D-60-46-D9-FC
000.000.000.000
[ Use Static IP ]

122
Planning and Installing the IBM Eserver X3 Architecture Servers
3.4.2 BMC IP address setup
The Baseboard Management Controller (BMC) is common in all three servers.
We strongly recommend configuring an IP address for the BMC, even when an
RSA is installed, so you can access and clear the BMC event log if it fills up. For
further information about how to access the BMC and perform tasks on it, please
refer to 5.3, “Baseboard Management Controller” on page 177.
The IP address configuration is done in the Configuration/Setup utility:
1.Press F1 during system startup when prompted to enter the
Configuration/Setup utility.
2.Select Advanced Setup → Baseboard Management Controller (BMC)
Settings → BMC Network Configuration.
3.Enter your static IP address, subnet mask, and a gateway.
4.Select Save Network Settings in BMC.
As with RSA the default logon credentials for the BMC are USERID and
PASSW0RD (with a zero). They can be changed by selecting Advanced
Setup → Baseboard Management Controller (BMC) Settings → User
Account Settings.
3.5 Storage options
This section discusses the available storage options for the x260, x366, and
x460.
3.5.1 Serial Attached SCSI (SAS) subsystem
The onboard disk subsystem is driven by an Adaptec AIC-9410 single chip Serial
Attached SCSI (SAS) controller. This controller connects to the internal SAS
drive bays via one or two backplanes.
Supported internal disks are described in 3.5.3, “Internal storage” on page 124.
For more information about the on-board SAS controller, see 1.11, “Serial
Attached SCSI” on page 31.
Restriction: The BMC cannot be configured via DHCP.
Note: Unlike the RSA II, you cannot administer the BMC with a Web browser.
The administrative interface is SMBridge, as described in 5.5, “OSA SMBridge
utility” on page 194.

Chapter 3. Hardware planning
123
3.5.2 ServeRAID-8i
The onboard SAS controller does not support any host RAID. To enable RAID
support, an optional ServeRAID-8i SAS RAID controller (part 13N2227) can be
installed in a dedicated slot (see Figure 3-21).
Figure 3-21 ServeRAID-8i
The ServeRAID-8i supports the following RAID levels for the internal disks:
RAID 0
RAID 1
RAID 1E
RAID 5
RAID 5EE
RAID 6
RAID 10
RAID 50
RAID 60 (the x366 and x460 do not support this RAID level because it
requires a minimum of eight drives)
The ServeRAID-8i has the following possible stripe sizes: 16 KB, 32 KB, 64 KB,
128 KB, 256 KB, and 512 KB. The default stripe size is 64 KB. The adapter
supports up to 24 logical drives. The ServeRAID-8i does not have any
connectors for external SAS devices.

124
Planning and Installing the IBM Eserver X3 Architecture Servers
The ServeRAID-8i supports two additional RAID levels compared to other
ServeRAID controllers:
RAID 6 is a multiparity type that requires a minimum of four drives, of which
two drives’ worth of capacity are consumed for redundancy. RAID-6 allows
the loss of two drives without data loss, but this loss is associated with a
substantial performance degradation.
RAID 60 is a hierarchical array made up of multiple RAID 6 arrays at the
lower level with data striped over these arrays (the 0 in 60) at the upper level.
Even though the ServeRAID-8i supports this RAID level, it is not supported on
the x460 because it requires a minimum of eight drives.
For a comparison of features of members of the ServeRAID family, see the
Technote ServeRAID Adapter Quick Reference, TIPS-0054, available from:
http://www.redbooks.ibm.com/abstracts/tips0054.html
If the system is not running applications with high microprocessor requirements,
software RAID may be a suitable option. Check with your selected OS vendor to
see what the vendor-specific requirements are. We recommend you use
hardware RAID for two reasons:
Software RAID uses the host CPUs, which may impact the performance of
the production applications.
Software RAID does not include any form of battery backup in case the server
fails in the middle of a write operation.
For hints and tips about using the ServeRAID-8i in the X3 Architecture servers,
see the publication xSeries 260, xSeries 366, xSeries 460, and xSeries MXE-460
FAQ — Hints and Tips, available at:
http://www.pc.ibm.com/support?page=MIGR-61395
3.5.3 Internal storage
Standard removable media are shown in Table 3-7 on page 125.
Note: When an optional ServeRAID-8i SAS RAID adapter is installed in the
system, use the existing internal cabling.

Chapter 3. Hardware planning
125
Table 3-7 Diskette, optical and tape drives
All three servers use Serial Attached SCSI (SAS) drives for internal disk storage
(see Table 3-8).
Table 3-8 Internal disk storage
The x260 has bays for six 3.5” SAS disk drives hot-swap drives, but supports
twelve with an optional 6-drive 3.5” SAS backplane, part number 13M7864.
Using the maximum number of hard drives, the storage maximums shown in
Table 3-9 can be achieved for the given disk drive capacities.
Table 3-9 Maximum capacities (no RAID or RAID-0)
The x260 also has two 5.25” half-height (1.6” high) drive bays that can also be
combined to form a single, full-height (3.2”) drive bay. These bays are primarily
used to install a tape drive. For a list of supported drives, see the ServerProven
page at:
http://www.ibm.com/servers/eserver/serverproven/compat/us/
Server Internal diskette drive Optical drive Tape option
x460 None
1
DVD-ROM None
x366 None
1
DVD-ROM None
x260 Standard CD-ROM Two half-high bays
Note 1: The USB diskette drive, part number 05K9276 is a supported option
Server Disks Number of bays Standard disks
1
x460 2.5” hot-swap SAS drives Six standard None
x366 2.5” hot-swap SAS drives Six standard None
x260 3.5” hot-swap SAS drives Six standard, six
optional
None
Note 1: Express models may include hard disks
Server / Disk drives 36.4 GB 73.4 GB 146.8 GB 300 GB
x260 436 GB 880 GB 1761 GB 3600 GB
x366 218 GB 440 GB 880 GB 1800 GB
x460 (per node) 218 GB 440 GB 880 GB 1800 GB

126
Planning and Installing the IBM Eserver X3 Architecture Servers
As well as the drive, clients will also need to install a SCSI adapter to attach to
the tape drive, as none is installed.
3.5.4 External storage options
To expand the storage capacity of each server, external storage devices can be
attached. Storage devices are available based on Ultra-320 SCSI, iSCSI, and
Fibre Channel. Since no external SCSI or Fibre Channel connections are
available as a standard, an optional SCSI adapter or Host Bus Adapter has to be
installed in order to connect to these external solutions.
IBM supports the following external storage devices:
The ServeRAID-6M SCSI RAID controller can attach up to two EXP400
Storage Expansion Unit. The EXP400 can be equipped with up to 14 Ultra320
146 GB disk drives, such that a maximum of approximately 4 TB disk space
can be attached.
IBM TotalStorage® DS300 iSCSI, using an optional iSCSI initiator adapter or
a software iSCSI initiator via a regular supported Ethernet adapter.
IBM TotalStorage DS400 Fibre Channel attached by the IBM TotalStorage
FC2-133 Host Bus Adapter (HBA).
Using the FC2-133 HBA, several further storage solutions, such as the IBM
TotalStorage DS family, can be attached.
Ultra320 SCSI Controller attaches external tape drives respectively, for
example, tape libraries.
ServerProven lists all supported adapters and external storage options at:
http://www.ibm.com/servers/eserver/serverproven/compat/us/
3.6 Power considerations
This section will cover several power related topics, such as available power
upgrade options, power redundancy, and recommendations for handling power
needs in different environments. All X3 Architecture servers support 220V and
110V power, but IBM recommends you use 220V power for the x366, x460 and
the MXE-460.
Note: The ServeRAID-8i is for internal SAS drives only. It does not offer
connection to external storage.

Chapter 3. Hardware planning
127
Power is delivered as follows:
The x260 and the x366 can be equipped with additional hot-swap power
supplies to provide a redundant power configuration.
The x460 and MXE-460 ship with two redundant hot-swap power supplies
standard.
Table 3-10 shows the default and the maximum configuration for each server.
Note that unlike the power supplies in the x366 and x460, the power output of the
power supply unit in the x260 is independent of the source voltage and is 775W
both at 110V and 220V.
Table 3-10 Hot-swap power supplies
3.6.1 Working in 110V environments
At 220V, all of these servers will operate with power supply redundancy when
both power supplies are installed in the x366, x460, and MXE-460. You get
redundancy in the x260 with three power supplies.
Server x260 x366 x460 MXE-460
Standard power supplies Two 775W One 1300W Two 1300W Two 1300W
Maximum supported Three 775W Two 1300W Two 1300W Two 1300W
Redundancy
requirements
Three power
supplies
Two power
supplies
Two power
supplies
Two power
supplies
Power output from each
power supply
775W at 110V
775W at 220V
650W at 110V
1300W at 220V
650W at 110V
1300W at 220V
650W at 110V
1300W at 220V
Part number 24R2656 13M7413 Not applicable Not applicable
A special note about x260 power: The x260 supports up to three power
supplies.
To provide sufficient power for all server configurations, two power
supplies are required (110V and 220V). This is the standard configuration.
If you require a redundant power supply (which we recommend), then you
should have three power supplies installed.

128
Planning and Installing the IBM Eserver X3 Architecture Servers
However, at 110V, the power supplies of the x366 and x460 may be insufficient
for heavily loaded configurations (that is, servers with many CPUs, DIMMs,
drives, and adapters):
If one power supply is installed, it may not supply enough power; in this case,
the OVERSPEC (over specification) LED will light on the operator panel
If two power supplies are installed, the second one may not provide power
redundancy; in this case, the NONRED (non-redundant) LED will light on the
operator panel.
The two standard power supplies of the x260 are sufficient to power all
configurations of the server and three power supplies is always enough to
provide power redundancy, both in 110V and 220V installations.
For example, the following x366 configuration can be powered by one power
supply and would run redundant with two power supplies:
One processor
Two PCI-X adapters
Three hard disk drives
Four 1 GB DIMMs
It is not a trivial exercise to calculate in advance the power requirements of a
specific configuration to determine how many power supplies are needed to drive
the server and how many to make the power redundant. As a result, there are
two options clients can take:
Use 220V power, either a supply direct from the utility company or via a
step-up transformer (if using the latter, ensure your power circuit has a
sufficient current rating).
Install the server with all options needed, and test this configuration under
real life conditions to see if the NONRED or OVERSPEC LEDs light up.
If neither the NONRED or the OVER SPEC LED on the light path diagnostic
are lit at any time, your server can operate properly with the power supplies
currently installed.
For simplicity of configuration, we recommend you use 220V power.
Summary: With 110V power, heavily configured x366 and the x460s may
require two power supplies for normal operation and may not be redundant.
220V power is recommended.

Chapter 3. Hardware planning
129
3.7 Performance tuning and optimization
The X3 Architecture servers are designed to deliver superior performance. This
section explains how the performance can be improved further by tuning certain
settings in the Configuration/Setup Utility or choosing the optimal way to install
hardware options.
3.7.1 Optimal memory module installation
The X3 servers support a maximum number of four memory cards. Up to four
memory DIMMs can be installed in each memory card. Each memory card is
driven by its own memory interface with a maximum bandwidth of 5.3 GBps.
Aspects of the memory configuration that can affect performance include:
Spread the installed DIMMs over as many memory cards as possible,
preferably up to the maximum of four memory cards. Because each memory
card has a dedicated link to the memory controller, the performance gain in
doing this can be significant. For example, with four DIMMs in a single
memory card, the x366 is about 50% slower than with four DIMMs across two
memory cards (two in each).
From a performance point of view, we do not recommend you operate the
server with only one memory card installed.
Use as many DIMMs as possible. The optimal configuration is to have all 16
slots in the server populated with DIMMs. The gains are not as significant as
with the use of memory cards (perhaps a 3% to 5% improvement) but this
method can still make a difference if this memory configuration is suitable for
your client.
3.7.2 Memory settings in BIOS
By default the Memory Array Settings is set to Redundant Bit Steering (RBS), as
shown in Figure 3-22, which provides better protection from multiple single-bit
errors.
Figure 3-22 Memory settings
Memory Card 1
Memory Card 2
Memory Card 3
Memory Card 4
Memory Array Setting
[ RBS (Redundant Bit Steering) ]
Memory Settings

130
Planning and Installing the IBM Eserver X3 Architecture Servers
However, if maximum of performance is required, choose High Performance
Memory Array (HPMA) in the menu Advanced Setup → Memory Settings. In
this mode, only one single-bit correctable error can be recovered in a chip select
group using ECC circuitry. See 3.2.5, “Memory configuration in BIOS” on
page 102 for details.
3.7.3 CPU settings in BIOS
By default, the X3 Architecture servers are optimized for database transaction
processing. This is achieved by enabling hardware prefetching on the
processors, which forces the processors to prefetch extra cache lines for every
request.
If you plan to run applications that do not take advantage of prefetch, such as
Java, file/print, or a Web server, then you can gain 10% to 20% by disabling
prefetch. To disable prefetch, select Advanced Setup → CPU Options and set
Processor Hardware Prefetcher to Disabled.
The default is Enabled. This setting affects all processors in the chassis.
For other CPU settings, refer to 3.1.2, “Processor configuration options” on
page 92.
3.7.4 PCI adapter placement
In x460 multi-node configurations, performance can also be impacted by the
installation of PCI cards as network adapters, Fibre Channel HBAs, and so on.
To distribute the load equally, we recommend you spread the placement of the
adapters across all the nodes.
3.8 ServerProven
The ServerProven program indicates the IBM and third party hardware
components and software that are supported by specific servers. More
information about this program can be found at:
http://www.ibm.com/servers/eserver/serverproven/compat/us/

Chapter 3. Hardware planning
131
3.9 Solution Assurance
Solution Assurance is a technical inspection of a completed solution design by
uninvolved third parties who are experts in its technology components, with the
purpose of answering three questions:
Will the solution work? That is, is it technically viable?
Can we implement it successfully? Is the implementation plan sound?
Will it meet the client’s requirements and expectations?
The key aspects of the Solution Assurance process are the following:
The inspection examines an entire solution, not merely the configuration of an
individual product. Solutions typically are comprised of a mixture of hardware,
software, and services. They have a life cycle that transcends that of a sales
opportunity. Clients must be able to continue to operate a solution in
production for its designated life span. Solution Assurance therefore
considers:
– Hardware components
– Software components
– Activities and services necessary to successfully implement the solution in
a client’s environment
– Activities and services necessary for the client to successfully operate the
solution in a production environment on a continuing basis
A key document in this process is the Solution Assurance Product Review
Guide.
The inspection concerns itself solely with the technical viability of the
proposed solution. It does not address aspects such as pricing, contracting,
terms and conditions, profitability, or assessing the client’s financial stability,
except insofar as these aspects can affect technical issues. These elements
are dealt with elsewhere in the Quality Assurance process. Neither does the
inspection address the availability of necessary resources, although it does
verify that all necessary resources have been identified.
The Solution Assurance process requires a completed solution. It occurs after
the designers deem the design finished.
The subject matter experts who act as Solution Assurance reviewers should
not be part of the solution design team. It is an important element of Solution
Assurance that those doing the technical assessment bring fresh eyes to the
solution so that they can examine not only the mechanics of the solution itself,
but also double-check all of the assumptions that were made in creating the
solution.

132
Planning and Installing the IBM Eserver X3 Architecture Servers
The technical reviewers must possess expertise in the technology
components. Most solutions involve many components, which implies that
most reviews will be conducted by a panel of reviewers that collectively cover
the necessary technical areas. The word
expert
also implies that the
reviewers must have credentials that command the respect of the design
team. Otherwise, their judgments will not carry weight.
The purpose of the three questions listed at the beginning of this introduction is
to form the terms of reference for the review panel. These questions attempt to
ascertain whether or not the solution will work as designed, whether there is a
feasible and complete implementation plan to implement the solution in the
client’s environment, and whether the combination of these two elements will
indeed satisfy all known client requirements and expectations.
Solution Assurance is not:
A session to design a solution
The first time technical details are considered
Detailed configuration assistance
A business case review
A client financial health review
A way to get an IBM IT Specialist involved
A transfer of responsibility for the solution to the Solution Assurance
reviewers
A cure-all that will catch all problems or replace the need for good solution
design
Not all solutions require a review by independent technical subject matter
experts, also known as an
expert

review
. The criteria used to determine when
that is necessary are subject to periodic change. The current criteria are
determined by referring to the Solution Assurance Web site or by using the
Trigger Tool, linked to from the Web site. If an expert review is needed, it should
be scheduled after the solution design, and before the proposal is due to be
delivered to the client.
Even if your proposal does not require a Solution Assurance Review, you should
still review the items in the Solution Assurance check list to ensure that you have
covered everything.
The Solution Assurance Web sites are as follows:
For IBM employees:
http://w3.ibm.com/support/assure

Chapter 3. Hardware planning
133
For Business Partners:
http://www.ibm.com/partnerworld/techsupport
Select Design Solutions on the left and then select Solution Assurance at
the bottom from the links provided. The above link requires a PartnerWorld®
ID and password. If you do not have this information or have forgotten it,
please contact PartnerWorld. For contact details, see:
http://www.ibm.com/partnerworld/pwhome.nsf/weblook/cpw_index.html
3.9.1 Solutions Assurance Reviews
Some level of Solution Assurance Review (SAR) should be performed on all IBM
solutions. The level of SAR (self, peer, or expert) should match the complexity of
the solution. For example, simpler solutions may need only a self review.
However, a combination of the client environment risk combined with the
complexity of the solution may require that an expert level SAR take place,
facilitated by a Quality Assurance practitioner, and supported by a team of
technical experts.
The three levels of Solution Assurance are:
Self
In a self review, the solution designer checks his or her own work using a
general or product-specific checklist, as appropriate.
Peer
A peer review is conducted by a peer or colleague of the solution design team
using available checklists, general and product-specific. The solution
designer explains the solution to the peer reviewers, who assess the
technical viability of the solution and record their action items and comments
in the checklist.
Expert
During an expert review, the solution design team explains the solution to one
or more subject matter experts. A Solution Assurance Quality Assurer
facilitates the expert review and records the results in the Solution Assurance
database used by that geography.
The xSeries SAR Trigger Criteria determine if an expert-level review is required
for an x460 solution. See the SA document SA200 for details. IBM employees
can use the following link:
http://w3.ibm.com/support/assure/assur30i.nsf/PubAllNum/SA200?OpenDocum
ent

134
Planning and Installing the IBM Eserver X3 Architecture Servers
For further information about what is required, refer to the following documents in
the Solution Assurance Web sites listed on page 132 for more information in your
specific geography:
For the Americas, reference document SA447 or contact TechXpress at
1-800-426-5525 in North America or 770-835-9700 in Latin America.
For EMEA, reference documents SA424 or SA359.
For Asia Pacific, reference document SA441 or go to:
http://w3.ibm.com/support/ap/
Select SA Contacts/SA Information.

