Clarification on the CA Bundl AIF Merge/Split process

Document ID : KB000120611
Last Modified Date : 08/11/2018
Show Technical Document Details
Introduction:
Clarification on the CA Bundl AIF Merge (job=CC50J074 PGM=DPMB074) process and the CA Bundl AIF Split (job=CC50J075 PGM=CC50J075) process.
Question:
Can we get some clarification on the CA Bundl AIF SPLIT (CC50J074 PGM=DPMB074) process and the CA Bundl AIF MERGE (CC50J075 PGM=CC50J075) process?
Environment:
z/OS
Answer:
You will want to make sure that you run a legitimate VSAM file backup utility before beginning the process. Run your backups and the merge/split process AFTER successful completion of your End of Day cycle and before collecting any NEW data into CA Bundl. At the time you run the backup, you will want to make sure that "ALL" CA Bundl related tasks and batch processed are STOPPED. And the LSERV task that services these CA Bundl files should be DOWN. These are the circumstances under which you have to run the MERGE/SPLIT anyway. 

Clarification on the process...

The ultimate goal of the AIF SPLIT is to keep the contents of, specifically, the "AIF" file, as small as possible for maximum performance. As the size and utilization of the "AIF" file increases, the longer it takes the daily CA Bundl processes to process indexes into and out of the "AIF" file.

After a SPLIT, the "AIF" file will only contain the "current year" archived indexes along with "future archives" which will be added ONLY to the new "AIF" file. The "AIF0 through AIF9" files will only contain all the "older" archive data. 

Before you can do a SPLIT, you are going to have to run a MERGE. The MERGE will initially combine "ALL" of the archive data, old and current, from your existing 11 files (AIF, AIF0-AIF9) into a SINGLE and temporary AIF file. 

The SPLIT will then create a NEW "AIF" file as well as NEW files "AIF0 through AIF9" and will REDISTRIBUTE all of the data in the "SINGLE and temporary AIF file" out across all 11 new AIFs. Again, after the SPLIT completes, the "AIF" file will only contain the "current year" and all "NEW" archive indexes. The AIF0 through AIF9 files will contain the older data. 

The DPMB074 jobs control parameter "B074PRM=" basically dictates how to distribute the OLDER data across the "AIF0 through AIF9" files. It's typically based on how many years of archived data you keep in CA Bundl. If you have 10 years worth of data, consider setting "B074PRM=S1YEARS". Since there are 10 files to store the data in, each file would have about a years worth of data in it. If you have 50 years worth of data, consider setting as"B074PRM=S5YEARS". Then each of the 10 files would have about 5 years worth of data in it. A maximum of nine years is allowed for the split year range. 

* NOTE *
The DPMB074 control parameter has an optional setting (using a 'C' instead of an 'S') that you can use if you only want to generate the "control report". No actual splitting will occur if you use the 'C'. This 'C' option will list the number of records that would be added to each file in simulation mode.
Additional Information:
You can access a PDF version of the Bundl Documentation at the following link: 

https://docops.ca.com/bookshelves/bundl

The DPMB074 MERGE and DPMB075 SPLIT programs and the SPLIT/MERGE process are documented in the Ca Bundl Administrators Guide.