Preparing your result...
Loading...
Press Esc to dismiss this message

Multi-Dump Analysis (02-Feb-2010)

Thumbnail
IP.com Prior Art Database Disclosure (Source: IPCOM)
Disclosure Number IPCOM000192795D dated 02-Feb-2010
Originally published in Prior Art Database
Disclosed by: IBM
Country: Undisclosed
Disclosure File: 1 pages / 22.1 KB / English (United States)

Currently, clients run a fair amount of software which includes regions on many of or every system in their SYSPLEX, such as VSAM RLS processing. When problems occur in SMSVSAM on one system, dumps are automatically triggered from each system in the plex. Debugging of the problem then consists of evaluation of each dump independently using the Interactive Problem Control System (IPCS). This can be a time consuming process, as some clients use upwards of 32 systems in a PLEX, and are likely to continue to grow as mainframe capabilities grow.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 79% of the total text.

Page 1 of 1

Multi-Dump Analysis

This invention works through an expansion of the current Interactive Program Control System (IPCS) with an additional data descriptor field. Currently, data inside IPCS is divided by ASID and dataspace. This would add another field on the internal handling of the data to specify system name. At load time, instead of loading a single dump, the program would allow the user to specify multiple dumps from multiple systems. IPCS would then discern from each dump the SYSTEM NAME, and then classify all of the data accordingly. Browsing of the data would include a new keyword to allow for selecting data from a particular system, such as SYSNAME. This would allow analysis commands to retrieve data from a single system or multiple systems with a single command.

Since IPCS would have the data from several different dumps merged into a single entity, it would allow the user to do additional analysis that is not easily available today. For example, today when a client views the logrec in a single dump in a series of dumps, they must view each dump separately and keep track of the events. With the merged dump, the client can view the combined logrec in chronological order to see the sequence of events as they unfolded across multiple systems. We do this by using the SYSNAME as a unique identifier on each line where the origin of the system changes. This same logic is also used to merge the MTRACE section of the dump...

(Source: IPCOM)
First page image
(Source: IPCOM)