© Copyright IBM Corp. 2006. All rights reserved.
135
Chapter 4.
Operating system
installation
This chapter describes operating system specifics and system hardware settings
that you may need to configure prior to installing an operating system. The topics
in this chapter are:
4.1, “Supported operating systems” on page 136
4.2, “Updating BIOS and firmware” on page 140
4.3, “Available memory in multi-node configurations” on page 141
4.4, “Microsoft Windows Server 2003” on page 142
4.5, “Datacenter offerings” on page 147
4.6, “Microsoft Windows 2000 Server” on page 152
4.7, “Red Hat and SUSE LINUX” on page 153
4.8, “VMware ESX Server 2.5.1” on page 155
The installation instructions for many of the supported operating systems can be
found at the OS installation matrix at:
http://www.pc.ibm.com/support?page=MIGR-4QLNTQ
4

136
Planning and Installing the IBM Eserver X3 Architecture Servers
4.1 Supported operating systems
For the latest operating system support information, refer to the ServerProven
operating system support matrix at:
http://www.pc.ibm.com/us/compat/nos/matrix.shtml
The operating systems shown in Table 4-1 are planned to be supported or are
now supported.
Table 4-1 Operating system support (current and planned)
Specific service packs or updates may be required, and some operating systems
may be limited as to the number of processors and nodes supported.
Operating system x260 x366 x460 MXE-460
Microsoft Windows Server 2003, Standard Edition
Yes
Yes
Yes
Yes
Microsoft Windows Server 2003, Standard for EM64T
Yes
Yes
Yes
Yes
Microsoft Windows Server 2003, Enterprise Edition
Yes
Yes
Yes
Yes
Microsoft Windows Server 2003, Enterprise for EM64T
Yes
Yes
Yes
Yes
Microsoft Windows Server 2003, Datacenter Edition No No
Yes
Yes
Microsoft Windows Server 2003, Datacenter for EM64T No No
Yes
Yes
Microsoft Windows 2000 Advanced Server
Yes
Yes
Yes
Yes
SUSE LINUX Enterprise Server 9 for x86
Yes
Yes
Yes
Yes
SUSE LINUX Enterprise Server 9 for EM64T
Yes
Yes
Yes
Yes
Red Hat Enterprise Linux 3 AS for x86
Yes
Yes
Yes
Yes
Red Hat Enterprise Linux 3 AS for EM64T
Yes
Yes
Yes
Yes
Red Hat Enterprise Linux 4 AS for x86
Yes
Yes
Yes
Yes
Red Hat Enterprise Linux 4 AS for EM64T
Yes
Yes
Yes
Yes
VMware ESX Server 2.5.1
Yes
Yes
Yes
Yes
Novell NetWare 6.5
Yes
Yes
Yes
Yes

Chapter 4. Operating system installation
137
4.1.1 Operating systems scalability
The X3 Architecture servers support (or plan to support) operating systems up to
32-way configurations, as described in Table 4-2.
Table 4-2 Operating systems scalability
4.1.2 Hyper-Threading support
As described in 1.8, “Processors” on page 18, Hyper-Threading Technology
allows a single physical processor to appear to the operating system and
applications as two logical processors. The logical processors share the core
processing engine of the physical processor and can execute code streams
concurrently.
Operating systems must be Hyper-Threading aware before they can see the
additional processors. If they are aware, they will see twice as many CPUs as
there really are.
Operating system 4-way 8-way 16-way 32-way
Windows Server 2003 SP1 Standard for x86
Yes No No No
Windows Server 2003 SP1 Standard for x86-64
Yes No No No
Windows Server 2003 SP1 Enterprise for x86
Yes
Yes No No
Windows Server 2003 SP1 Enterprise for x86-64
Yes
Yes No No
Windows Server 2003 Datacenter for x86
Yes
Yes
Yes
Yes
Windows Server 2003 Datacenter for x86-64
Yes
Yes
Yes
Yes
Red Hat Enterprise Linux AS 3 for x86
Yes
Yes No No
Red Hat Enterprise Linux AS 3 for x86-64
Yes
Yes No No
Red Hat Enterprise Linux AS 4 for x86
Yes
Yes
Yes No
Red Hat Enterprise Linux AS 4 for x86-64
Yes
Yes
Yes No
SUSE LINUX Enterprise Server 9 for x86
Yes
Yes
Yes No
SUSE LINUX Enterprise Server 9 for x86-64
Yes
Yes
Yes No
VMware ESX Server 2.5.1
Yes
Yes
Yes No
Windows 2000 Server, Advanced Server
Yes
Yes No No
NetWare 6.5
Yes No No No

138
Planning and Installing the IBM Eserver X3 Architecture Servers
In addition, to fully take advantage of Hyper-Threading, the operating system’s
scheduler must allocate resources with the full knowledge of which logical
processors are part of the one physical CPU. For example, consider a server
with two physical processors with Hyper-Threading enabled (that is, four logical
processors). If the operating system 's scheduler is unaware of Hyper-Threading
(“Yes” in the Hyper-Threading column in Table 4-3 on page 139), it would treat all
four logical processors the same. As a result, if two processes are eligible to run,
the scheduler might put those processes on the two logical processors that are
part of the one physical processor. Thus, one physical CPU would be busy while
the other CPU is idle, leading to poor overall performance.
Simply enabling Hyper-Threading might not guarantee improved overall system
performance. To get a benefit from Hyper-Threading, the operating system and
server applications need to be capable of detecting the additional logical
processors and spawning multiple threads that can exploit the additional
processing power.
In addition to considering whether the operating system you are installing
supports Hyper-Threading, there can be licensing implications to weigh before
enabling Hyper-Threading technology.
For a more detailed discussion of Hyper-Threading Technology, refer to:
http://www.intel.com/technology/hyperthread
Table 4-3 on page 139 lists the level of support for Hyper-Threading technology
provided by the operating systems.
In the Hyper-Threading column:

Yes
indicates that the operating system recognizes the logical processors and
can execute threads on them, but is not optimized for Hyper-Threading. From
a licensing perspective, if Hyper-Threading is enabled, the operating system
will need to be licensed for twice the number of physical processors to take
full advantage of the processors’ capabilities.

Optimized
indicates that the operating system recognizes the logical
processors and that the operating system code has been designed to take full
advantage of the technology. From a licensing perspective, the logical
processors do not count toward the number of processors for which an
operating system is licensed.

Chapter 4. Operating system installation
139
Table 4-3 Supporting Hyper-Threading technology by operating systems
Hyper-Threading Technology is enabled by default on x260, x366, and x460. To
disable it, if necessary, do the following:
1.Press F1 during system startup to enter the System Configuration Utility.
2.From the main menu, select Advanced Setup → CPU Options. Figure 4-1
appears.
Figure 4-1 Hyper-Threading setting
Operating system Release Hyper-Threading
Windows 2000 Advanced Server SP4 Yes
Windows Server 2003, Standard Edition SP1 Optimized
Windows Server 2003, Standard for EM64T SP1 Optimized
Windows Server 2003, Enterprise Edition SP1 Optimized
Windows Server 2003, Enterprise for EM64T SP1 Optimized
Windows Server 2003, Datacenter Edition Initial Optimized
Windows Server 2003, Datacenter for EM64T Initial Optimized
Red Hat Enterprise Linux AS for x86 3 Update 2 Optimized
Red Hat Enterprise Linux AS for EM64T 3 Update 2 Optimized
Red Hat Enterprise Linux ES for x86 3 Update 5 Optimized
Red Hat Enterprise Linux ES for EM64T 3 Update 5 Optimized
SUSE LINUX Enterprise Server for x86 9 Optimized
SUSE LINUX Enterprise Server for EM64T 9 Optimized
VMware ESX Server 2.5.1 Yes
NetWare 6.5 Optimized
Hyperthreading Technology
Clustering Technology
Processor Adjacent Sector Prefetch
Processor Hardware Prefetcher
Processor Execute Disable Bit
[ Enabled ]
[ Logical Mode ]
[ Enabled ]
[ Enabled ]
[ Disabled ]
CPU Options

140
Planning and Installing the IBM Eserver X3 Architecture Servers
3.With Hyper-Threading Technology selected, press the right arrow key to
change the value to Disabled.
4.Save changes and exit the System Configuration Utility.
On x460 multi-node configurations, all the nodes in a partition must have the
same Hyper-Threading setting. This will need to be set individually on each node.
For descriptions of the other options in this menu, see 3.1.2, “Processor
configuration options” on page 92.
4.2 Updating BIOS and firmware
We recommend you check the BIOS and firmware levels on the items listed
below and, as part of your installation procedure, update them to the most
current revision in the following order:
System BIOS. On x460 multi-node configurations, you can update the BIOS
version on all chassis from the primary node. You will be prompted as shown
in Figure 4-2.
Figure 4-2 BIOS update on an x460 multi-node configuration
Onboard diagnostics. On x460 multi-node configurations, you will need to
update the Diagnostics on each node separately. If you have an RSA II
installed, you can use the Remote Console feature with a remote diskette and
boot each node from the diagnostics diskette.
BMC firmware. On x460 multi-node configuration, you will need to update the
firmware separately on each node. If you have an RSA II installed, you can
use the Remote Console feature with a remote diskette and boot each node
from the BMC firmware diskette.
Remote Supervisor Adapter II SlimLine firmware. You can update the RSA
firmware from its Web interface. The update involves uploading two separate
PKT files before you restart the adapter. In the x460 multi-node configuration,
you should update the Diagnostics separately on each node
Additional devices if installed, such as ServeRAID adapters and Fibre
Channel host bus adapters.
Would you like to automatically update each NODE to the BIOS image
contained on this diskette? [Y/N]
Note: If you choose no, you will be prompted to confirm update
for each NODE detected.

Chapter 4. Operating system installation
141
The latest BIOS and firmware code can be found at the driver matrix index page:
http://www.pc.ibm.com/support?page=MIGR-4JTS2T
Follow the installation instructions provided with each package.
If you use Microsoft Windows Datacenter, use the driver and firmware matrix
located on the following page:
http://www.pc.ibm.com/support?page=MIGR-4P7RG3
There are a number of methods you can use to update this code, including:
Local diskette drive (x260) or external USB diskette drive, part number
05K9276 (x366 and x460)
The Remote Console feature of the RSA II SlimLine adapter (when installed),
along with the remote diskette feature. See the redbook IBM Eserver
xSeries and BladeCenter Server Management, SG24-6495 for details.
Using the bootable UpdateXpress CD-ROM (if supported; see the list of
servers), available from:
http://www.pc.ibm.com/support?page=MIGR-53046
4.3 Available memory in multi-node configurations
As described in 1.9, “XceL4v cache” on page 27, the X3 Architecture servers use
a new technology called XceL4v Dynamic Server Cache as an replacement of
the L4 cache used in x440 and x445 servers.
In the x260, x366, and single-node configurations of the x460, the XceL4v cache
works as a snoop filter to reduce traffic between the two front-side buses and it
uses the memory integrated into the chipset for storage.
In multi-node configurations of x460, 256 MB of main memory in each node is
allocated to the XceL4v cache, in addition to building in a chipset for use as an
L4 cache for saving memory addresses of data, which are currently located in all
the processor’s caches; this reduces traffic between different nodes.

142
Planning and Installing the IBM Eserver X3 Architecture Servers
The available memory in multi-node configuration will be less then installed
according to Table 4-4.
Table 4-4 Amount of subtracted memory in multi-node configuration
For example, when you boot a two-node configuration, you will see the
messages shown in Figure 4-3.
Figure 4-3 Memory allocated to the XceL4v scalability cache
4.4 Microsoft Windows Server 2003
There are three different Microsoft Windows Server 2003 releases:
Windows Server 2003: For 32-bit and EM64T/AMD64 systems
Windows Server 2003, x64: For EM64T and AMD64-based servers
Microsoft Windows Server 2003 64-bit: For Itanium systems
The X3 Architecture servers use Intel Xeon MP processors with EM64T 64-bit
extensions. As such, the 32-bit releases and x64 releases of Windows Server
2003 are supported. The 64-bit Itanium release, which Microsoft refers to as
Number of nodes Total amount of memory allocated to XceL4v
One 0
Two 512 MB
Four 1 GB
Eight 2 GB
Chassis Number Partition Merge Status Installed Memory
1 Primary 8GB
2 Merged 8GB
Partition merge successful
16384 MB Memory: Installed
00512 MB Memory: Consumed by Scalability
8 Processors Installed

Chapter 4. Operating system installation
143
64-bit
as opposed to
x64
, are not supported on the X3 Architecture servers. The
Itanium release only runs on servers with the Itanium 2 processor, such as the
xSeries 455.
Each of the above three releases come in three or four editions: Standard,
Enterprise, Datacenter, and Web. These editions are compared in Table 4-5.
Datacenter Edition is discussed in more detail in 4.5, “Datacenter offerings” on
page 147.
The latest supported levels of Windows can be found at the ServerProven site:
http://www.ibm.com/servers/eserver/serverproven/compat/us/nos/matrix.shtml
Table 4-5 Features of the Windows Server 2003 family
Features Standard
Edition
Enterprise
Edition
Datacenter
Edition
Web
Edition
Edition availability
32-bit release Yes Yes Yes Yes
x64 release (EM64T & AMD64) Yes Yes Yes No
64-bit release (Itanium) Yes Yes Yes No
Scalability
Processors 1-4 1-8 1-32 1-2
Number of x460 nodes One Two Eight One
Memory — 32-bit 4 GB 32 GB 64 GB 2 GB
Memory — x64 (64-bit) 32 GB 1 TB
1
1 TB
1
N/A
Hyper-Threading Yes Yes Yes Yes
Hot-add memory No Yes Yes No
NUMA support No Yes Yes No
Directory Services
Active Directory Yes Yes Yes Limited
Metadirectory support No Yes Yes No
Security Services
Internet Connection Firewall Yes Yes No Yes
PKI, Certify services, Smart
Cards
Limited Yes Yes Limited

144
Planning and Installing the IBM Eserver X3 Architecture Servers
Terminal services
Remote Desktop for
Administration
Yes Yes Yes Yes
Terminal Server Yes Yes Yes N/A
Terminal server Session
Directory
No Yes Yes No
Clustering technologies
Network Load Balancing Yes Yes Yes Yes
Cluster Service No 1-8 nodes 1-8 nodes No
Communication and networking
VPN Support Yes Yes Yes Limited
Internet Authentication Service
(IAS)
Yes Yes Yes No
Network Bridge Yes Yes Yes No
Internet Connection Sharing Yes Yes No No
IP6 Yes Yes Yes Yes
File and print services
Distributed File System (DFS™) Yes Yes Yes Yes
Encrypting File System (EFS) Yes Yes Yes Yes
Shadow Copy Restore Yes Yes Yes Yes
Removable and Remote
Storage
Yes Yes Yes No
Management services
Remote Installation services
(RIS)
Yes Yes Yes No
Remote OS Installation Yes Yes Yes Yes
Note 1: The x460 is limited to 64 GB per node. This means 128 GB in a two-node
configuration and 512 GB in an eight-node configuration.
Features Standard
Edition
Enterprise
Edition
Datacenter
Edition
Web
Edition

Chapter 4. Operating system installation
145
With regard to Hyper-Threading, Windows Server 2003 operating systems
understand the concept of physical processors versus logical processors. In the
case of Windows Server 2003, only physical processors will count against the
license limit. For example, Windows Server 2003, Standard Edition running on a
two-way system with Hyper-Threading enabled will recognize and use the
processing capabilities of both physical and logical processors. However, only
the two physical processors will be counted for licensing purposes.
A custom Hardware Abstraction Layer (HAL) is required for all Windows
implementations, when installed on a multi-node x460. This HAL is included in
Service Pack 1 for all supported 32-bit versions and also in 64-bit versions of
Windows Server 2003.
4.4.1 Installing Windows Server 2003
Instructions on how to install Windows Server 2003 and the necessary drivers
are available on the xSeries support site in the support documents listed in
Table 4-6.
The latest supported levels of Windows can be found at the ServerProven site:
http://www.ibm.com/servers/eserver/serverproven/compat/us/nos/matrix.shtml
Table 4-6 xSeries support documents for Windows Server 2003
For installation instructions not listed, see:
http://www.pc.ibm.com/support?page=MIGR-4QLNTQ
Tip: The URLs associated with these support documents are of the following
form, where xxxxx is the number listed in the table:
http://www.pc.ibm.com/support?page=MIGR-xxxxx
In the PDF version of this redbook, the links in the table below are clickable.
Server Driver matrix
Windows Server 2003
installation instructions
x64 editions 32-bit editions
xSeries 460
MIGR-59923 MIGR-60676 MIGR-61178
xSeries 366
MIGR-59142 MIGR-59608 MIGR-62043
xSeries 260
MIGR-61523 MIGR-62497 MIGR-62498
All servers
MIGR-4JTS2T MIGR-4QLNTQ MIGR-4QLNTQ

146
Planning and Installing the IBM Eserver X3 Architecture Servers
As well as the normal method of using the Windows installation CD-ROM, you
can also install Windows using ServerGuide, available from:
http://www.pc.ibm.com/support?page=MIGR-4ZKPPT
Pre-install information
The key points to the installation are as follows:
You should update all firmware before starting. Download them from the
above driver matrix Web page. This includes:
– System BIOS
– Diagnostics
– BMC firmware
– RSA II firmware (if installed)
– ServeRAID firmware (if installed)
If you are using a ServeRAID-8i RAID controller, use the supplied CD to
configure an array or download the latest configuration CD from the above
driver matrix page.
In BIOS, ensure the following parameters are set:
– In CPU Options, ensure that the Clustering Technology parameter is set to
Logical Mode, as shown in Figure 3-1 on page 92.
– In Advanced Settings → RSA II Settings, ensure that the OS USB
Selection setting is set to Other OS, as shown in Figure 5-3 on page 167.
If you plan to install Windows using a regular Windows installation CD, you
will need a USB diskette drive to supply the necessary boot device drivers.
If you are planning to boot from the internal SAS drives and will be installing
using the Windows installation CD-ROM, you will need to press F6 when you
see the Setup is inspecting your computer's hardware configuration
message, and then insert a driver diskette. The required drivers are one of
the following:
– Adaptec SAS drivers if you are booting from the internal SAS drives and
you do not have a ServeRAID-8i installed
– ServeRAID-8i driver if you are booting from the internal SAS drives and
you do have a ServeRAID-8i installed (SAS driver is not needed.)
Download these drivers from the above driver matrix URLs. If you are using
ServerGuide to install Windows, you do not need to obtain these drivers
separately.
If you do not have a USB diskette drive, but you do have the RSA II adapter
installed in the server, you can also install Windows remotely using the

Chapter 4. Operating system installation
147
remote console and remote media functions of the RSA II. In this instance,
you would put the Windows CD-ROM in the server and the driver diskette in
the drive of a remote workstation and control the server remotely via the RSA
II Web interface. See 3.5, “Remote console and remote media”, in the IBM
Redbook IBM Eserver xSeries and BladeCenter Server Management,
SG24-6495 for details.
Post-installation information
The key points to the installation are as follows:
After installation, you will need to install additional drivers. Consult the
post-install steps in the installation instructions (from the above URLs). In
addition, you will need to install:
– The OSA IPMI driver (for the BMC)
– The RSA II driver
If you are installing a 32-bit version of Windows and you have more than 4 GB
of RAM installed, you should add the /PAE switch to the boot.ini file once
installation is complete, so that the operating system can access the memory
about the 4 GB line (see the last line in Example 4-1).
Example 4-1 boot.ini example for accessing more than 4 GB memory
[boot loader]
timeout=3
default=multi(0)disk(0)rdisk(1)partition(1)\WINDOWS
[operating systems]
multi(0)disk(0)rdisk(1)partition(1)\WINDOWS="Windows Server 2003, Enterprise" /fastdetect /PAE
4.5 Datacenter offerings
Windows Server 2003, Datacenter Edition is for highly scalable network
operating systems designed for mission-critical, enterprise-wide applications.
Datacenter Edition is suited for four-node x460 configurations: It enables full use
of the large number of processors and large amounts of RAM that can be
installed. High-volume online transaction processing, large-scale data
warehousing, and scientific simulations are some of the applications for which
Datacenter is optimized.
The complete IBM Datacenter Solution Program can be found at:
http://www.ibm.com/servers/eserver/xseries/windows/datacenter.html

148
Planning and Installing the IBM Eserver X3 Architecture Servers
The x460 is fully certified to run Windows Server 2003, Datacenter Edition, up to
an 8-node configuration. See the Windows Server Catalog for details:
http://www.microsoft.com/windows/catalog/server
There are two options available for clients who want to run Windows Server
2003, Datacenter Edition:
IBM Datacenter High Availability Program Offering
IBM Datacenter Scalable Offering
Option 1: IBM Datacenter High Availability Program Offering
The Datacenter High Availability Program Offering delivers a fully configured,
Datacenter-certified, preinstalled hardware and software operating system
solution on 4-way through 32-way server configurations that support up to 8 node
Microsoft Cluster Certified solutions for a tightly controlled, end-to-end supported
environment for maximum availability.
This end-to-end offering will deliver a fully configured, certified, and preinstalled
system for clients who want to maintain a tightly controlled environment for
maximum availability. To maintain this high availability, the solution must be
maintained as a certified configuration.
Table 4-7 on page 149 shows the models for both 32-bit and 64-bit versions of
the operating system.
With the High Availability Program Offering, the client selects memory, disks, and
other options and IBM installs them at the factory. The x460 Datacenter models
come with four processors, but no memory or disks, in order to provide maximum
flexibility in configuration. The operating system is preloaded at the factory. Also
shipped with the system is a recovery CD, OS documentation, and a 4-CPU
Certificate of Authenticity (COA) to license the system.
The MXE-460 models have no processors, memory, or disks (allowing the client
to specify any combination), but have a 4-CPU Certificate of Authenticity (COA)
to license the system.
Unlike the x440 and x445 IBM Datacenter Offerings, it is no longer necessary to
order the operating system separately. With the x460, the operating system is
included when you order the system.
Important: If you plan to implement Datacenter on a 32-way (8-node) x460
configuration, see 3.3.4, “Export regulations for 32-core complexes” on
page 118. Gaining approvals can take several months.

