Question & Answer
Why are the number for lookasides and reads in my CICS Performance Analyzer for z/OS (CICS PA) report on LSRPOOL EOD statistics much smaller than what OMEGAMON XE for CICS and the CICS STAT transaction displayed online the previous day? These statistics should not be reset so I don not know why the CICS PA post-processing shows such small numbers. The only thing I can think of is that our nightly automation that runs in batch to close the files, process them, and re-open them may have reset something since they all belong to the same LSR POOL. I would like to understand why the numbers are off.
The key piece of information is that your batch automation closed all the files using the LSRPOOL. When this occurs, the LSRPOOL is discarded and CICS records the statistics up to that time as unsolicited statistics (USS). If you re-run your CICS PA statistics report for USS instead of EOD, you should see the larger values you saw the previous day online with OMEGAMON and the STAT transaction.
This is described in the CICS Transaction Server documentation topic Interpreting LSR pool statistics at this link:
When the last open file in an LSR pool is closed, the pool is deleted. The subsequent unsolicited statistics (USS) LSR pool record written to SMF can be mapped by the DFHA08DS DSECT.
CICSPA CICS PA
02 February 2016