SYSVIEW Performance Management Option for DB2 for z/OSDatabase Management for DB2 for z/OS - Performance SuiteDatabase Management for DB2 for z/OS - SQL Performance SuiteInsight Performance Monitor for DB2 UDB for z/OS
Issue/Introduction
Looking to debug a job that runs 8-12 minutes sometimes, and 8-12 hours others using the following probes. The problem is that they'are only getting about 5-10 seconds worth of data using these NEVENTS values:
HPRACCT = 1500
HPRSTDTL = 50000
HPRSTLST = 15000
HPRSTSML = 15000
How to increase the time-frame of data being collected without using more than 2gb worth of memory?
Environment
Release : 20.0
Component : CA Insight Database Performance Monitor for DB for z/OS
Resolution
To keep the traces running longer, increase the NEVENTS value and specifying NOWRAP=NO in the HPRACCT, HPRSTDTL, HPRSTLST and HPRSTSMLrequests in Hghlevel.REQUEST library. The max value for NEVENTS is 65535. The changes to the above requests will require to increase the address space region size for the Data Collector increase storage usage. The SHOW display in DIAGNOSTIC mode will show how much storage each started request is using. Also, it may not be necessary to make the changes to all the requests - for example, the HPRACCT request will only have one record for each of the threads traced so it may not need to have NEVENTS increased, or maybe it only needs to be increase by a small amount.