Chapter 4. Operating system installation
149
Table 4-7 Certified models for the High Availability Program Offering (Datacenter preloaded)
Note: Processors, memory, HDDs, ServeRAID, and so forth, must be ordered
as part of the Datacenter solution and will be configured in second-level
manufacturing before client shipment. The processors in all nodes must match
exactly. The HA models will be configured and tested in manufacturing.
Server Model Standard processors Cores/socket L2/L3 cache Std mem/disk
With 32-bit Windows Server 2003 Datacenter Edition preloaded
x460 8872-2Ax 4x 3.00 GHz Xeon MP Single-core 1 MB / 8 MB None / None
x460 8872-3Ax 4x 3.33 GHz Xeon MP Single-core 1 MB / 8 MB None / None
MXE-460 8874-1Ax (Match the x460) Single-core 1 MB / 8 MB None / None
x460 8872-5Ax 4x 2.66 GHz Xeon 7020 Dual-core 1+1 MB / 0 None / None
x460 8872-6Ax 4x 3.00 GHz Xeon 7040 Dual-core 2+2 MB / 0 None / None
MXE-460 8874-2Ax (Match the x460) Dual-core Varies None / None
With 64-bit Windows Server 2003 Datacenter Edition preloaded
x460 8872-2Bx 4x 3.00 GHz Xeon MP Single-core 1 MB / 8 MB None / None
x460 8872-3Bx 4x 3.33 GHz Xeon MP Single-core 1 MB / 8 MB None / None
MXE-460 8874-1Bx (Match the x460) Single-core 1 MB / 8 MB None / None
x460 8872-5Bx 4x 2.66 GHz Xeon 7020 Dual-core 1+1 MB / 0 None / None
x460 8872-6Bx 4x 3.00 GHz Xeon 7040 Dual-core 2+2 MB / 0 None / None
MXE-460 8874-2Bx (Match the x460) Dual-core Varies None / None
Note: Table 4-7 lists models ending in x. The
x
changes according to
geography. For example, 8872-3Ax will be 8872-3AU in the USA and
8872-3AG in EMEA. For country-specific models, check the announcement
letter.

150
Planning and Installing the IBM Eserver X3 Architecture Servers
Software Update Subscription
Clients should also purchase the annual Software Update Subscription for the
IBM Datacenter High Availability Program Offering. The Software Update
Subscription provides periodic updates to the Microsoft Windows Datacenter
operating system for maintenance and new versions, which you license for a
period of one year. This subscription also includes IBM updates to firmware and
device drivers certified by IBM and Microsoft for use with the Datacenter
Solution.
Subscriptions are cumulative, meaning that if you have a 16-way solution, you
will require four 4-CPU subscriptions (see Table 4-8).
Table 4-8 Software Update Subscriptions for the Datacenter High Availability Program
IBM builds, tests, and provides the complete certified package of these
components. IBM provides program updates as they become available for a
period of one year. A program update can contain a new version, release,
supplements, or service packs, as IBM determines, announced during the
subscription term. IBM does not guarantee that updates will be announced
during the annual term.
Option 2: IBM Datacenter Scalable Offering
The Datacenter Scalable Offering is ideal for clients who already have a
well-managed IT infrastructure and just want a Windows operating system that
scales from 4-way to 32-way and offers maximum performance and scalability in
a non-clustered environment.
Table 4-9 on page 151 shows the system models for both 32-bit and 64-bit
versions of the operating system. With this option, the x460 models come with
four processors, and no memory or disks. The system is shipped with the
Datacenter installation CD, OS documentation, recovery CD and a 4-CPU
Certificate of Authenticity (COA) to license the system.
The MXE-460 comes with no processors, memory or disks, and a 4-CPU
Certificate of Authenticity (COA) to license the system.
Software Update Subscriptions Order number
Windows Server 2003, Datacenter Edition, 32-bit, 1-4 CPUs 4816-GAX
Windows Server 2003, Datacenter Edition, 64-bit, 1-4 CPUs 4816-HAX
Tip: The Maintenance Update Subscription is no longer needed and has been
withdrawn. Maintenance is now included in the Software Update Subscription.

Chapter 4. Operating system installation
151
Unlike the x440 and x445 IBM Datacenter Offerings, it is no longer necessary to
order the operating system separately. With the x460, the operating system is
included when you order the system.
Table 4-9 Certified models for the Datacenter Scalable offering without operating system
Software Update Subscription
Clients should also purchase the annual Software Update Subscription for the
IBM Datacenter Scalable Offering. For this offering, the Software Update
Subscription provides the client with operating system upgrade protection,
allowing the client to upgrade to the next Datacenter operating system release
for no additional charge. The Software Update Subscription is a one-year
license.
Server Model Standard processors Cores/socket L2/L3 cache Std mem/disk
With 32-bit Windows Server 2003 Datacenter Edition preloaded
x460 8872-2Dx 4x 3.00 GHz Xeon MP Single-core 1 MB / 8 MB 4 GB / None
x460 8872-3Dx 4x 3.33 GHz Xeon MP Single-core 1 MB / 8 MB 4 GB / None
MXE-460 8874-1Dx (Match the x460) Single-core 1 MB / 8 MB None / None
x460 8872-5Dx 4x 2.66 GHz Xeon 7020 Dual-core 1+1 MB / 0 4 GB / None
x460 8872-6Dx 4x 3.00 GHz Xeon 7040 Dual-core 2+2 MB / 0 4 GB / None
MXE-460 8874-2Dx (Match the x460) Dual-core Varies None / None
With 64-bit Windows Server 2003 Datacenter Edition preloaded
x460 8872-2Ex 4x 3.00 GHz Xeon MP Single-core 1 MB / 8 MB 4 GB/None
x460 8872-3Ex 4x 3.33 GHz Xeon MP Single-core 1 MB / 8 MB 4 GB/None
MXE-460 8874-1Ex (Match the x460) Single-core 1 MB / 8 MB None/None
x460 8872-5Ex 4x 2.66 GHz Xeon 7020 Dual-core 1+1 MB / 0 4 GB / None
x460 8872-6Ex 4x 3.00 GHz Xeon 7040 Dual-core 2+2 MB / 0 4 GB / None
MXE-460 8874-2Ex (Match the x460) Dual-core Varies None / None

152
Planning and Installing the IBM Eserver X3 Architecture Servers
Subscriptions are cumulative, meaning that if you have a 16-way solution, you
will require four CPU subscriptions (see Table 4-10).
Table 4-10 Software Update Subscription for the Datacenter Scalable Offering
4.6 Microsoft Windows 2000 Server
Instructions on how to install Windows 2000 Server and the necessary drivers
are available on the xSeries support site in the support documents listed in
Table 4-11.
Table 4-11 xSeries support documents for Windows 2000 Server
For installation instructions not listed, see:
http://www.pc.ibm.com/support?page=MIGR-4QLNTQ
Tip: Unlike the Datacenter High Availability Program Offering, maintenance
updates are not part of the Software Update Subscription. Instead, they are
available directly from Microsoft.
Software Update Subscriptions Order number
Windows Server 2003, Datacenter Edition, 32-bit, 1-4 CPUs 4818-GAX
Windows Server 2003, Datacenter Edition, 64-bit, 1-4 CPUs 4818-HAX
Tip: The URLs associated with these support documents are of the following
form, where xxxxx is the number listed in the table:
http://www.pc.ibm.com/support?page=MIGR-xxxxx
In the PDF version of this redbook, the links in the table below are clickable.
Server Driver matrix OS install instructions
xSeries 460
MIGR-59923
MIGR-62044
xSeries 366
MIGR-59142 MIGR-59215
xSeries 260
MIGR-61523
All servers
MIGR-4JTS2T MIGR-4QLNTQ

Chapter 4. Operating system installation
153
As well as the normal method of using the Windows installation CD-ROM, you
can also install Windows using ServerGuide, available at:
http://www.pc.ibm.com/support?page=MIGR-4ZKPPT
The prerequisites and important notes for installing Windows 2000 Server are
similar to those of Windows Server 2003, as listed in 4.4.1, “Installing Windows
Server 2003” on page 145.
You will need to install Service Pack 4 for Windows 2000 Server to bring the
operating system up to a supported level.
In addition, for multi-node x460 configurations, you will need to a custom
hardware abstraction layer (HAL) code, available at:
http://www.pc.ibm.com/support?page=MIGR-61296
You will also need an external diskette drive to install the HAL during the
operating system installation.
4.7 Red Hat and SUSE LINUX
Red Hat Enterprise Linux Advanced Server 3.0 update 5 and SUSE LINUX
Enterprise Server 9.0 SP2 are designed to exploit the capabilities of the X3
Architecture servers, including support for Hyper-Threading technology and the
64-bit extensions.
The latest supported levels of Red Hat and SUSE LINUX products can be found
at the ServerProven site:
http://www.ibm.com/servers/eserver/serverproven/compat/us/nos/matrix.shtml

154
Planning and Installing the IBM Eserver X3 Architecture Servers
Instructions on how to install Linux and the necessary drivers are available on
the xSeries support site in the support documents listed in Table 4-12.
Table 4-12 xSeries support documents for Linux
For installation instructions not listed, see:
http://www.pc.ibm.com/support?page=MIGR-4QLNTQ
The key points to installing Linux are as follows:
You should update all firmware before starting. Download them from the
above driver matrix Web page. This includes:
– System BIOS
– Diagnostics
– BMC firmware
– RSA II firmware (if installed)
– ServeRAID firmware (if installed)
If you are using a ServeRAID-8i RAID controller, use the supplied CD to
configure an array or download the latest configuration CD from the above
driver matrix page.
Tip: The URLs associated with these support documents are of the following
form, where xxxxx is the number listed in the table:
http://www.pc.ibm.com/support?page=MIGR-xxxxx
In the PDF version of this redbook, the links in the table below are clickable.
Server Driver matrix
OS Installation instructions
RHEL 3 RHEL4 SLES 9
xSeries 460
MIGR-59923 MIGR-62630 MIGR-62632 MIGR-62720
xSeries 366
MIGR-59142 MIGR-59187
MIGR-62583
MIGR-62584 MIGR-59157
MIGR-62688
xSeries 260
MIGR-61523 MIGR-62582 MIGR-62616 MIGR-62689
All servers
MIGR-4JTS2T MIGR-4QLNTQ MIGR-4QLNTQ MIGR-4QLNTQ

Chapter 4. Operating system installation
155
In BIOS, ensure the following parameters are set:
– In CPU Options, ensure that the Clustering Technology parameter is set
as follows (see Figure 3-1 on page 92):
• Special Mode for 64-bit RHEL 3.0 when using single-core processors
• Logical Mode for all other Linux distributions (including 32-bit RHEL)
and for all distributions on dual-core processors
– Select Advanced Settings → RSA II Settings and ensure that the OS
USB Selection setting is set to Linux OS, as shown in Figure 5-3 on
page 167.
You may need a USB diskette drive to supply the necessary boot device
drivers.
At the time of writing, PCI Hot Plug was not supported with RHEL 3 on these
systems.
4.7.1 Additional information for SLES 9
The following is of importance when installing SUSE LINUX Enterprise Server 9:
If the server contains a Remote Supervisor Adapter II, in the boot screen,
press F2 and scroll to chose 1024 x 768 resolution.
In the boot screen, choose Installation. SLES 9 Update 2 already has all the
necessary drivers; simply follow the installation wizard. If you are installing a
version of SLES 9 older than Update 2, follow the instructions documented in
Table 4-12 on page 154 to obtain the necessary driver diskettes.
If you are installing to a server that contains a Remote Supervisor Adapter II,
select Skip at any graphics configuration screens that are displayed
throughout the rest of the installation. After installing, do not reconfigure the
monitor or any other graphical settings.
If you must change the graphics settings, type the following command to start
the configuration utility:
sax2 -m 0=fbdev
4.8 VMware ESX Server 2.5.1
VMware ESX Server is virtual machine software for consolidating and
partitioning servers. It is a cost-effective, highly scalable virtual machine platform
with advanced resource management capabilities. VMware ESX Server is used
to minimize the total cost of ownership of the server infrastructure by maximizing
server manageability, flexibility, and efficiency across the enterprise.

156
Planning and Installing the IBM Eserver X3 Architecture Servers
ESX Server supports two-way SMP virtual machines. This support is made with
the add-on component VMware ESX Virtual SMP.
Note the following items if you are using ESX Server:
ESX Server must be ordered in conjunction with a supported xSeries server.
It cannot be ordered separately. If you want to order ESX Server separately,
contact VMware directly.
Each VMware ESX Server includes a one-year Software Update Subscription
Service from VMware. During this period, product and maintenance updates
are made available to registered owners. After the first year, this becomes a
fee-based offering.
In addition to the Software Update Subscription Service, support services are
provided on a fee basis through Support Line offerings which, in certain
countries, may be mandatory. VMware ESX Server is a supported product
under the Microsoft and Linux Support Line offerings. For additional
information, visit:
http://www.ibm.com/eserver/xseries/vmware
VMware ESX 2.5.1 supports the following maximum configuration:
Storage
– 16 host bus adapters
VMware File System (VMFS)
– VMFS-2: Up to 64 TB
– VMFS-1: Up to 2 TB
CPU
– Up to 16 physical processors per system
– Up to 80 virtual CPUs in virtual machines
– Up to 200 registered virtual machines
Memory
– 64 GB of RAM
Adapters
– Up to 64 adapters of all types,
– Up to 8 Gigabit Ethernet or 16 10/100 Ethernet ports per system
Table 4-13 on page 157 represents lists of supported guest operating systems in
ESX Server 2.5. SMP-capable means that guest operating system supports
VMware Virtual SMP for dual virtual CPU or more.

Chapter 4. Operating system installation
157
Table 4-13 Supported guest operating systems in VMware ESX 2.5.1
VMware ESX Server 2.5.1 supports IBM Director 4.21 or higher.
4.8.1 Support for applications running on ESX Server
Ensure that the applications you plan to run on VMware ESX Server are
supported by the application vendor.
Microsoft
See the following Microsoft support Web site for details about their support of
applications and operating systems running on ESX Server:
http://support.microsoft.com/kb/897615/
IBM software
If you are running IBM software, such as WebSphere, Lotus®, and Tivoli®
products on VMware ESX Server, you must have an IBM Remote Technical
Support ServicePac® or IBM VMware Support Line agreement through the
IBM Support Line or the IBM equivalent. You must have a current Software
Maintenance Agreement in order to receive support for the IBM software
products in this environment. Individual IBM software can announce a level of
client support beyond that described. If applicable, information about the
added support will be included in the specific product announcement letter.
Guest operating system SMP-capable
Windows Server 2003 (Enterprise, Standard, and Web Editions) Yes
Windows XP Professional (Service Pack 1 or 2) No
Windows 2000 Server (Service Pack 3 or 4) Yes
Windows 2000 Advanced Server (Service Pack 3 or 4) Yes
Windows NT® 4.0 (Service Pack 6a) No
Red Hat Linux 7.2, 7.3, 8.0, 9.0 No
Red Hat Enterprise Linux (AS) 2.1 (Update 5) and 3.0 (Update 3) Yes
SUSE Linux 8.2, 9.0, and 9.1 No
SUSE Linux Enterprise Server (SLES) 8 and 9.0 Yes
Novell NetWare 6.5 (Support Pack 1 and 2) 5, 6.0 (Support Pack 5)
and 5.1 (Support Pack 6 and 7)
No
FreeBSD 4.9 No

158
Planning and Installing the IBM Eserver X3 Architecture Servers
4.8.2 Pre-install information
Before you begin installing ESX Server, do the following:
You should update all firmware before starting. Download them from the
above driver matrix Web page. This includes:
– System BIOS
– Diagnostics
– BMC firmware
– RSA II firmware (if installed)
– ServeRAID firmware (if installed)
Table 4-14 xSeries support documents for VMware ESX Server
If you are using a ServeRAID-8i RAID controller, use the supplied CD to
configure an array or download the latest configuration CD from the above
driver matrix page.
In BIOS, ensure the following parameters are set:
– In CPU Options, ensure that the Clustering Technology parameter is set to
Logical Mode, as shown in Figure 3-1 on page 92.
– Select Advanced Settings → RSA II Settings and ensure that the OS
USB Selection setting is set to Linux, as shown in Figure 5-3 on
page 167.
ESX Server 2.5.1 includes the necessary drivers for both SAS and
ServeRAID-8i, so no additional drivers are typically needed to install the
operating system.
Tip: The URLs associated with these support documents are of the following
form, where xxxxx is the number listed in the table:
http://www.pc.ibm.com/support?page=MIGR-xxxxx
In the PDF version of this redbook, the links in Table 4-14 are clickable.
Server Driver matrix Installation instructions
xSeries 460
MIGR-59923
MIGR-60546

xSeries 366
MIGR-59142
xSeries 260
MIGR-61523
All servers
MIGR-4JTS2T MIGR-4QLNTQ

Chapter 4. Operating system installation
159
4.8.3 Installing ESX Server
Instructions on how to install ESX Server 2.5.1 are available at:
http://www.pc.ibm.com/support?page=MIGR-60546
The upgrade patches and the latest releases of drivers can be found on the
VMware Web site:
http://www.vmware.com/download/
Tip: Be sure to check the Issues section (5.0) in this installation guide. At the
time of writing, there was additional information there for multi-node x460
installations.

160
Planning and Installing the IBM Eserver X3 Architecture Servers

© Copyright IBM Corp. 2006. All rights reserved.
161
Chapter 5.
Management
Like the rest of the xSeries family of servers, the X3 Architecture servers have a
number of features that aid in the systems management of the servers. Topics
covered in this chapter are:
5.1, “IBM Director” on page 162
5.2, “Remote Supervisor Adapter II SlimLine” on page 164
5.3, “Baseboard Management Controller” on page 177
5.4, “Integrating the service processors with IBM Director” on page 187
5.5, “OSA SMBridge utility” on page 194
5.6, “Predictive Failure Analysis (PFA)” on page 220
5.7, “IBM Dynamic System Analysis” on page 220
5.8, “Partition management” on page 223
5

162
Planning and Installing the IBM Eserver X3 Architecture Servers
5.1 IBM Director
IBM Director is a client/server workgroup manager. With IBM Director, you can
view and track the hardware configuration of remote systems in detail and
monitor the usage and performance of critical components, such as processors,
disks, and memory.
All X3 Architecture servers require IBM Director Version 4.22 or later. The latest
version of IBM Director can be downloaded from:
http://www.pc.ibm.com/support?page=SERV-DIRECT
IBM Director is designed to manage an industry-standard server environment
and supports a variety of operating systems, including Windows, Linux, and
NetWare. The latest version, IBM Director 5.10, supports the full range of
IBM Eserver platforms, including Linux on zSeries, iSeries, pSeries, xSeries,
and BladeCenter. AIX on pSeries systems and i5/OS® on iSeries is also
supported.
IBM Director supports many industry standards, such as DMI, CIM, WMI, SNMP,
TCP/IP, IPX™, SNA, NetBIOS, SLIP, XML, and HTTP, among others.
The basic management functions included in IBM Director provide the ability to
get servers up and running quickly and smoothly, and to simplify the ongoing
hardware management. Additional powerful tools are available as extensions to
IBM Director to further strengthen its management capabilities.
For more information about IBM Director, see Implementing Systems
Management Solutions Using IBM Director, SG24-6188, found at:
http://www.redbooks.ibm.com/abstracts/sg246188.html
Other IBM Director plug-ins such as Virtual Machine Manager and Scalable
Systems Manager can be downloaded at:
http://www.ibm.com/servers/eserver/xseries/systems_management/xseries_s
m/dwnl.html
Integration with enterprise managers
Many clients already have an enterprise management system to manage all of
their computer systems using centralized alerting, problem determination, and
inventory. Because most of these enterprise solutions do not provide the
hardware-level information that is available from systems such as the xSeries
family, it is important to have tools such as IBM Director to enable integration
with these large enterprise solutions.

Chapter 5. Management
163
IBM Director offers a graphical user interface for easy local and remote access,
control, and smooth integration into higher levels of workgroup or enterprise
management tools, including:
Tivoli NetView®
Tivoli Enterprise™
HP OpenView
Microsoft System Management Server (SMS)
Computer Associates Unicenter
NetIQ
BMC Patrol (Morse Systems)
The NetIQ and BMC Patrol integration modules are available from those
companies.
By letting IT administrators view the hardware configuration of remote systems in
detail and monitor the usage and performance of critical components such as
processors, hard disk drives, power supplies, cooling fans, voltage regulator
modules (VRMs), and memory, IBM Director can help you manage your server
more efficiently and can help you control many of the hidden costs of operation.
For further information, see the IBM Redbook Integrating IBM Director with
Enterprise Management Solutions, SG24-5388, available at:
http://www.redbooks.ibm.com/abstracts/sg245388.html
The latest IBM Director upward integration modules are available at:
http://www.pc.ibm.com/support?page=SERV-DIRECT
Restriction: The IBM Director extension Scalable Systems Manger available
for xSeries 445 and 455 does
not
work with xSeries 460 and MXE 460.

164
Planning and Installing the IBM Eserver X3 Architecture Servers
5.2 Remote Supervisor Adapter II SlimLine
The Remote Supervisor Adaptor II SlimLine (RSA II), shown in Figure 5-1, is a
systems management card that ships with every x460 and MXE 460 and is
available as an option for the x260 and x366. This section describes how to
utilize the RSA II and its features for beneficial systems management.
Figure 5-1 Remote Supervisor Adapter II SlimLine
The most useful functions and features of the RSA II are:
Automatic notification and alerts
The RSA II automatically sends different types of alerts and notifications to
another server, like IBM Director, SNMP destination, or as e-mail directly to a
user by using SMTP.
Continuous health monitoring and control
The RSA II monitors all important system parameters like temperature,
voltage, and so on, continuously. If a fan fails, for example, the RSA II forces
the remaining fans to increase speed to compensate for the failing fan.
Web interface
The RSA II is managed via its built-in Web interface. From it, you can control
the server (power, remote media, and KVM), check sensors (fans,
temperature, and so on) and set monitors.
Event log
You can get access to the event logs of the server and the power-on-self-test
(POST) log and export them while the server is up and running.
Remote control
The RSA II card offers full remote control, including mouse, from power-up
and Setup/Diagnostics panels, all the way through to the operating system
running as normal. In fact, combined with the remote media function, you boot
the server and remotely install an operating system from a remote CD-ROM
using this feature.

Chapter 5. Management
165
Operating system failure screen capture
When the operating system hangs, for example, with a blue screen, you can
do a screen capture for support purposes. Additionally, the RSA II stores the
last failure screen in memory so you can refer to it later.
Remote media
As a part of the remote control feature, the remote media capability lets you
use diskette drives, diskette images, optical drives (such as DVD or
CD-ROM), or optical drive images of the system where the Web interface of
RSA II is running on the remote PC, and make them appear to be local drives
on the server.
Remote power control
The RSA II supports remote power control to power on, power off, or restart
the server with or without operating system shutdown over LAN or even WAN
connection.
The RSA II SlimLine family does not occupy a PCI slot. It is a small circuit board
that plugs into a dedicated socket on the I/O planar.
5.2.1 Installing the RSA II SlimLine
The RSA II SlimLine is installed by default in the x460 and MXE-460. However, if
you purchase the card for an x366 or x260 you will need to install it. Follow the
instructions in the Installation Guide that ships with the server. No additional
cabling is required.
Note: At the time of writing, support for optical images (ISO files) was
being rolled out to all servers that support RSA II for Windows only. Linux
support will come at a later date.

166
Planning and Installing the IBM Eserver X3 Architecture Servers
5.2.2 Connectivity
Once the RSA II SlimLine is installed, two ports on the rear of the server are
enabled. The location of these ports on the x366 is shown in Figure 5-2:
Service processor Ethernet port is for TCP/IP-based communication.
Service processor serial port is for serial/modem communication.
Figure 5-2 The RSA II SL ports (x366 shown)
5.2.3 Network settings
After installing the adapter in your server, you have to configure the network
settings to connect to the RSA II using the Web interface or telnet.
By default, the RSA II adapter is configured to look for a DHCP server to obtain
an IP address, and if none is available, to use the IP address 192.168.70.125.
To view the current DHCP-assigned address, enter Setup (press F1 at boot time
when prompted) and select Advanced Settings → RSA II Settings. The
assigned address will be listed on the second line, as shown in Figure 5-3 on
page 167.
If possible, we recommend you change this to always use a static IP address or
at least set up a reservation in your DHCP server. This is changed in Setup by
selecting Advanced Settings → RSA II Settings (see Figure 5-3 on page 167).
Tip: If the RSA II SlimLine is not installed, these two ports are non-functional.
Service processor
serial port
System serial port
Keyboard
Mouse
Service processor
Ethernet port
2x Gigabit
Ethernet
2x USB
IXA RS-485
(not service processor)
Video
The RSA II SL
uses the two
service processor
ports.

Chapter 5. Management
167
To configure the network settings, do the following:
1.Boot your server and press F1 to go to the BIOS settings.
2.Select Advanced Setup → RSA II Settings.
Figure 5-3 RSA II Settings in server BIOS
3.Change the DHCP control to Use Static IP by using the right and left arrow
keys. We recommend that you use a static IP address for the RSA II so that
you can still get access even if DHCP problems occur.
4.Fill in the IP address you want to assign to the RSA II, the network’s subnet
mask, and the standard gateway. Contact your network administrator for
details.
5.Select Other OS for Windows operating system or Linux OS as OS USB
Selection. Use the right and left arrow keys for selection.
The purpose of this selection is to prevent a known problem with Linux and its
generic human interface device (HID) driver. Linux cannot establish USB
communication with the RSA II using the generic HID (which Windows uses).
By selecting Linux OS here, it makes the RSA II appear as an OEM HID
instead of generic HID, which then functions properly.
Now select Save the Values and Reboot RSA II, with the arrow keys and press
Enter. Exit the utility.
Tip: To check the network connection of RSA II use the PING command from
another system connected to the network.
RSA II Settings
RSA II MAC Address
DHCP IP Address
DHCP Control
Static IP Settings
Static IP Address
Subnet Mask
Gateway
OS USB Selection
Save Values and Reboot RSA II
[ Other OS ]
[ 009.042.171.238 ]
[ 255.255.255.000 ]
[ 009.042.171.003 ]
00-09-6B-9E-08-5C
000.000.000.000
[ Use Static IP ]

168
Planning and Installing the IBM Eserver X3 Architecture Servers
5.2.4 Web interface
The RSA II can be controlled from a Web browser. As described in the previous
section, the adapter is configured by default to look for a DHCP server to obtain
an IP address, and if none is available, to use the IP address 192.168.70.125.
You can view the assigned address from the BIOS in the Advanced Settings
menu. Once you have the address, enter it into a Web browser.The default user
ID and password are USERID and PASSW0RD (where 0 is the number zero).

Chapter 5. Management
169
Figure 5-4 RSA II Web interface

170
Planning and Installing the IBM Eserver X3 Architecture Servers
5.2.5 Updating firmware
After the card is installed in the server, you will need to flash it with firmware that
is customized for your model, the x366 or x260. Download the firmware from the
Remote Supervisor Adapter II Family — Firmware and Drivers Quick Reference,
TIPS0534. In particular, download the Packet files for your server:
http://www.redbooks.ibm.com/abstracts/tips0534.html
The steps to update the firmware using the PKT files and the Web interface are
as follows:
1.Click the Packet files link for your server from the URL above.
2.Download the EXE file and save it to a local directory.
3.Run the EXE to extract the files. After extracting the file, take a few minutes to
read the readme.txt. The following files should be in your directory. Note that
there are two PKT files (Figure 5-5); you will need to perform the firmware
update procedure twice, once for each file.
Figure 5-5 Files of RSA II firmware update package
4.Connect to the RSA II using a Web browser by simply entering the IP address
in the address bar.
5.Log on to the RSA II with the default user USERID and PASSW0RD (with a
zero, not the letter O) as the password (unless you have changed it). For
security reasons, you should change the standard password before you put
your server into production.
Tip: IBM provides the packet (PKT) files so you can update the firmware using
the Web interface as we are describing here. In addition to that, there are also
Windows and Linux tools that let you update the firmware from within the
operating system running on the server. All three are listed in the above URL.
Tip: The firmware package suitable for update via a Web browser is
delivered as a ZIP file, and the ZIP file contains only two PKT files.

Chapter 5. Management
171
6.If this is a new adapter (x260 or x366) or new server (x460 or MXE-460), you
will be prompted to update the firmware (Figure 5-6). If not, select Tasks →
Firmware Update.
Figure 5-6 RSA II firmware update
7.Click Browse to select the first of two files for firmware update.
You should select the files in the correct order for updating the firmware. First,
select RAETBRUS.PKT (RSA Boot ROM) and then RAETMNUS.PKT (RSA
Main Application). Restart RSA only after applying both files.
8.To update, click Update. The file is now transferred to the RSA II.
9.Click Continue to flash the RSA II.
10.When prompted,
do not
restart the RSA adapter. You will do this after loading
the second firmware PKT file.
11.Repeat steps 6-9 for the second PKT file.
12.Restart the adapter by selecting ASM Control → ASM Restart.
5.2.6 Installing the device driver
Once you install your operating system, you will need to also install the driver for
the RSA II SlimLine adapter.
Download the driver via the appropriate link from the Remote Supervisor Adapter
II Family — Firmware and Drivers Quick Reference, TIPS0534, found at:
http://www.redbooks.ibm.com/abstracts/tips0534.html
Tip: If you have an RSA II, you only need to install the RSA II driver. You do
not need to also install the BMC drivers (although doing so will not have any
negative impact).

172
Planning and Installing the IBM Eserver X3 Architecture Servers
Windows service installation
The installation of the RSA II server software package is unlike the driver
installations of older systems management adapters. It is done by executing the
downloaded executable file.
The installation is as follows:
1.Execute the downloaded EXE file on the server with the RSA II.
2.Optionally, click Change to specify an alternate temporary folder for the
installation files.
3.The installation process starts automatically after the files are copied.
4.Follow the instructions.
5.When the installation finishes, you can delete the files in the temporary folder.
To determine if the installation was successful, check the services for the IBM
Remote Supervisor Adapter II by selecting Start → All Programs →
Administrative Tools → Services. Scroll down to the service IBM Remote
Supervisor Adapter II and verify that the status is started (see Figure 5-7).
Figure 5-7 RSA II service in Windows Server 2003
Note: If you have not already done so, change the setting OS USB Selection
to Other OS in the system BIOS, as shown in Figure 5-3 on page 167.

Chapter 5. Management
173
Linux daemon installation
To install the Linux daemon for the RSA II, first download it from the Remote
Supervisor Adapter II Family — Firmware and Drivers Quick Reference,
TIPS0534, found at:
http://www.redbooks.ibm.com/abstracts/tips0534.html
1.Select the correct rpm package for your Linux distribution (Red Hat or SUSE
LINUX).
2.Review the appropriate readme file of the rpm package for prerequisites and
installation steps.
3.Copy the downloaded file to a folder of the Linux server, for example,
/tmp/inst.
4.Install the daemon (for example, SUSE, where xx is the version) by running:
rpm -ivh ibmusbasm-1.xx.i386.rpm
Now you can check to see if the daemon is running. Use the ps command, as
shown in Example 5-1.
Example 5-1 Command to verify the RSA daemon is running
linux:~ # ps -ef | grep ibmasm
root 11056 1 0 10:47 pts/1 00:00:00 /sbin/ibmasm
root 11060 11056 0 10:47 pts/1 00:00:00 /sbin/ibmasm
root 11062 10996 0 10:48 pts/1 00:00:00 grep ibmasm
linux:~ #
If /sbin/ibmasm appears in the list, the daemon is running. The ibmusbasm
daemon is started automatically during the boot process of the operating system.
To start the daemon manually, use the command ibmspup. To stop the daemon,
enter ibmspdown.
Note: At the time of writing, drivers were not available for some distributions.
Note: If you have not already done so, change the setting OS USB
Selection to Linux OS in the system BIOS, as shown in Figure 5-3 on
page 167.

174
Planning and Installing the IBM Eserver X3 Architecture Servers
5.2.7 Remote console and media
To manage servers from a remote location, you often need more than just
keyboard-video-mouse (KVM) redirection. For example, for the remote
installation of an operating system or patches, you may need remote media to
connect a CD-ROM or diskette to the server.
The RSA II SlimLine offers the ability to make available a local diskette,
CD-ROM, or image to a remote server and have that server treat it as a local
USB-attached device.
Using remote media requires USB support from the operating system while the
OS is up and running or during the installation of the OS. Remote media works
with the following operating systems:
Windows Server 2003
Windows 2000 Server with Service Pack 4 or later
Red Hat Enterprise Linux AS 3, but not for OS installation
SUSE LINUX Enterprise Server 8, but not for OS installation
A Java runtime is required, which can be installed by going to:
http://www.java.com/en/download/manual.jsp
In the remote control window, there is a set of buttons that simulate specific
keystrokes and the video speed selector, as shown in Figure 5-8. The slider is
used to limit the bandwidth that is devoted to the remote console display on your
computer.
Figure 5-8 RSA II: Remote control buttons
Reducing the video speed can improve the rate at which the remote console
display is refreshed by limiting the video data that must be displayed. You can
Tip: It is possible to mount more than one remote drive concurrently. For
example, you could mount a CD-ROM and a diskette or diskette image.

Chapter 5. Management
175
reduce, or even stop, video data to allow more bandwidth for remote disk, if
desired. Move the slider left or right until you find the bandwidth that achieves the
best results.
Each of the buttons represents a key or a combination of keys. If you press a
button, the corresponding key stroke sequence will be send to the server. If you
require additional buttons, click Preferences, where you can modify or create
new key buttons.
The button bar can be detached by clicking anywhere in the grey background
and dragging. Drop the button bar to create a separate window.
Figure 5-9 Detached button bar
The Preferences link also lets you specify your keyboard and enable mouse
synchronization (that is, ensure the mouse pointer on the remote system
precisely follows the local mouse pointer). The following keyboard types are
supported:
US 104-key keyboard
Belgian 105-key keyboard
French 105-key keyboard
German 105-key keyboard
Italian 105-key keyboard
Japanese 109-key keyboard
Spanish 105-key keyboard
UK 105-key keyboard
For more information about the RSA II SlimLine, see Chapter 3, “Remote
Supervisor Adapter II” in IBM Eserver xSeries and BladeCenter Server
Management, SG24-6495.

176
Planning and Installing the IBM Eserver X3 Architecture Servers
5.2.8 TCP/UDP ports used by the RSA II SlimLine
The RSA II is using several TCP/UDP ports for communication. If the
communication with the RSA II passes through firewalls, it is important to know
which ports you have to enable on the firewalls to communicate with the RSA.
Table 5-1 shows the default ports. Remember when you change the ports in the
RSA you have to change them in the firewalls too.
Table 5-1 User configurable TCP/IP ports used by the RSA II
Some other ports are fixed and cannot be changed, as shown in Table 5-2.
Table 5-2 Fixed TCP/IP ports used by the RSA II
5.2.9 MIB files
The RSA II supports SNMP from many management tools, including IBM
Director. If you require MIB files, these can be found on the RSA II firmware
update for your server, in the ZIP file that also includes the PKT files. Links to the
appropriate firmware files are listed in technote Remote Supervisor Adapter II
Family — Firmware and Drivers Quick Reference, TIPS0534, available from:
http://www.redbooks.ibm.com/abstracts/tips0534.html
Port name Port number Description
http 80 (default) Web server HTTP connection (TCP)
https 443 (default) SSL connection (TCP)
telnet 23 (default) Telnet command-line interface connection (TCP)
SSH 22 (default) Secure Shell (SSH) command-line interface (TCP)
SNMP Agent 161 (default) SNMP get/set commands (UDP)
SNMP Traps 162 (default) SNMP traps (UDP)
Port number Description
427 SLP connection (UDP)
1044 Remote disk function (TCP)
1045 Persistent remote disk (disk on card) (TCP)
2000 Remote Console video redirect (TCP)
6090 IBM Director commands (TCP)
7070-7074 Partition management (TCP)

Chapter 5. Management
177
5.3 Baseboard Management Controller
The Baseboard Management Controller (BMC) is standard on all three X3
Architecture servers, and provides basic management features and
environmental monitoring. When errors or warnings occur, it controls the light
path diagnostics panel to aid in problem determination. It also provides limited
text-based remote control using Serial-over-LAN.
The BMC is based on an Hitachi 2166 chip and implements Version 1.5 of the
IPMI specification. The spec document is available from:
ftp://download.intel.com/design/servers/ipmi/IPMIv1_5rev1_1-012904ma
rkup.pdf
This integrated BMC has the following functions:
Monitoring of system voltages
Battery voltage monitor
System temperature monitors
Fan speed control
Fan tachometer monitor
Power Good signal monitor
System ID and planar version detection
System power control
System reset control
NMI detection
SMI detection and generation
Serial Port text redirection using SOL
Remind button detection
System LEDs control (power, HDD activity, alert, and so on)
Control of Lightpath LED
5.3.1 Connectivity
The BMC communicates via port 1 of the integrated Ethernet controller of each
of the X3 Architecture servers. To communicate with the BMC, you would attach
a standard Ethernet cable.
Figure 5-10 on page 178 shows the location of the port on the x366.
By default, the BMC is configured with an IP address of 10.1.1.97 (some servers
may have a default BMC address of 0.0.0.0). You will need to set a valid static IP
address using Setup in the BIOS or IBM Director. DHCP is currently not
supported.

178
Planning and Installing the IBM Eserver X3 Architecture Servers
Figure 5-10 The BMC is accessed using the system Gigabit Ethernet port 1 (x366
shown)
The servers have an RJ-45 port labeled “Systems Management”. This port is the
Remote Supervisor Adapter II SlimLine Ethernet connector, and for the x260 and
x366, is only active when the RSA II SL is installed.
In addition, there is a port labeled RS-485 and this port is only for use to connect
the server to an Integrated xSeries Adapter (IXA) if one is installed in the server.
They are
not
used to form an ASM interconnect network; the interconnect
network is not supported.
5.3.2 Updating the BMC firmware
It is a best practice to ensure that the BMC firmware is at the latest level to avoid
any issues.
The steps are as follows:
1.Download the firmware appropriate for your server. Refer to the IBM
Technote IBM Eserver xSeries BMC — Firmware and Drivers Quick
Reference, TIPS0532, and click the link for the firmware for the appropriate
server:
http://www.redbooks.ibm.com/abstracts/tips0532.html
Tip: Even if an RSA II SlimLine is installed in the server, you still access the
BMC using system Ethernet port 1.
Service processor
serial port
System serial port
Keyboard
Mouse
Service processor
Ethernet port
2x Gigabit
Ethernet
2x USB
IXA RS-485
(not service processor)
Video
The BMC is
accessed directly
via the system
Gigabit Ethernet
port 1.
Note: Updating the firmware of the BMC does not change any user settings.

Chapter 5. Management
179
Alternatively, you can navigate to the appropriate download page from the
software matrix:
http://www.pc.ibm.com/support?page=MIGR-4JTS2T
2.The BMC firmware update is available in bootable diskette form (an EXE file
that you run to extract the files to diskette) or as a ISO image file for CD
creation.
3.Insert the media into the server and boot the server. You may need to use
POST/BIOS setup to configure the correct boot devices.
4.Once your system starts, a Ramdrive will be created, the appropriate files will
be copied to this Ramdrive, and the BMC firmware update will execute
automatically.
5.3.3 Configuring the BMC in BIOS
Within the system BIOS, you are able to configure the following settings:
IP address
Subnet mask
Default gateway
The default IP address is 10.1.1.97 (some servers may have a default BMC
address of 0.0.0.0).
If you want to modify the default user ID and password, either use IBM Director
(the preferred method) or the bmc_cfg command-line tool supplied on the
firmware diskette.
The bmc_cfg command-line tool is described in detail in Chapter 2, “Baseboard
Management Controller” in IBM Eserver xSeries and BladeCenter Server
Management, SG24-6495. To get to a command prompt where you can run
bmc_cfg, boot to the firmware diskette/CD-ROM and exit from the firmware
update when prompted.
To configure the BMC using the BIOS, follow these steps:
1.Reboot the server if currently running. During post, press F1 to enter the
Configuration and Setup utility.
2.Select Advanced Setup.
Tip: If you have an RSA II also installed in your server, then you do not need
to enable the BMC network interface by supplying a static IP address. You can
perform all management functions from the RSA II Web interface or IBM
Director.

180
Planning and Installing the IBM Eserver X3 Architecture Servers
3.Select Baseboard Management Controller (BMC) Settings. The menu is
as shown in Figure 5-11.
Figure 5-11 BMC Settings panel in BIOS
4.Select BMC Network Configuration.
Figure 5-12 BMC network settings in BIOS
5.Enter the appropriate IP address, subnet mask, and gateway addresses, and
then select Save Network Settings in BMC.
Further parameters will need to be adjusted if you plan to use console redirection
and Serial over LAN. This can be used with a tool such as OSA SMBridge (see
5.5, “OSA SMBridge utility” on page 194).
5.3.4 Event Log
The BMC maintains a separate event log that contains, for example, entries like
power events, events from environmental sensors, and chipset specific entries.
You can access the BMC System Event Log (SEL) via the menu shown in
Figure 5-11 or using tools such as OSA SMBridge. This event log records all the
hardware alerts for the server. The event log displays one event per screen. Use
Baseboard Management Controller (BMC) Settings
BMC Firmware Version
BMC POST Watchdog
BMC POST Watchdog Timeout
System-BMC Serial Port Sharing
BMC Serial Port Access Mode
Reboot System on NMI
BMC Network Configuration
BMC System Event Log
01.05
[ Disabled ]
[ 5 Minutes ]
[Enabled ]
[Dedicated ]
[ Enabled ]
BMC Network Configuration
BMC MAC Address 00-11-25-3F-B7-AD
IP Settings
IP Address
Subnet Mask
Gateway
Save Network Settings in BMC
[ 009.042.171.054 ]
[ 255.255.255.000 ]
[ 009.042.171.003 ]

Chapter 5. Management
181
the Get Next Entry and Get Previous Entry links to page through the events, as
shown in Figure 5-13.
Figure 5-13 BMC Event Log
The system event log has room for 512 entries. You will be alerted if the log
reaches 75 percent or 90 percent full. However, unlike the RSA II or BladeCenter
management module, once the log is full, new entries are not saved and the LOG
LED on the light path diagnostics panel and the Information LED on the front
panel will be lit. You will need to clear the log in this instance using tools such as
SMBridge (command-line interface or SOL interface) or IBM Director.
Tip: You may notice that some events have a date stamp of 2070. When
power is removed from the server and later restored, the BMC clock resets to
1970 (which the BIOS displays as 2070). As soon POST completes and the
operating system is passed control, the BMC clock is updated with the correct
time from BIOS.
The incorrect time does not affect the timestamps of the events as received by
IBM Director. The events as seen in the IBM Director Event Log have the
correct time (the time the events were received) as long as the IBM Director
Management Server has the correct time.
BMC System Event Log
Get Next Entry
Get Previous Entry
Clear MBC SEL
Entry Number= 00001 / 00056
Record ID= 0001
Record Type= 02
Timestamp= 2005/08/04 10:18:42
Entry Details:Generator ID= 0020
Sensor Type= 09
Assertion Event
Power Unit
Sensor Specific Type
Power off/Power down
Sensor Number= 01

182
Planning and Installing the IBM Eserver X3 Architecture Servers
When an RSA II SlimLine is installed in the server, all events in the BMC System
Event Log are also made available to the RSA II. The RSA II also maintains a
separate log, and when viewed by accessing the RSA II, you will see both RSA II
and BMC-based events. However, the reverse is not true. You cannot view the
RSA II event log by viewing the event log in BIOS. The BIOS event log only
shows BMC-based events.
5.3.5 Remote control
The BMC supports remote control using the OSA SMBridge utility and Serial
over LAN. This provides a text-only console interface that lets you control BIOS
screens and specific operating system consoles. Both Linux and Windows
provide such text-only consoles (see 5.5, “OSA SMBridge utility” on page 194).
5.3.6 Installing the BMC device drivers
The device drivers are required to provide operating support and inband
communication with IBM Director. This section describes how to install the IPMI
device drivers on Windows and Linux platforms. The required device drivers are
listed in Table 5-3.
Table 5-3 IPMI required device drivers
The device drivers must be installed in a specific order or the installation will fail.
The order is as follows:
1.IPMI device driver
2.IPMI library (mapping layer)
3.IPMI ASR service
Tip: If you have an RSA II SL installed in your server, you do not need to
install the BMC drivers as described in this section. However, doing so will not
cause problems.
Device driver Additional comments
IPMI device drivers Required for in-band communication with IBM
Director.
IPMI Library This is the OSA BMC IPMI mapping layer. Includes
the BMC Mapping Layer, which maps the
dot.commands to IPMI commands. Required for
in-band communication with IBM Director.
ASR Server Restart software Required for ASR functionality.

Chapter 5. Management
183
To download the drivers appropriate for your server, refer to the IBM Technote
IBM Eserver xSeries BMC — Firmware and Drivers Quick Reference,
TIPS0532, and click the link for the firmware for the appropriate server:
http://www.redbooks.ibm.com/abstracts/tips0532.html
Alternatively, you can navigate to the appropriate download page from the
software matrix:
http://www.pc.ibm.com/support?page=MIGR-4JTS2T
Installing the device drivers on Windows
This section describes how to install the drivers under Windows.
IPMI device driver
To install the OSA IPMI device driver, follow these steps:
1.Run Setup.exe. After the usual initial windows, you will be prompted to select
a driver parameter, as in Figure 5-14.
Figure 5-14 Driver parameters
2.Click No.
Clicking No means that you want the driver to query SMBIOS record 38 for
the necessary device driver parameters. Clicking Yes means that you will
manually set these parameters, and this should only be used in unusual
situations as directed by IBM support. If you manually set an invalid
parameter, you may cause Windows to crash or reboot.
3.Click Next to begin the installation. When it completes, you will be prompted
to reboot the server, but the installer will not do this automatically.

184
Planning and Installing the IBM Eserver X3 Architecture Servers
IPMI mapping layer (library) files
To install the IPMI mapping layer (library) files, do the following:
1.Ensure that the IPMI device driver is installed before installing this software.
2.Download the EXE from the above Web site and run it.
3.Follow the on-screen instructions.
4.Reboot the server if the installation procedure prompts you to do so.
IPMI ASR service
To install the ASR service, do the following:
1.Ensure that the IPMI device driver and IPMI library files are installed before
installing this software.
2.Download the EXE from the above Web site and run it.
3.Follow the on-screen instructions.
4.Reboot the server if the installation procedure prompts you to do so.
Installing the device drivers on Linux
This section describes how to install the drivers under Linux.
IPMI device driver
To install the OSA IPMI device driver, launch a UNIX shell and enter the
following command to build and install the driver module on your system:
rpm -i osa_ipmi-x.x.x-x.i386.rpm
If you upgrade your Linux kernel, you should uninstall and then
recompile/re-install the OSA IPMI device driver. To rebuild the driver, change to
directory /usr/osa/osa_ipmi-x.x.x-x and enter:
sh build_osadrv
Notes:
You should install the necessary source code packages on your system. We
recommend that you install them in /usr/src.
Make sure you have gcc 3.2 available. You may also need to upgrade the
binutils package.
When installing on 32-bit SUSE LINUX 8.2, which has gcc Version 3.3
20030226 (pre-release) installed, insmod does not work without the -f option.
You should manually add the -f option to the insmod command in the
/sbin/ipmi_load script for the force loading. You may, however, receive a
warning that the kernel is tainted. Other gcc versions may cause the same
problem.

Chapter 5. Management
185
To uninstall the OSA IPMI device driver, enter one of the following:
rpm -e osa_ipmi-x.x.x-x
rpm -e osa_ipmi
See the README.TXT file available with the driver for more information.
IPMI mapping layer (library) files
The IBM mapping layer software is installed and removed via the Linux RPM
package management tool. Ensure that you have first installed the IPMI driver.
If this is an upgrade to an existing software package, remove the old version first
with the command:
rpm -e ibmsp6a
Depending upon your system's configuration, you may see messages about
missing files. These may be ignored.
To install the IPMI mapping layer (library) files, issue the following commands:
rpmbuild --rebuild ibmsp6a-x.xx-y.src.rpm
Followed by:
cd /usr/src/package-dir/RPMS/architecture
rpm -ivh ibmsp6a-x.xx-y.architecture.rpm
Where:
package-dir is the distribution-specific name of the RPM build directory
(usually "redhat" or "packages").
architecture is the architecture of the kernel in use (i386™, i586, or x86_64).
For example, to install the rpm on an x86_64 SUSE LINUX, the commands are:
rpmbuild --rebuild ibmsp6a-x.xx-y.src.rpm
cd /usr/src/packages/RPMS/x86_64
rpm -ivh ibmsp6a-x.xx-y.x86_64.rpm
EM64T and AMD64 note: On x86_64 kernels, this RPM will build a 64-bit
shared object and a 32-bit compatibility shared object. Before installing the
RPM on an x86_64 kernel, make sure the 32-bit compatibility development
packages are installed.

186
Planning and Installing the IBM Eserver X3 Architecture Servers
IPMI ASR service
This section describes how to install the ASR (ibmipmiasr) RPM.
Before installing, make sure your server has both the IPMI device driver and the
IBM Mapping Layer Software installed.
The system that the source rpm file is to be run on must have Linux
development/build capability.
If this is an upgrade to an existing software package, remove the old version first
with the command:
rpm -e ibmipmiasr
Depending upon your system's configuration, you may see messages about
missing files; these may be ignored.
To install the source rpm, execute the following command:
rpm -ivh ibmipmiasr-x.xx-y.i386.rpm
Once the installation is complete, check the log file /var/log/message. A
successful installation will write the following message to the log:
IBM IPMI ASR application loaded
To uninstall the binary rpm, execute the following command:
rpm -e ibmipmiasr
RPM will unload the ASR application, and remove all ibmipmiasr-related files
from your system.
5.3.7 Ports used by the BMC
The BMC uses several TCP/UDP ports for communication, as shown in
Table 5-4. If the communication with the BMC passes firewalls, it is important to
know which ports you have to enable on the firewalls to communicate properly.
Table 5-4 TCP/IP ports used by the BMC
Port number Description
623 IPMI communications to SMBridge and Director
664 IPMI communications (secondary)
161 SNMP get/set commands
162 SNMP traps and PET alerts to Director

Chapter 5. Management
187
5.4 Integrating the service processors with IBM Director
If you have the IBM Director Server installed in your environment, you will most
likely want to manage the service processors (BMC and RSA II) with IBM
Director. In this section, we describe how to configure the user ID, passwords,
and alert-forwarding settings.
5.4.1 Adding the service processor to IBM Director
If you have the service processor device driver and the IBM Director Agent
installed on the server, then IBM Director can automatically discover the service
processor as a Physical Platform. Consequently, you do not need to manually
add it as described in these steps.
To manually add the BMC or RSA to Director as a managed object, do the
following:
1.From the Director console, right-click the middle pane in a blank area.
2.Select New→ Physical Platform.
3.Enter the appropriate details for the service processor you want to add and
select OK, as shown in Figure 5-15.
Figure 5-15 Add physical platforms window
Tip: If you have an RSA II also installed in your server, then you do not need
to enable the BMC network interface by supplying a static IP address. You can
perform all management functions from the RSA II Web interface or IBM
Director.
Tip: If you server has both an RSA and a BMC installed, only the RSA will be
added to Director. You cannot add both service processors in the same server
to Director.

188
Planning and Installing the IBM Eserver X3 Architecture Servers
4.Once the service processor is discovered, it will appear as a Physical
Platform object on the Director console.
5.IBM Director attempts to access the BMC using the default
USERID/PASSW0RD combination.
If you have deleted or changed the default USERID/PASSW0RD
combination, then a small padlock icon will appear next to the device.
Right-click the device and click Request Access, and enter a valid user ID
and password.
5.4.2 Adding users
In Director 5.10, the recommended way to add users to the service processors is
to use the Server Configuration Manager. This is a task new to Director 5.10.
With it, you create a profile and then apply it to multiple systems at once. It does
have some limitations, however:
You can only add new users. You cannot replace existing users, including the
default USERID/PASSW0RD combination.
Since the RSA II supports 12 users at most and the BMC supports four users
at most (one of which is reserved), then this means you can only add 10 and
two, respectively.
You cannot delete users. This function, however, is planned to be added in a
future release of IBM Director.
To add a user to one or more service processors, do the following:
1.Double-click Server Configuration Manager.
2.Click Create a new profile, as shown in Figure 5-16 on page 189. You can
also edit an existing profile or import a profile from an XML file (which you
probably based on an XML file you previously exported).

Chapter 5. Management
189
Figure 5-16 Server Configuration Manager: Quick Start menu

190
Planning and Installing the IBM Eserver X3 Architecture Servers
3.Enter a name for the profile (for example, spadmin userid) and click OK.
You can select any of three components, as shown in Figure 5-17:
– IP Address Pool: Only applies to BladeCenter Management Modules
– Service Processor Logon Profile: To add a new service processor user ID
– Service Processor Network Configuration: To configure DHCP and SNMP
properties
Figure 5-17 Select components window
We select only the Service Processor Login Profile. Click OK to continue.
4.Enter a suitable user name, password, and authority level, as shown in
Figure 5-18 on page 191. If you wish to replace the default
USERID/PASSW0RD combination, check the check box Replace default
user ID and password.

Chapter 5. Management
191
Figure 5-18 Server Configuration Manager: adding a login profile
5.Select File → Save to save the profile, and then close the editor.

192
Planning and Installing the IBM Eserver X3 Architecture Servers
6.A profile will now appear in the tasks pane in the Director console (see
Figure 5-19). You can now drag and drop that profile to one or more service
processors (Physical Platforms) or groups to either execute the task now or
schedule it to run later.
7.You will see the status of each system as it is processed.
Figure 5-19 Results of applying the task against a service processor
8.Once complete, you can view the detailed results by double-clicking any of
the systems listed in the bottom half of the window. To see the status of an
individual system, double-click it in the lower pane of the Execution History
window.
If successful you will see:
-VENDOR- spconfig: Creating a new account for user spadmin
-VENDOR- spconfig: Setting the user's password
-VENDOR- spconfig: Setting the user's authority level
Note that if you already have the maximum number of users already defined,
then you will get an error:
-VENDOR- spconfig: User does not already exist and maximum number of
users has already been reached
Note: Server Configuration Manager adds users to the highest numbered
available slot first. For example, with the RSA II, the first user it adds will be
user ID 12 if that slot is available.

Chapter 5. Management
193
For the RSA II (with a limit of 12 users, including USERID), you will need to go to
the Web interface to delete an existing users. For the BMC (with a limit of three
users, including USERID), you will need to delete the user using the MPCLI
(management processor command-line interface). You can also use the MPCLI
with the RSA II as well.
Using the MPCLI, you can view the user IDs already defined using the
getdialinentry command and you can delete existing user IDs using the
setdialinentry command (see Figure 5-20 for an example).
Figure 5-20 MPCLI commands to read and delete a service processor user ID
Configuring alert forwarding
To configure the service processors to send all alerts to IBM Director 5.10, use
the following MPCLI commands (where, in our example, 9.42.171.229 is the
address of our Director 5.10 server):
setalertentry -index 1 -type director.all
setalertentry -index 1 -ipaddress 9.42.171.229
You can use the following command to verify the setting
getalertentry -index 1
Tip: To delete a user ID, set the user ID to a blank with the parameter -id ""
logged on ip=9.42.171.192>getdialinentry -index 11
Dialin Configuration:
Entry Number: 11
Id: blah
Last login: Tue Nov 30 00:00:00 EST 1999
Dialback enabled: false
Dialback number: null
Authority: supervisor
logged on ip=9.42.171.192>setdialinentry -index 11 -id ""
true
logged on ip=9.42.171.192>getdialinentry -index 11
Dialin Configuration:
Entry Number: 11
Id:
Last login: Tue Nov 30 00:00:00 EST 1999
Dialback enabled: false
Dialback number: null
Authority: supervisor

194
Planning and Installing the IBM Eserver X3 Architecture Servers
For the RSA II adapter, you can also use the Web interface to configure alert
forwarding by setting it to type “Director Comprehensive”.
5.5 OSA SMBridge utility
The OSA System Management Bridge (SMBridge) is a utility that lets you
perform certain remote management functions on servers with a BMC service
processor, such as the X3 Architecture servers. It allows the administration of
servers using IPMI1.5 protocol and the Serial Over LAN (SOL) protocol via either
the server’s Ethernet or serial interfaces. The primary function of SMBridge is to
provide remote control of the text-mode console via Ethernet.
There are two ways to use the SMBridge utility: as a telnet server and as a direct
command-line interface to the BMC. These are shown in Figure 5-21.
Figure 5-21 The two modes of the SMBridge utility
Telnet server connection
Used as a telnet server, SMBridge is started as a background service or
daemon on a system on your network (typically not the server with the BMC).
Client running
telnet client
Remote server with
BMC service
processor
Windows/Linux
server running
SMBridge telnet
server
In-band
telnet
session
(port 623)
Out-of-band
IP session
to BMC with
connect

command
Client running
smbridge.exe CLI
Remote server
with BMC service
processor
Out-of-band
IP or serial
session to
BMC
Telnet server mode CLI mode

Chapter 5. Management
195
You connect to the telnet server, then from there, connect to the BMC via the
server Ethernet port.
SMBridge uses the Serial Over LAN protocol to let the administrator remotely
control text-mode tasks, such as POST messages, BIOS setup, and
text-mode tasks with operating systems. Tasks you can perform are:
– Establish a text-mode console session with the remote server.
– Power on, power off (immediate and graceful), or reboot a server.
– Turn on/off the blinking system identifier.
– Display the current power status.
– Display the event log.
Any standard telnet client application, such as HyperTerminal on Microsoft
Windows or telnet on Linux, can be used to access the server’s features.
The SOL protocol coupled with the remote system’s BIOS console redirection
allows administrators to view and change the BIOS settings over LAN. Linux
serial console and Microsoft’s Emergency Messaging Service (EMS)/Special
Administration Console (SAC) interfaces can also be accessed over LAN
using SOL.
This is discussed further in 5.5.3, “Connecting via the telnet server” on
page 201.
Command-line interface
Used in this way, SMBridge lets an administrator perform the following tasks
on a remote BMC service processor to:
– Power on, power off (immediate and graceful), or reboot a server.
– Turn on/off the blinking system identifier.
– Display the current power status.
– Display or clear the event log.
The CLI lets you do all but the remote console feature offered by the telnet
server.
This is discussed further in 5.5.7, “Connecting via the command-line
interface” on page 217.
SMBridge can be downloaded from:
http://www.ibm.com/pc/support/site.wss/MIGR-57729.html

196
Planning and Installing the IBM Eserver X3 Architecture Servers
The current version supports the following operating systems:
Red Hat Linux 7.2
Red Hat Linux 8.0
Red Hat Linux 9.0
Red Hat Enterprise Linux 3.0
Microsoft Windows XP
Microsoft Windows 2000 Professional
Microsoft Windows 2000 Server
Microsoft Windows Server 2003
The OSA System Management Bridge User’s Guide is available at:
http://www.ibm.com/pc/support/site.wss/MIGR-57816.html
5.5.1 Configuring BIOS
Before SMBridge can be used to manage a remote server via SOL, the BMC and
BIOS of the remote server must have the following settings configured:
1.Enter the BIOS Setup by pressing F1 when prompted during boot.
2.If you have not done so already, configure the static IP address, subnet mask,
and gateway of the BMC, as described in 5.3.3, “Configuring the BMC in
BIOS” on page 179.
3.From the main menu, select Devices and I/O Ports. Set the following:
– Set field Serial Port A to Auto-configure.
– Set field Serial Port B to Auto-configure.
4.Select Remote Console Redirection. Set the following:
– Remote Console Active to Active
– Remote Console Text Emulation to VT100/VT220
– Remote Console Keyboard Emulation to VT100/VT220
– Remote Console Active After Boot to Enabled
– Remote Console Flow Control to Hardware
The result is shown in Figure 5-22 on page 197.
Note: This procedure disables PXE boot on Gigabit port 1 on the server. If you
plan to use PXE, you will need to connect Gigabit port 2 to your network and
ensure that your remote install procedure is configured to use that port.

Chapter 5. Management
197
Figure 5-22 Remote Console Redirection settings to enable SOL
5.Press Esc twice to return to the main menu, then select Start Options. Set
the following:
– Planar Ethernet 1 PXE to Disabled
– Planar Ethernet 2 PXE to Enabled
– Planar Ethernet PXE/DHCP to Planar Ethernet 2
– Run PXE only on Selected Planar NIC to Enabled
Note that you will most likely only have some of these options on your server.
6.Press Esc to return to the main menu, select Advanced Options, and then
select Baseboard Management Controller (BMC) Settings. Set the
following:
– System-BMC Serial Port Sharing to Enabled
– BMC Serial Port Access Mode to Dedicated
7.Save the BIOS settings and reboot the server.
5.5.2 Installation
This section describes how to install the SMBridge utility on both Windows and
Linux platforms. The latest version of the utility is available from:
http://www.ibm.com/pc/support/site.wss/MIGR-57729.html
********************************************************
* Remote Console Redirection *
********************************************************
* Remote Console Active [ Enabled ] *
* Remote Console COM Port [ COM 1 ] *
* Remote Console Baud Rate [ 19200 ] *
* Remote Console Data Bits [ 8 ] *
* Remote Console Parity [ None ] *
* Remote Console Stop Bits [ 1 ] *
* Remote Console Text Emulation [ VT100/VT220 ] *
* Remote Console Keyboard Emulation [ VT100/VT220 ] *
* Remote Console Active After Boot [ Enabled ] *
* Remote Console Flow Control [ Hardware ] *
********************************************************

198
Planning and Installing the IBM Eserver X3 Architecture Servers
Microsoft Windows
The procedure both installs the CLI, and it installs and enables the telnet server.
1.Run Setup, agree to the license, and specify an installation directory.
2.You will now be prompted to enter an IP address and TCP/IP port number, as
shown in Figure 5-23.
Figure 5-23 IP address and port number window
These values are as follows:
– IP specifies the server IP address that SMBridge will bind to.
Since a server may have multiple valid IP addresses, SMBridge allows
you to restrict access to it via a single IP address. Specify INADDR_ANY
as the IP address if any of the multiple IP addresses can be bound to
SMBridge. Specify 127.0.0.1 or localhost as the IP address if SMBridge
Tip: This installation is normally run on the server you plan to have act as the
telnet server, as shown in Figure 5-21 on page 194. If you plan to use the CLI,
you do not actually need to install the tool, since the only files you need to run
the CLI are smbridge.exe and smbridge.cfg.
Consequently, if you plan to use the CLI on other systems, you can either:
Copy the files smbridge.exe and smbridge.cfg from the telnet server.
Install SMBridge and then disable the service.

Chapter 5. Management
199
should only accept local connections. Specify a specific IP address if only
this IP address should be bound to SMBridge.
– Port specifies the server port number that SMBridge will listen on.
3.Next, you are asked to specify timeout values for telnet sessions (in minutes)
and the power-off command (in seconds) (Figure 5-24).
Figure 5-24 Session and power off timeout settings window
These values are as follows:
– Session specifies the number of minutes without any keyboard activity
before an established telnet session is ended.
– PowerOff specifies the number of seconds to wait for an IPMI power off
command (graceful or forced) to complete. If time has exceeded the
timeout value and the power status is still on, an error code will be
returned to indicate that the power off command may have failed.
4.Click Next to confirm your choices, and then begin the installation. Once the
installation is complete, click Finish to end the installer.
Note: These two values will be recorded in the smbridge.cfg file for
automatic startup of SMBridge as a service daemon.

200
Planning and Installing the IBM Eserver X3 Architecture Servers
The OSA SMBridge service is started automatically and is configured to start
every time the server starts. You can change this by selecting Control Panel →
Administrative Tools → Services.
Installation on a Linux platform
To install SMBridge on Linux, follow these steps:
1.Log in as root.
2.If the SMBridge RPM file is on CD, then insert the CD into the drive and enter
the following commands to mount the drive and change to the root directory
of the CD:
mount /mnt/cdrom
cd /mnt/cdrom
3.Run the installation with the following command (substitute the file name of
the rpm file you have if it is different):
rpm -i osasmbridge-1.0.3-1.i386.rpm
When the installation process has finished successfully, files are copied to the
following directories:
/etc/init.d/smbridge
/etc/smbridge.cfg
/usr/bin/smbridge
/var/log/smbridge
/var/log/smbridge/LICENSE
Additionally, the symbolic link /usr/sbin/smbridge is created.
The text file /etc/smbridge.cfg contains a number of SMBridge runtime
parameters that you should review and modify if necessary.
To start the daemon, navigate to directory /etc/int.d and use the following
command to start or stop the OSA SMBridge daemon service:
smbridge start
smbridge stop
You can also start the daemon using the command:
smbridge -d -c config-file
Where config-file is the name of the file containing the telnet server configuration.
By default, it is /etc/smbridge.cfg. See Appendix C, “Configure OSA SMBridge”,
in the OSA System Management Bridge User’s Guide for more details about this
file, which can be found at:
http://www.pc.ibm.com/support?page=MIGR-57816

Chapter 5. Management
201
5.5.3 Connecting via the telnet server
As shown in Figure 5-21 on page 194, you can use SMBridge as a telnet
interface (or a “bridge”) to the BMC. You connect to a telnet server (where you
installed SMBridge), and from there you connect to the BMC using a Serial Over
LAN (SOL) connection.
Using a SOL connection, you can perform over a LAN connection all the tasks
that you would normally only be able to do while connected directly to the
server’s serial port:
Change the BIOS settings.
Linux serial console.
Emergency Messaging Service (EMS) from Microsoft.
Special Administration Console (SAC) from Microsoft.
Information about Microsoft EMS is available from:
http://www.microsoft.com/resources/documentation/WindowsServ/2003/stand
ard/proddocs/en-us/EMS_topnode.asp
The SAC commands you can perform are described here:
http://www.microsoft.com/resources/documentation/WindowsServ/2003/stand
ard/proddocs/en-us/EMS_SAC_commands.asp
With BIOS console redirection to serial port enabled on the remote server,
applications that use the BIOS to read and write to the system console will have
their I/O redirected to the serial port. With SOL, the BMC firmware then reads the
data written to the serial port and transmits it to the SMBridge as LAN packets.
SMBridge then forwards the data to the telnet client as TCP/IP packets.
One SMBridge session supports one SOL session with one BMC at a time.
Telnet clients
To access the SMBridge telnet server, connect using port 623 (or the port you
selected during installation if you changed it from this default). For example:
telnet smbridge-server 623
Note: By default, the SMBridge telnet server listens on port 623.

202
Planning and Installing the IBM Eserver X3 Architecture Servers
Telnet clients that support VT100 terminal emulation can be used to access the
BMC via SMBridge, including the following:
The telnet command-line utility in Windows
Additional information:
– Our testing showed that on a Windows XP system, pressing F1 on the
keyboard correctly sent us to the remote server (for example, to enter the
BIOS Setup). If your telnet client does not work this way (such as the
telnet command in Windows Server 2000), you can simulate the F1 and
F2 keys in two ways. To simulate F1, either press:
• Esc, then Numeric+1 (that is, the 1 key on the numeric keypad)
• Esc, then Shift+O, then Shift+p
To simulate F2, either press:
• Esc, then Numeric+2 (that is, the 2 key on the numeric keypad)
• Esc, then Shift+O, then Shift+q
HyperTerminal in Windows
Additional information:
– HyperTerminal supplied with Windows Server 2003, Enterprise Edition
may not work correctly. You should upgrade to the Private version.
– HyperTerminal supplied with Windows 2000 Server displays random
characters and loses some text. We recommend that you do not use this
client.
– When creating a new connection, select TCP/IP (WinSock) in the
Connect window using the drop-down menu. Enter the IP address of the
telnet server and specify port 623 (or the port you specified when you
installed SMBridge).
– Turn off automatic line wrapping by selecting File → Properties →
Setting → ASCII → Setup and unchecking the Wrap lines that exceed
terminal width check box.
– Configure the connection to emulate a VT100 terminal in the Properties
window.
The telnet command as part of csh or ksh in Linux.
The F1 and F2 keys may not work correctly, especially outside of an
X-Windows environment. You may be able to reconfigure X-Windows to
generate VT100 keystrokes. With KDE, for example, use Settings to
reconfigure the keyboard.
Tip: We recommend that you use this telnet client when running Windows.
It is a standard implementation of telnet and also supports colors.

Chapter 5. Management
203
Connecting
In this section, we describe the process by referring to servers in our lab. We
have installed the SMBridge telnet server on the system at address
9.42.171.121, and the remote server has a BMC configured to use address
9.24.171.237. Our BMC uses the default USERID/PASSW0RD authentication.
To connect to the BMC, do the following (using our example addresses):
1.Issue the following to connect to the telnet server on port 623:
telnet 9.42.171.121 623
2.You will be given the prompt shown in Figure 5-25.
Figure 5-25 Username prompt
3.Enter an administrator user ID and password that is valid for the telnet server
(for example, Administrator or root).
4.You will then see the welcome message shown in Figure 5-26.
Figure 5-26 Welcome message
Username:
Note: If the SMBridge telnet server is running on the local system, you will
not see this prompt, as SMBridge uses the authority of the current user
logged on.
Username:Administrator
Password:
Administrator login successful.
OSA System Management Bridge (SMBridge), Version 1.0.3.1
Copyright (c) 2004 - OSA Technologies, an Avocent Company. All
Rights Reserved.
SMBridge>

204
Planning and Installing the IBM Eserver X3 Architecture Servers
5.Connect to the BMC on the server you wish to manage using the connect
command, as shown in Figure 5-27.
Figure 5-27 Connect to the BMC
The user ID and password here are ones that have previously been
configured as users able to log into the BMC.
If the command was successful, you will be returned to the SMBridge
command prompt.
6.You can now issue commands against the remote BMC, as described below.
7.To exit, enter the exit command.
Available commands
These commands listed in Table 5-5 are a super-set of those of the
command-line interface. To get detailed help about a command, issue the help
command. For example:
help power
Table 5-5 SMBridge telnet subcommands
SMBridge>connect -ip 9.42.171.237 -u USERID -p PASSW0RD
SMBridge>
Tip: Most of the commands available to the telnet interface are the same as
those used in the CLI. The additional telnet commands are console, sol, and
reboot.
Subcommand Description and syntax
console Starts a Serial Over LAN (SOL) session with the BMC, displaying the
text that has been redirected from the console to the serial port. There
are no parameters.
When you enter the console command, you will see:
Activating remote console now.
Remote console is now active and ready for user input.
To return to the telnet session, press the tilde key followed by the
period key, as in:
~.

Chapter 5. Management
205
sol Used to enable or disable Serial Over LAN and to configure serial
parameters to match the Console Redirection parameters of the
remote server’s BIOS. The options are:
sol enable
sol disable
sol config [-baud baud_rate] [-priv privilege_level] [-retry
count retry_count] [-retry interval retry_interval]
reboot Performs the equivalent of a power off (graceful shutdown), power on,
then starts the remote console. The options are:
reboot
reboot -force
Note that the x236, x336, and x346 do not support the graceful
shutdown option. The -force parameter is required on these servers.
sysinfo Displays general system information related to the server and BMC.
The options are:
sysinfo fru
sysinfo id
id is the default if no parameter is specified.
identify Controls the blue identification LED on the front panel of the server.
The options are:
identify on [-t <seconds>]
identify off
on is the default if no parameter is specified.
power Controls the power options of the server. The options are:
power status
power on
power cycle
power reset
power off [-force]
status is the default if no parameter is specified.
Note that the x236, x336, and x346 do not support the graceful
shutdown option. The -force parameter is required on the server.
Subcommand Description and syntax

206
Planning and Installing the IBM Eserver X3 Architecture Servers
5.5.4 Configuring Windows Server 2003 to support SOL
When you connect to the BMC using the SMBridge telnet server, you can
remotely control the text console. With SOL, this also includes operating systems
such as Windows Server 2003 and Linux.
Windows Server 2003 has two components that work with SMBridge and the
BMC to provide out-of-band access to the operating system:
Microsoft Emergency Messaging Service (EMS)
Microsoft Special Administration Console (SAC)
Information about Microsoft EMS is available from:
http://www.microsoft.com/resources/documentation/WindowsServ/2003/stand
ard/proddocs/en-us/EMS_topnode.asp
The SAC commands you can perform are summarized in Table 5-6.
Table 5-6 Windows Server 2003 Special Administration Console (SAC) commands
sel Performs operations with the System Event Log (SEL). The options
are:
sel status
sel get
set get -last <n>
sel get -begin <index1> -end <index2>
sel get -begin <index1> -max <count>
sel clear
sel set -time <YYYY/MM/DD hh:mm:ss>
status is the default if no parameter is specified.
help Displays general help about all commands or help about a specific
command.
Subcommand Description and syntax
Command Description
ch Lists all channels.
cmd Creates a command-prompt channel. You will be asked to log
on.
crashdump Manually generates a Stop error message and forces a memory
dump file to be created.
d Dumps the current kernel log.

Chapter 5. Management
207
To exit SOL and return to the SMBridge prompt, press the tilde key and the
period key (that is, ~.).
f Toggles the information output of the t-list command, which
shows processes only, or shows processes and threads.
i Lists the TCP/IP details of all network interfaces, and lets you
configure the IP address, subnet mask, and gateway of a given
network interface. To change the parameters, specify them as:
network# IPaddress subnet gateway
id Displays identification information about the server.
k
pid
Ends the given process. pid is the process identification number
you specify.
L
pid
Lowers the priority of a process (and any associated child
processes) to the lowest possible level.
lock Restricts access to Emergency Management Services
command-prompt channels. You must provide valid logon
credentials to unlock a channel.
m
pid mb
-allow Limits the memory usage of a process (and any associated
child processes) to a specified number of megabytes. mb is the
number of megabytes you specify.
p Causes t-list command output to pause after displaying one
full screen of information.
r
pid
Raises the priority of a process and any associated child
processes by one level.
restart Restarts the server.
s Displays or sets the system time. To set the time, use the
format:
mm/dd/yyyy hh:mm
shutdown Shuts down and powers off the server. Terminates the console
session and returns you to the SMBridge prompt.
t Lists the processes and threads that are currently running.
? or help Lists the available commands.
Command Description

208
Planning and Installing the IBM Eserver X3 Architecture Servers
For more information, see:
http://www.microsoft.com/resources/documentation/WindowsServ/2003/stand
ard/proddocs/en-us/EMS_SAC_commands.asp
To enable EMS on a Windows Server 2003, do the following.
1.Log in to Windows as an administrator.
2.Launch a command prompt and enter the command bootcfg. This command
should produce the output shown in Figure 5-28.
Figure 5-28 Output from the bootcfg command
3.Examine the output. If there is more than one boot entry, then you will need to
determine the default entry by looking at the default line under Boot Loader
Settings and determining whether Boot Entry has a matching Path value. In
our case, there is only one boot entry, 1.
4.Issue the following command, substituting your boot entry number in the /id
parameter if it is not 1, as in our example:
bootcfg /ems on /port com1 /baud 19200 /id 1
Figure 5-29 shows the ouput of this command.
Figure 5-29 Changing the boot configuration
C:\>bootcfg /ems on /port com1 /baud 19200 /id 1
SUCCESS: Changed the redirection port in boot loader section.
SUCCESS: Changed the redirection baudrate in boot loader section.
SUCCESS: Changed the OS entry switches for line "1" in the BOOT.INI
file.
C:\>bootcfg
Boot Loader Settings
--------------------
timeout:30
default:multi(0)disk(0)rdisk(0)partition(1)\WINDOWS
Boot Entries
------------
Boot entry ID: 1
OS Friendly Name: Windows Server 2003, Enterprise
Path: multi(0)disk(0)rdisk(0)partition(1)\WINDOWS
OS Load Options: /fastdetect

Chapter 5. Management
209
5.Reissue the bootcfg command to see the result. The changes in our example
are highlighted, as shown in Figure 5-30 on page 209.
Figure 5-30 The bootcfg command after enabling EMS
6.Reboot the server to have the changes take effect.
Note: To turn EMS off again, issue the following command:
bootcfg /ems off /id 1
Where 1 is the boot entry you have modified in the above steps. Reboot to
bring the changes online.
C:\>bootcfg /ems on /port com1 /baud 19200 /id 1
SUCCESS: Changed the redirection port in boot loader section.
SUCCESS: Changed the redirection baudrate in boot loader section.
SUCCESS: Changed the OS entry switches for line "1" in the BOOT.INI
file.
C:\>bootcfg
Boot Loader Settings
--------------------
timeout: 30
default: multi(0)disk(0)rdisk(0)partition(1)\WINDOWS
redirect: COM1
redirectbaudrate:19200
Boot Entries
------------
Boot entry ID: 1
OS Friendly Name: Windows Server 2003, Enterprise
Path:multi(0)disk(0)rdisk(0)partition(1)\WINDOWS

210
Planning and Installing the IBM Eserver X3 Architecture Servers
Once you have rebooted and engaged the SMBridge console (see the console
command in Table 5-5 on page 204), you will see the EMS console, as shown in
Figure 5-31.
Figure 5-31 Microsoft Emergency Messaging Service console
You can now issue the various SAC commands described in Table 5-6 on
page 206. For example, to start a command prompt, the commands are as
follows (see Figure 5-32 on page 211).
<?xml version="1.0"?>
<machine-info>
Computer is booting, SAC started and initialized.
<processor-architecture>x86</processor-architecture>
<os-version>5.2</os-version>
<os-build-number>3790</os-build-number>
SAC>
Tip: After you start the SMBridge console, if you only get a blank screen,
press Enter a few times to get the SAC> prompt.

Chapter 5. Management
211
Figure 5-32 SAC commands to launch a command prompt channel
To close the command prompt channel, enter exit. To leave the channel open
and return to the SAC prompt, press Esc+Tab+0 (the number zero key) (three
keys in sequence). To leave the remote console and return to SMBridge, press
tilde+period (that is, ~.)
SAC>cmd
The Command Prompt session was successfully launched.
SAC>
EVENT: A new channel has been created. Use "ch -?" for channel help.
Channel: Cmd0002
SAC>ch
Channel List
(Use "ch -?" for information About using channels)
# Status Channel Name
0 (AV) SAC
1 (AV) Cmd0002
SAC>ch -si 1
Name: Cmd0002
Description: Command Prompt
Type: <Esc><tab>0 to return to the SAC channel.
Use any other key to view this channel.
Please enter login credentials.
Username:
Domain:
Password:
Attempting to authenticate...
Microsoft Windows [Version 5.2.3790]
(C) Copyright 1985-2003
C:\WINDOWS\system32>

212
Planning and Installing the IBM Eserver X3 Architecture Servers
5.5.5 Configuring Red Hat Linux to support SOL
You must configure Red Hat Linux to expose the Linux initialization (booting)
process. This enables users to log in to the Linux console through an SOL
session and directs output to the serial console. The following instructions are for
Red Hat Enterprise Linux ES 2.1 or 3.0 to enable SOL:
1.Log in as root.
2.Modify the /etc/inittab file by adding the following line to the end of the # Run
gettys in standard runlevels section to enable users to log in at the SOL
console:
7:2345:respawn:/sbin/agetty -h ttyS1 19200 vt102
3.Modify the /etc/securetty file by adding the following line to enable users to
log in as root at the SOL console:
ttyS1
For LILO users
(GRUB users, please go to step 1 on page 214):
1.Modify the /etc/lilo.conf file:
a.Add -Monitor to the first default line.
b.Comment out the map line.
c.Comment out the message line.
d.In the first Image section, append -Monitor to the label line, and append
the following line:
append="console=ttyS1,19200n8 console=tty1"
e.Add the following lines between the two Image sections
# This will allow you to Interact with the OS boot via SOL
image=/boot/vmlinuz-2.4.9-e.12smp
label=linux-Interact
initrd=/boot/initrd-2.4.9-e.12smp.img
read-only
root=/dev/hda6
append="console=tty1 console=ttyS1,19200n8"
The result is shown in Figure 5-33 on page 213. The changes are highlighted.

Chapter 5. Management
213
Figure 5-33 Changes to the lilo.conf file
2.Enter lilo to store and activate the new LILO configuration.
3.Restart Linux.
When the operating system starts to boot, you will now see a LILO boot: prompt
instead of the usual GUI interface. Pressing the Tab key while at this prompt will
display the boot options. To load the operating system in interactive mode, you
would enter:
linux-Interact
prompt
timeout=50
default=linux-Monitor
boot=/dev/hda
#map=/boot/map
install=/boot/boot.b
#message=/boot/message
linear
# This will allow you to only Monitor the OS boot via SOL
image=/boot/vmlinuz-2.4.9-e.12smp
label=linux-Monitor
initrd=/boot/initrd-2.4.9-e.12smp.img
read-only
root=/dev/hda6
append="console=ttyS1,19200n8 console=tty1"
# This will allow you to Interact with the OS boot via SOL
image=/boot/vmlinuz-2.4.9-e.12smp
label=linux-Interact
initrd=/boot/initrd-2.4.9-e.12smp.img
read-only
root=/dev/hda6
append="console=tty1 console=ttyS1,19200n8"
image=/boot/vmlinuz-2.4.9-e.12
label=linux-up
initrd=/boot/initrd-2.4.9-e.12.img

214
Planning and Installing the IBM Eserver X3 Architecture Servers
For GRUB users:
1.Modify the /boot/grub/grub.conf file as follows:
a.Comment out the splashimage line.
b.Add the following comment before the first title line:
# This will allow you to only Monitor the OS boot via SOL
c.Append SOL Monitor to the first title line.
d.Append the following text to the end of the kernel line of the first title
section:
console=ttyS1,19200 console=tty1
e.Add the following lines between the two title sections:
# This will allow you to Interact with the OS boot via SOL
title Red Hat Linux (2.4.9-e.12smp) SOL Interactive
root (hd0,0)
kernel /vmlinuz-2.4.9-e.12smp ro root=/dev/hda6 console=tty1
console=ttyS1,19200
initrd /initrd-2.4.9-e.12smp.img
The result is shown in Figure 5-34 on page 215. The changes are highlighted.
2.Restart Linux.

Chapter 5. Management
215
Figure 5-34 Changes to the grub.conf file
5.5.6 Configuring SUSE LINUX to support SOL
You must configure SUSE LINUX to expose the Linux initialization (booting)
process. This enables users to log in to the Linux console through an SOL
session and directs output to the serial console. The following instructions are for
SUSE LINUX Enterprise Server 8.0 to enable SOL:
1.Log in as root.
2.Modify the /etc/inittab file by adding the following line to the end of the
#getty-programs for the normal runlevels section to enable users to log in
at the SOL console:
7:2345:respawn:/sbin/agetty -h ttyS1 19200 vt102
#grub.conf generated by anaconda
#
# Note that you do not have to rerun grub after making changes to this file
# NOTICE: You have a /boot partition. This means that
# all kernel and initrd paths are relative to /boot/, eg.
# root (hd0,0)
# kernel /vmlinuz-version ro root=/dev/hda6
# initrd /initrd-version.img
#boot=/dev/hda
default=0
timeout=10
#splashimage=(hd0,0)/grub/splash.xpm.gz
# This will allow you to only Monitor the OS boot via SOL
title Red Hat Enterprise Linux ES (2.4.9-e.12smp) SOL Monitor
root (hd0,0)
kernel /vmlinuz-2.4.9-e.12smp ro root=/dev/hda6 console=ttyS1,19200 console=tty1
initrd /initrd-2.4.9-e.12smp.img
# This will allow you to Interact with the OS boot via SOL
title Red Hat Linux (2.4.9-e.12smp) SOL Interactive
root (hd0,0)
kernel /vmlinuz-2.4.9-e.12smp ro root=/dev/hda6 console=tty1 console=ttyS1,19200
initrd /initrd-2.4.9-e.12smp.img
title Red Hat Enterprise Linux ES-up (2.4.9-e.12)
root (hd0,0)

216
Planning and Installing the IBM Eserver X3 Architecture Servers
3.Modify the /etc/securetty file by adding the following line after the tty6 line to
enable users to log in as root at the SOL console:
ttyS1
4.Modify the /boot/grub/menu.lst file as follows:
a.Comment out the gfxmenu line.
b.Add the following comment line before the first title line:
# This will allow you to only Monitor the OS boot via SOL
c.Append SOL Monitor to the first title line.
d.Append the following text to the kernel line of the first title section:
console=ttyS1,19200 console=tty1
e.Add the following lines between the first two title sections:
# This will allow you to Interact with the OS boot via SOL
title linux SOL Interactive
kernel (hd0,1)/boot/vmlinuz root=/dev/hda2 acpi=oldboot vga=791
console=tty1 console=ttyS1,19200
initrd (hd0,1)/boot/initrd
The result is shown in Figure 5-35 on page 217. The changes are highlighted.
5.Restart Linux.

Chapter 5. Management
217
Figure 5-35 Changes to the menu.lst file
5.5.7 Connecting via the command-line interface
OSA SMBridge also supplies a command-line tool, smbridge.exe, which lets you
perform a subset of the functions that you can perform using the telnet server.
Specifically, the tasks that are missing are the ability to remotely control the text
console of the server via SOL. The CLI does, however, let you connect to the
server via a serial connection.
In CLI mode, SMBridge supports out-of-band access through the LAN or serial
port to one server at a time. However, multiple IPMI sessions can run
simultaneously on the same remote server. LAN connections are via Ethernet
and serial connections are typically via a null modem.
#gfxmenu (hd0,1)/boot/message
color white/blue black/light-gray
default 0
timeout 8
# This will allow you to only Monitor the OS boot via SOL
title linux SOL Monitor
# Note: The following "kernel" line is all one line, not two separate lines
# The text has wrapped in this example
kernel (hd0,1)/boot/vmlinuz root=/dev/hda2 acpi=oldboot vga=791
console=ttyS1,19200 console=tty1
initrd (hd0,1)/boot/initrd
# This will allow you to Interact with the OS boot via SOL
title linux SOL Interactive
# Note: The following "kernel" line is all one line, not two separate lines
# The text has wrapped in this example
kernel (hd0,1)/boot/vmlinuz root=/dev/hda2 acpi=oldboot vga=791 console=tty1
console=ttyS1,19200
initrd (hd0,1)/boot/initrd
title floppy
root
chainloader +1
title failsafe
kernel (hd0,1)/boot/vmlinuz.shipped root=/dev/hda2 ide=nodma apm=off vga=normal

218
Planning and Installing the IBM Eserver X3 Architecture Servers
To run SMBridge in CLI mode, simply open a command prompt/shell prompt at
the directory where SMBridge is installed and issue the smbridge command.
With Windows, SMBridge is installed by default in c:\Program Files\OSA.
With Linux, it is installed by default in /usr/sbin.
The syntax is as follows for Ethernet or Serial connectivity.
For Ethernet connections:
smbridge -ip address -u user -p password subcommand
Where
-ip address is the IP address or host name of the remote server.
-u user -p password is a valid service processor user ID and password
(default USERID/PASSW0RD).
For Serial connections:
smbridge -com serialport [-baud baudrate] [-flow flowcontrol] -u user
-p password subcommand
Where:
-com serialport specifies the serial port on remote server. In Windows
systems, it can be 1 for COM1, 2 for COM2, and so on. In Linux systems, it
can be ttyS0, ttyS1, and so on.
-baud baudrate specifies the baud rate you wish to communicate at, such as
9600 and 19200. It should match the one set in BIOS of the remote server (in
the Remote Console Redirection window). If not specified, it defaults to
19200.
-flow flowcontrol specifies the flow control. If not specified, it defaults to CTS
(hardware flow control). The options are:
– CTS: Hardware flow control
– XON: Software flow control
– NONE: No flow control
The valid subcommands and the syntax of those commands is listed in Table 5-7
on page 219. For more information about the syntax, issue the -help command.
For example:
smbridge -help power

Chapter 5. Management
219
Table 5-7 SMBridge CLI subcommands
Subcommand Description and syntax
sysinfo Displays general system information related to the server and BMC.
The options are:
sysinfo fru
sysinfo id
id is the default if no parameter is specified.
identify Controls the blue identification LED on the front panel of the server.
The options are:
identify on [-t <seconds>]
identify off
on is the default if no parameter is specified.
power Controls the power options of the server. The options are:
power status
power on
power cycle
power reset
power off [-force]
status is the default if no parameter is specified.
Note: The x236, x336, and x346 do not support the graceful
shutdown option. The -force parameter is required on these servers.
sel Performs operations with the System Event Log (SEL). The options
are:
sel status
sel get
set get -last <n>
sel get -begin <index1> -end <index2>
sel get -begin <index1> -max <count>
sel clear
sel set -time <YYYY/MM/DD hh:mm:ss>
status is the default if no parameter is specified.

220
Planning and Installing the IBM Eserver X3 Architecture Servers
5.6 Predictive Failure Analysis (PFA)
PFA is a mechanism developed by IBM, that periodically measures critical
components such as hard drives, power supplies, fans, processors, and memory
to reduce unscheduled system downtime.
If a predefined threshold of one of these components is exceeded, for example,
the number of single bit errors on a memory module, a PFA alert will be issued.
This enables you to replace the component before a failure actually occurs.
If a PFA alert is raised, the Error LED on the light path diagnostic panel will be lit
and the BMC logs an event that can be forwarded to IBM Director. The RSA II
can also be configured to forward alerts via e-mail or to an IBM Director server
(see also 5.2, “Remote Supervisor Adapter II SlimLine” on page 164).
5.7 IBM Dynamic System Analysis
IBM Dynamic System Analysis (DSA) is an information gathering tool, which
collects and analyzes system information in order to provide the appropriate data
to the IBM support in case of problems.
5.7.1 Introduction
DSA can be downloaded from the following Web site:
http://www.ibm.com/servers/eserver/xseries/systems_management/dsa.html
DSA collects information about the following aspects of a system:
System configuration
Installed applications and hot fixes
Device drivers and system services
Network interfaces and settings
Performance data and running process details
Hardware inventory, including PCI information
Vital product data and firmware information
SCSI device sense data
EXA chipset uncorrectable error register information
ServeRAID configuration

Chapter 5. Management
221
Application, system, security, ServeRAID, and service processor system
event logs
As the time of writing this redbook, the following operating systems are
supported by DSA: Microsoft Windows 2000 Server, Advanced Server and
Datacenter Server with Service Pack 3 or later, Microsoft Windows 2003
Standard, Enterprise, and Web Editions. DSA versions that support 64-bit
Windows editions and Linux will be available in the future.
Two versions of DSA are available:
DSA Portable Edition is meant to be run from portable media, such as a
memory key or CD-ROM. It runs from the command prompt, unpacks to and
runs from a temporary directory, but does not actually install on the system.
DSA Installable Edition provides a permanent installation of DSA onto a
system. You can run DSA Installable Edition from a command line, from the
Start menu, as an IBM Director task, or in conjunction with UpdateXpress.
DSA can be downloaded here:
http://www.ibm.com/servers/eserver/xseries/systems_management/dsa.html
Tip: The current and previous versions of DSA can be downloaded from:
http://www.ibm.com/pc/support/site.wss/document.do?lndocid=SERV-DSA

222
Planning and Installing the IBM Eserver X3 Architecture Servers
Figure 5-36 shows an example of the DSA output in HTML format.
Figure 5-36 DSA results Web page
5.7.2 Usage
DSA can be used in several ways, depending on the desired result. In the
following examples, we refer to the DSA Portable Edition, which is simply
launched by executing the downloaded program from a command-line prompt.
The Installable Edition is installed by default in C:\Program Files\IBM\DSA\. Once
installed, you can launch it in a variety of ways:
From the Start menu
From a command prompt
As an IBM Director task
In conjunction with UpdateXpress
The option switches listed for the DSA Portable Edition apply to the installable
edition as well.

Chapter 5. Management
223
When DSA is started without any additional command-line parameters, a
compressed XML file is created, which can be sent via e-mail to an IBM support
agent:
C:\temp>dsa101p.exe
To create a human readable output in HTML format instead of the XML file, use
the following parameters:
C:\temp>dsa101p.exe /s /a /x /v
Servers that have an Internet connection can create and upload the compressed
XML file via FTP to a secured server, so that is available for the IBM support
agent for analysis. This is done using the following command:
C:\temp>dsa101p.exe /t
It is also possible to parse a compressed XML created on an other machine file
to inspect the output using the following command:
C:\temp>dsa101p.exe /s /a /x /i C:\temp\ 88721RC_KK00011_20050704-081500
For further details about the command options, see the DSA readme.txt that
accompanies the downloadable files in the above URLs.
5.8 Partition management
Scalable partitions can be managed using the RSA Web GUI. This is described
in 3.3.2, “Scalable system setup” on page 110 and 3.3.3, “Partitioning” on
page 116.

224
Planning and Installing the IBM Eserver X3 Architecture Servers

© Copyright IBM Corp. 2006. All rights reserved.
225
AC alternating current
ACPI advanced control and power
interface
ALU arithmetic logic unit
AMD Advanced Micro Devices
APIC Advanced Programmable
Interrupt Controller
AS Australian Standards
ASCII American National Standard
Code for Information
Interchange
ASIC application-specific integrated
circuit
ASM Advanced System
Management
ASR automatic server restart
AWE Address Windowing
Extensions
BI business intelligence
BIOS basic input output system
BMC baseboard management
controller
CD compact disk
CD-ROM compact disk read only
memory
CE Conformité Européene
CIM Common Information Model
CLI command-line interface
COA Certificate of Authenticity
COM Component Object Model
CPU central processing unit
CRM Customer Relationships
Management
CSG Chip Select Group
Abbreviations and acronyms
CTP composite theoretical
performance
CTS clear to send
DB database
DBA database administrator
DHCP Dynamic Host Configuration
Protocol
DIMM dual inline memory module
DMI Desktop Management
Interface
DNS Domain Name System
DP dual processor
DPF Database Partitioning Feature
DRAM dynamic random access
memory
DSA digital signature algorithm
EB exabyte
ECC error checking and correcting
EIDE enhanced IDE
EMEA Europe, Middle East, Africa
EMS Emergency Messaging
Service
ERO Export Regulation Office
ERP enterprise resource planning
EXA Enterprise X-Architecture
FAMM full array memory mirroring
FQDN fully qualified domain names
GB gigabyte
GPR general purpose register
GRUB grand unified bootloader
GUI graphical user interface
HAL hardware abstraction layer
HAM hot-add memory

226
Planning and Installing the IBM Eserver X3 Architecture Servers
HBA host bus adapter
HDD hard disk drive
HID human interface device
HPMA high performance memory
array
HTML Hypertext Markup Language
HTTP hypertext transmission
protocol
I/O input/output
IBM International Business
Machines Corporation
IBS International Business
System
ICC International Competency
Center
ID identifier
IP Internet Protocol
IPMI Intelligent Platform
Management Interface
IPX internetwork packet exchange
ISO International Organization for
Standards
ISU Industry Solutions Utilities
IT information technology
ITSO International Technical
Support Organization
IXA Integrated xSeries Adapter
JDE JD Edwards
KB Kilobyte
KDE K Desktop Environment
KVM keyboard video mouse
LAN local area network
LDAP Lightweight Directory Access
Protocol
LED light emitting diode
LILO Linux loader
MAC media access control
MB megabyte
MIB management information
base
MIOC Memory and I/O Controller
MP multiprocessor
MPA Management Processor
Assistant
MTOPS millions composite theoretical
operations per second
MXE modular expansion enclosure
NIC network interface card
NMI non-maskable interrupt
NONRED not redundant, the
non-redundant
NUMA Non-Uniform Memory Access
OEM other equipment manufacturer
OLAP online analytical processing
OLTP online transaction processing
OS operating system
PAE Physical Address Extension
PC personal computer
PCI peripheral component
interconnect
PFA Predictive Failure Analysis
PKT packet
POST power on self test
PXE Pre-boot-execution
RAC Real Application Clusters
RAID redundant array of
independent disks
RAM random access memory
RBS redundant bit steering
RHEL Red Hat Enterprise Linux
ROLAP Relational Online Analytical
Processing
ROM read-only memory
RPM Red Hat Package Manager
RSA Remote Supervisor Adapter

Abbreviations and acronyms
227
SA solution assurance
SAC Special Administration
Console
SAN storage area network
SAPR Solution Assurance Product
Review
SAR Solution Assurance Review
SAS Serial Attached SCSI
SCM Supply Chain Management
SCON server consolidation
SCSI small computer system
interface
SDRAM static dynamic RAM
SEL System Event Log
SIMD Single Instruction Multiple
Data
SL SlimLine
SLES SUSE LINUX Enterprise
Server
SLIP Serial Line Internet Protocol
SMB server message block
SMBIOS system management BIOS
SMI Structure of Management
Information
SMP symmetric multiprocessing
SMS System Managed Space
SMTP simple mail transfer protocol
SNA systems network architecture
SNMP Simple Network Management
Protocol
SOL Serial over LAN
SPEC Standard Performance
Evaluation Corporation
SPORE ServerProven Opportunity
Request for Evaluation
SQL structured query language
SRAT Static Resource Allocation
Table
SSE Streaming SIMD Extensions
SSL Secure Sockets Layer
TB terabyte
TCO Total Cost of Ownership
TCP/IP Transmission Control
Protocol/Internet Protocol
UDB DB2 Universal Database™
URL Uniform Resource Locator
USB universal serial bus
VMFS virtual machine file system
VRM voltage regulator module
WAN wide area network
WMI Windows Management
Instrumentation
XML Extensible Markup Language
XON transmitter on

228
Planning and Installing the IBM Eserver X3 Architecture Servers

© Copyright IBM Corp. 2006. All rights reserved.
229
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
You can search for, view, or download Redbooks, Redpapers, Hints and Tips,
draft publications and Additional materials, as well as order hardcopy Redbooks
or CD-ROMs, at this Web site:
ibm.com/redbooks
The following Redbooks, Redpapers, and Hints and Tips are relevant. Note that
some of the documents referenced here may be available in softcopy only.
IBM Director 5.10, SG24-6188
IBM Eserver xSeries 260 Solution Assurance Product Review Guide,
REDP-4007
IBM ^ xSeries and BladeCenter Server Management, SG24-6495
IBM ^ xSeries BMC — Firmware and Drivers Quick Reference,
TIPS0532
Implementing VMware ESX Server 2.1 with IBM TotalStorage FAStT,
SG24-6434
Integrating IBM Director with Enterprise Management Solutions, SG24-5388
Introducing Windows Server x64 on IBM Eserver xSeries Servers,
REDP-3982
Remote Supervisor Adapter II Family — Firmware and Drivers Quick
Reference, TIPS0534
ServeRAID Adapter Quick Reference, TIPS0054
Server Consolidation with the IBM ^ xSeries 440 and VMware ESX
Server, SG24-6852
Server Consolidation with VMware ESX Server, REDP-3939
Tuning IBM ^ xSeries Servers for Performance, SG24-5287
VMware ESX Server: Scale Up or Scale Out?, REDP-3953

230
Planning and Installing the IBM Eserver X3 Architecture Servers
Other publications
These publications are also relevant as further information sources. These are
available from the URLs listed below or generally from:
http://www.pc.ibm.com/support
IBM ^ xSeries Solution Assurance Trigger Criteria (Worldwide), found
at:
http://w3.ibm.com/support/assure/assur30i.nsf/PubAllNum/SA200?OpenDo
cument
xSeries 260, xSeries 366, xSeries 460, and xSeries MXE-460 FAQ — Hints
and Tips, found at:
http://www.pc.ibm.com/support?page=MIGR-61395
xSeries Configuration and Options Guide, found at:
http://www.pc.ibm.com/support?page=SCOD-3ZVQ5W
Solving POST 1801 PCI Allocation Errors, found at:
http://www.pc.ibm.com/support?page=MIGR-61663
OSA System Management Bridge User’s Guide, found at:
http://www.ibm.com/pc/support/site.wss/MIGR-57816.html
xSeries 460 and MXE-460 publications
IBM ^ xSeries 460 and MXE 460 Enablement Kit for export-restricted
configurations, found at:
http://www.pc.ibm.com/support?page=MIGR-60747
IBM ^ xSeries 460 Type 8872 and xSeries MXE 460 Type 8874
Installation Guide, found at:
http://www.pc.ibm.com/support?page=MIGR-59865
IBM ^ xSeries 460 Type 8872 and xSeries MXE 460 Type 8874
Problem Determination and Service Guide, found at:
http://www.pc.ibm.com/support?page=MIGR-59864
IBM ^ xSeries 460 Type 8872 and xSeries MXE 460 Type 8874
User's Guide, found at:
http://www.pc.ibm.com/support?page=MIGR-59863
Rack Installation Instructions, found at:
http://www.pc.ibm.com/support?page=MIGR-59862

Related publications
231
xSeries 366 publications
IBM ^ xSeries 366 Type 8863 Installation Guide, found at:
http://www.pc.ibm.com/support?page=MIGR-58972
IBM ^ xSeries 366 Type 8863 Problem Determination and Service
Guide, found at:
http://www.pc.ibm.com/support?page=MIGR-58887
IBM ^ xSeries 366 Type 8863 User's Guide, found at:
http://www.pc.ibm.com/support?page=MIGR-58973
Rack Installation Instructions, found at:
http://www.pc.ibm.com/support?page=MIGR-58974
xSeries 260 publications
IBM ^ xSeries 260 Type 8865 Installation Guide, found at:
http://www.pc.ibm.com/support?page=MIGR-61926
IBM ^ xSeries 260 Type 8865 Problem Determination and Service
Guide, found at:
http://www.pc.ibm.com/support?page=MIGR-61239
IBM ^ xSeries 260 Type 8865 User's Guide, found at:
http://www.pc.ibm.com/support?page=MIGR-61110
Online resources
The following Web sites and URLs are also relevant as further information
sources.
IBM Web pages
Here are some useful IBM web links.
Solutions
IBM-SAP alliance
http://www.ibm.com/solutions/sap/us/en
IBM-SAP alliance (IBM internal network users only)
http://w3.ncs.ibm.com/solutions/SAP

232
Planning and Installing the IBM Eserver X3 Architecture Servers
SAP sizing documents
http://www.ibm.com/support/techdocs/atsmastr.nsf/PubAllNum/PRS261
IBM hardware sizing for applications
http://www.ibm.com/servers/solutions/finder/portal/hw-sizing.html
SAP Benchmark results
http://www.sap.com/solutions/benchmark
IBM-Siebel alliance
http://www.ibm.com/solutions/businesssolutions/siebel
IBM-Siebel alliance (IBM internal network users only)
http://w3.ncs.ibm.com/solutions/Siebel
Siebel sizing document
http://www.ibm.com/support/techdocs/atsmastr.nsf/PubAllNum/PRS593
IBM SAP International Competence Center (IBM internal network users only)
http://w3.ncs.ibm.com/solution.nsf/SOL/BTIY-5PBJ73
IBM-VMware alliance
http://www.ibm.com/eserver/xseries/vmware
IBM Datacenter Solutions with High Performance xSeries Servers
http://www.ibm.com/servers/eserver/xseries/windows/datacenter.html
Configurations
ServerProven
http://www.ibm.com/servers/eserver/serverproven/compat/us
ServerProven operating system compatibility matrix
http://www.ibm.com/servers/eserver/serverproven/compat/us/nos/matrix.shtml
DataCenter Supported Configurations
http://www.pc.ibm.com/support?page=MIGR-4P7RG3
IBM Export Regulations Office (IBM internal network users only)
http://w3.ibm.com/chq/ero
Solution Assurance (IBM internal network users only)
http://w3.ibm.com/support/assure
Solution Assurance (PartnerWorld users only)
http://www.ibm.com/partnerworld/techsupport

Related publications
233
Solution Assurance Trigger Criteria (IBM internal network users only)
http://w3.ibm.com/support/assure/assur30i.nsf/PubAllNum/SA200
Support and downloads
AP Technical Sales Support (IBM internal network users only)
http://w3.ibm.com/support/ap
Contact PartnerWorld
http://www.ibm.com/partnerworld/pwhome.nsf/weblook/cpw_index.html
xSeries driver matrix
http://www.pc.ibm.com/support?page=MIGR-4JTS2T
UpdateXpress
http://www.pc.ibm.com/support?page=MIGR-53046
OS installation instructions for xSeries systems
http://www.pc.ibm.com/support?page=MIGR-4QLNTQ
Installing VMware ESX Server 2.5.1
http://www.pc.ibm.com/support?page=MIGR-60546
Windows 2000 Server HAL for the x460
http://www.pc.ibm.com/support?page=MIGR-61296
ServerGuide
http://www.pc.ibm.com/support?page=MIGR-4ZKPPT
IBM Director
http://www.pc.ibm.com/support?page=SERV-DIRECT
Dynamic System Analysis
http://www.ibm.com/pc/support/site.wss/document.do?lndocid=SERV-DSA
SMBridge BMC utility
http://www.pc.ibm.com/support?page=MIGR-57729
IBM Director downloads
http://www.ibm.com/servers/eserver/xseries/systems_management/xserie
s_sm/dwnl.html
System Management Bridge Baseboard Management Controller CLI and
Remote Console Utility - Servers
http://www.ibm.com/pc/support/site.wss/MIGR-57729.html

234
Planning and Installing the IBM Eserver X3 Architecture Servers
Intel Web pages
EM64T technology
http://www.intel.com/technology/64bitextensions
Hyper-Threading Technology
http://www.intel.com/technology/hyperthread
Execute Disable bit
http://www.intel.com/cd/ids/developer/asmo-na/eng/149308.htm
Intelligent Platform Management Interface Specification V1.5, found at:
ftp://download.intel.com/design/servers/ipmi/IPMIv1_5rev1_1-012904ma
rkup.pdf
Microsoft Web pages
Microsoft Support on VMware
http://support.microsoft.com/kb/897615
Windows Server Catalog
http://www.microsoft.com/windows/catalog/server
Microsoft Emergency Messaging Service
http://www.microsoft.com/resources/documentation/WindowsServ/2003/st
andard/proddocs/en-us/EMS_topnode.asp
Special Administration Console commands
http://www.microsoft.com/resources/documentation/WindowsServ/2003/st
andard/proddocs/en-us/EMS_SAC_commands.asp
Others
“x86 Servers Brace for a Hurricane” by David Kanter, found at:
http://www.realworldtech.com/page.cfm?ArticleID=RWT042405213553
“Error Correcting Memory, Part 1”, by David Wang, which describes memory
hot-swapping, scrubbing, and bit steering:
http://www.realworldtech.com/page.cfm?ArticleID=RWT121603153445
VMware Download Center
http://www.vmware.com/download
Java Runtime Environment Version 5.0 Update 6 Manual Download
http://www.java.com/en/download/manual.jsp

Related publications
235
Help from IBM
IBM Support and downloads
ibm.com/support
IBM Global Services
ibm.com/services

236
Planning and Installing the IBM Eserver X3 Architecture Servers

© Copyright IBM Corp. 2006. All rights reserved.
237
Index
Numerics
05K9276, USB diskette drive 141
110V power 127
13M7409, memory card 95
13M7414, 2.3m scalability cable 107
13M7416, 2.9m scalability cable 107
13N0694, 3.16 GHz Cranford processor 90
13N0695, 3.66 GHz Cranford processor 90
13N0713, 3.33 GHz Potomac processor 90
13N0714, 3.00 GHz Potomac processor 90
13N0715, 2.83 GHz Potomac processor 90
13N2227, ServeRAID-8i 123
19K4164, USB-parallel port cable 35
25R8941, 2.67 GHz Paxville processor 90
25R8942, 3.00 GHz Paxville processor 90
30R5145, two 4 GB DIMMs 96
31R1558, export enablement CD-ROM 119
32R0719, Tower To Rack Conversion Kit 14
39Y6580, Dual Core X3 Upgrade Kit 19, 91
40K2450, memory card for 32-core 119
4816, Software Update Subscriptions 150
64-bit computing, benefits 26
64-bit processor 23
73P2865, two 512 MB DIMMs 96
73P2866, two 1 GB DIMMs 96
73P2867, two 2 GB DIMMs 96
8863 13
8865 14
8872 12, 149, 151
8874 12, 149, 151
A
Active Memory 29
Adaptec AIC-9410 31, 122
adjacent sector prefetch 93
AMD64 23
application integration 58
applications 52, 68
scalability 68
B
Baseboard Management Controller
See BMC
benchmarks 45, 69
benefits 44
BIOS
BMC configuration 122, 179
Clustering Technology 93
ESX Server 158
High Performance Memory Array 29
Hyper-Threading 92
memory configuration 29, 102
OS USB Selection 121
Periodic SMI Generation 121
Processor Adjacent Sector Prefetch 93
Processor Exectue Disable Bit 94
Processor Hardware Prefetcher 94
processor settings 92
Remote Console Redirection 196
RSA II 166
RSA II setup 120
SMBridge configuration 196
updating 140
Windows Server 2003 146
block diagram
SAS 32
X3 Architecture 8
blue screen of death 165
BMC 37, 177
add to IBM Director 187
add users 188
alert forwarding 193
BIOS settings 179
clock 181
connectivity 177
drivers 182
event log 180
features 177
firmware 140, 178
IBM Director integration 187
MPCLI 193
ports 186
remote control 182
SEL 180
SMBridge 194
TCP/IP ports 186

238
Planning and Installing the IBM Eserver X3 Architecture Servers
BMC configuration 122
bootcfg command 208
Broadcom Ethernet 35
Business Intelligence 53
buttons 15
C
cabling 107
cache 19, 28
XceL4v 27
Calgary 8
Certificate of Authenticity 148
chassis 4
Chassis ID 113
Chipkill memory 30
chipset
comparison 42
SAS 33
X3 Architecture 3
XceL4v 27
Clustering Technology 93
comparison 4, 45
compatibility mode, EM64T 24
complex 5
composite theoretical performance 118
connectors 15
console redirection 196
consolidation 56
cores 5
CPUs 18
Cranford 4, 18
memory addressing 27
create a partition 112
CRM 53
D
data integration 58
database applications 53
Datacenter
See Windows Datacenter
Datacenter High Availability Program Offering 148
Datacenter Scalable Offering 150
DB2 80
DDR2 memory 28, 96
default IP address, RSA II 166
default stripe size 123
diagnostics 140
differences 4
DIMMs 9, 28, 96
diskette drive 141
driver matrix 141
drivers
BMC 182
IPMI device drivers 182
Red Hat 154
RSA II 171
RSA II Linux driver 173
SAS 34
ServeRAID 34
SUSE LINUX 154
Windows 2000 Server 152
Windows Server 2003 145
DS300 126
Dual Core X3 Upgrade Kit 19
dual-core processors 5, 20
Dynamic System Analysis 220
E
eCommerce 54
EIDE interface 35
EM64T 23
Emergency Messaging Service 201
enterprise management 162
EnterpriseOne 73
ERP 52
event log 180
expanders, SAS 32
Export Regulation Office 119
export regulations 118
express models 13
F
far memory 10
fault tolerance 98
FC2-133 HBA 126
features 3, 44
firmware 106
BMC 178
RSA II 170
front view 15
frontside bus 21
G
Gallatin
memory addressing 27

Index
239
General Purpose Registers 24
grocery store analogy 10
H
HAL
Windows 2000 Server 153
Windows Server 2003 145
hardware prefetch 94
High Performance Memory Array 29
hot-add memory 30, 100
hot-replace
memory 30
hot-swap memory 99
Hurricane 8–9
Hyper-Threading 21, 92
optimized 138
OS support 137
I
IA-32e mode, EM64T 25
IBM Director 162
add users 188
adding service processor 187
alert forwarding 193
integration with RSA and BMC 187
MPCLI 193
padlock icon 188
Server Configuration Manager 188
upward integration 162
IBM Export Regulation Office 119
identical processors 18
installation
Red Hat 154
SUSE LINUX 154
VMware ESX Server 158
Windows 2000 Server 152
Windows Server 2003 145
Integrated xSeries Adapter 178
interleaving 9, 28
IPMI device drivers 182
Itanium 2
64-bit implementation 23
memory addressing 27
J
J.D. Edwards 73
Java runtime 174
K
KVM 174
L
latency 10
light path diagnostics 38, 220
local memory 10
logical processors 21
M
Maintenance Update Subscription 150
matrix 141
memory 28
64-bit addressing 24
Active Memory 29
addressability by processor 27
BIOS settings 102
Chipkill memory 30
disabling during POST 30
far 10
hot-add 30, 100
hot-swap 30, 99
interleaving 28
latency 10
local 10
memory card 13M7409 95
memory mirroring 30, 98
Memory ProteXion 29, 101
memory scrubbing 29
multi-node configurations 106
NUMA 10
performance 129
ports 9
remote 10
request queue 10
rules 97
throughput 9
use by XceL4v cache 141
merge timeout 114
merging nodes 107
MIB files, RSA II 176
Microsoft SQL Server 79
Microsoft Windows Server support 136
MIOC 8
models 12
express models 13
Windows Datacenter 149, 151
x260 14

240
Planning and Installing the IBM Eserver X3 Architecture Servers
x366 13
x460 12
modular expansion enclosure 5
MPCLI
add user 193
alert forwarding 193
MTOPS 118
multi-node
cables 107
concept 4
configurations 104
MXE-460 5
comparison with x460 66
drive bays 33
export regulations 118
front view 15
memory cards 28
memory configuration 96
models 12
OS support 136
power supplies 127
processor upgrades 90
rear view 16
removable media 125
x460 comparison 66
mySAP 72
N
NetWare support 136
Nocona 23
memory addressing 27
nodes 4, 67
NUMA 9
O
operating systems 136
Hyper-Threading 137
Red Hat 153
scalability 137
SUSE LINUX 153
VMware ESX Server 155
Windows 2000 Server 152
Windows Datacenter 147
Windows Server 2003 142
Opteron 23
memory addressing 27
Oracle 77
OS USB Selection 121
OSA SMBridge 194
Other OS setting, RSA II 167
overview 1
P
PAE switch 147
parallel ports 35
partial merge 115
partial mirroring 98
partitioning 6, 67, 116
Paxville 4, 18, 20
memory addressing 27
pay-as-you-grow 65
PCI bridge 9
PCI scan order 35
PCI-X slots 34
PeopleSoft 77
performance 129
CPU settings 130
memory 9
memory cards 129
memory settings 129
number of DIMMs 129
PCI-X in multi-node 130
Periodic SMI Generation 121
PFA 220
physical consolidation 58
PKT files
RSA II 170
ports
BMC 186
memory 9
RSA II 166, 176
SAS 32
SMBridge 198
positioning 41, 66
Potomac 4, 18
memory addressing 27
power control 107
RSA II 165
SMBridge 205
power supply considerations 126
primary nodes 7
Primary Scalable Node 113
Processor Adjacent Sector Prefetch 93
processor cache 19
Processor Exectue Disable Bit 94
Processor Hardware Prefetcher 94

Index
241
processor tray 19
processors 18
64-bit 23
BIOS options 92
Clustering Technology 93
dual-core 20
dual-core upgrade 91
hardware prefetch 94
Hyper-Threading 21, 92, 137
memory addressability 27
multi-node configurations 106
Processor Adjacent Sector Prefetch 93
rules 91
Q
queuing time 11
R
RAID levels 123
rear view 15
Red Hat 153
drivers 154
installation 154
RSA II drivers 173
SMBridge installation 200
SMBridge support 196
SOL, enabling 212
support 136
telnet 202
Redbooks Web site 229
Contact us xi
redundant bit steering 101
remote console redirection 196
remote control
BMC 182
RSA II 174
remote memory 10
Remote Supervisor Adapter II SlimLine 35–36
request queue 10
RSA II 164
add to IBM Director 187
add users 188
alert forwarding 193
alerts 164
BIOS settings 166
blue screen of death 165
configuration 120
connectivity 166
default IP address 166
default user 170
DHCP 167
driver 171
event log 164
features 164
firmware 140, 170
health monitoring 164
IBM Director integration 187
installation 173
installing 165
IP address 166
Java runtime 174
KVM 174
MIB files 176
MPCLI 193
multi-node configuration 110
network settings 166
OS USB Selection 121
Other OS setting 167
partitioning 6
Periodic SMI Generation 121
ports 176
remote control 174
remote media 165
SNMP MIB files 176
static address 167
TCP ports 176
video speed 174
Web interface 168
Windows drivers 172
rules
hot-add memory 100
memory 97
memory mirroring 98
multi-node configurations 112
processors 91
RXE-100
support 13
S
SAC commands 206
SAP 71
SAS 31
scalability 9, 137
scalability directory 27
Scalability Enablement/Memory Adapter 119
Scalable Partition Number 113

242
Planning and Installing the IBM Eserver X3 Architecture Servers
Scalable Systems Manger 163
scale-up 42
scale-up versus scale-out 82
scan order, PCI 35
scrubbing, memory 29
SEL 180
Serial Attached SCSI 31
Serial over LAN 194
SMBridge 204
serial port 35
Server Configuration Manager 188
server consolidation 54, 56
ServeRAID-6M 126
ServeRAID-8i 33, 123
ServerProven 130, 136
shopping cart analogy 10
Siebel 75
SMBridge 194
authentication 203
BIOS settings 196
bootcfg command 208
CLI 195, 217
commands 204, 218
connecting 203
console command 204
console redirection 196
daemon 200
EMS 201, 206
EMS, enabling 208
event log 206
exiting SOL 207
F1 key in telnet 202
GRUB 214
installation 197
LILO 212
Linux support 196
ports 198
power control 205
PXE boot 196
Red Hat, enabling SOL 212
remote console redirection 196
SAC 206
SAC commands 206
SAC support 201
Serial over LAN 204
service 200
telnet clients 201
telnet server 194, 201
timeout values 199
Windows SOL 206
SMP 9
snoop filter 27
Software Update Subscription
Datacenter 150–151
Solution Assurance 131
Special Administration Console 201
Special Mode 93
SQL Server 79
SRAT table 11
SSE 24
stripe sizes 123
Supply Chain Management 53
support
VMware, applications running on 157
SUSE LINUX
drivers 154
installation 154
RSA II drivers 173
RSA II information 155
SOL, enabling 215
support 136
system boards 3
T
target applications 52
TCP/IP ports
BMC 186
RSA II 176
Total Cost of Ownership 61
TotalStorage DS300/400 126
Tower To Rack Conversion Kit 14
two-way interleaving 9
U
Ultra320 SCSI 31, 126
upward integration 162
USB diskette drive 141
USB ports 35
V
video PCI adapters 35
VMware ESX Server 155
BIOS settings 158
installation 159
Microsoft support 157
pre-install 158

Index
243
support 136
VRMs 19, 91
W
Web interface 168
Windows
bootcfg command 208
EMS 201, 208
Hyper-Threading, affect of 23
RSA II drivers 172
SAC 201, 206
SMBridge
installation 198
support for 196
SOL support 206
Windows 2000 Server
drivers 152
HAL 153
installation 152
Windows Datacenter 147
Datacenter High Availability Program Offering
148
Datacenter Scalable Offering 150
x460 models 149, 151
Windows Server 2003
BIOS settings 146
boot from internal drives 146
drivers 145
installation 145
PAE switch 147
support 142
X
x255 comparison 50
x260
drive bays 33
front view 17
memory cards 28
memory configuration 96
OS support 136
power supplies 127
processor upgrades 90
processors 18
rack models 14
rear view 18
removable media 125
target applications 54
tower models 14
tower-to-rack conversion 14
x255 comparions 50
X3 Architecture 1, 42
x365 comparison 46
x366
drive bays 33
dual-core upgrade 19
front view 16
memory cards 28
memory configuration 96
models 13
OS support 136
power supplies 127
processor upgrades 90
processors 18
rear view 17
removable media 125
target applications 55
x365 comparison 46
x445 comparison 48
x460 66
drive bays 33
dual-core upgrade 19
export regulations 118
front view 15
memory cards 28
memory configuration 96
models 12, 149, 151
OS support 136
power supplies 127
processor upgrades 90
processors 18
rear view 16
removable media 125
Scalability Enablement/Memory Adapter 119
target applications 55
Windows Datacenter models 149, 151
x445 comparison 48
XA-64e chipset 8
XceL4v 27, 43
memory use 141
Xeon 7020/7040 20
Xeon MP 18
XpandOnDemand 43, 64, 104

244
Planning and Installing the IBM Eserver X3 Architecture Servers

(0.2”spine)
0.17”<->0.473”
90<->249 pages
Planning and Installing the IBM Eserver X3 Architecture Servers



®
SG24-6797-00 ISBN 0738492620
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
Planning and Installing
the IBM Eserver X3
Architecture Servers
Covers the xSeries
460, xSeries 366, and
xSeries 260
Learn the technical
details of these
high-performance
servers
How to configure,
install, and manage
the servers
The IBM Eserver X3 Architecture servers are the new
third-generation Enterprise X-Architecture servers from IBM.
The xSeries 366 was announced in February 2005, the
xSeries 460 was announced in May, and the xSeries 260 was
announced in August. These X3 Architecture servers are ideal
for random commercial workloads with high processor and
memory bandwidth requirements.
Although aimed at different client sets, they share common
components and options. The x460 is targeted at clients who
need to implement a large single image “scale-up”
configuration of up to 32 processors and 512 GB of RAM. The
x366 is targeted at the high-performance environment where
rack space is a premium. The x260 is for clients who need
high performance but also require large amounts of internal
disk storage.
This IBM Redbook provides a detailed technical description of
the three servers and explains how to plan, install, configure,
and manage these high-performance servers running 32-bit
and 64-bit versions of Windows Server 2003, Red Hat
Enterprise Linux, SUSE Linux Enterprise Server, and VMware
ESX Server.
Back cover
First page image
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.