Why is AVG TASK CPU TIME increasing so much after migration from CICS 4.1 to CICS TS 1.3 ?
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
Due to the new system logger feature, we have a significant performance degradation in our production CICS after migration. After changing the LGDFINT parameter from 30 ms (default) to 5ms we got back our response times as we had in CICS 4.1, but CPU time is still up about 50-70%. Setting AKPFREQ higher has no effect.
When we start with CICS in the morning, the avg CPU time is ok and then the avg CPU time per task increases during the whole day. How could this happen?
You may well end up raising a PMR (problem) on this - but to do so you will have to do a lot more research. The only evidence you quote is woefully inadequate for a comment.
Are all transactions slowing down or just some of them? Does the slow down just involve more CPU or is elapsed time increasing? Is there an increase in I/O activity? Is DB2 or VSAM involved? Are you getting Short on Storage messages? Are there some transactions started which never run? What is going on in MRO-linked regions? Do you have long running transactions? Does the Logger involve the Coupling Facility? Is there sufficient Disk Space available for the Logger? What do the logger stats say? Are transactions correctly terminating? Does Storage Isolation affect things? etc. etc. etc.
I strongly suggest you take a SNAP early on in the day and then one later on, and start looking at the contents. Two possibilities come to mind: There could be some sort of Storage Leak in either CICS or your Application - you should be able to see this by looking at the Storage SubPools - so scanning activity takes longer the more things are in a chain. The other idea is that you have not got the logger setup correct and so whatever application-type activity is occuring is taking longer & longer to journal.
Dig Deeper on IBM system z and mainframe systems
Related Q&A from Robert Crawford
For better mainframe capacity planning, how do I convert CPU hours to MIPS? And is there a way to calculate the relationship between MIPS and MSUs?continue reading
I have two years of experience in mainframe technology, currently working as a mainframe developer. I want to change to Java technology.continue reading
I want to replicate DB2 from the mainframe to an AIX box since it's cheaper and the copy can be used for testing. Is this possible?continue reading
Have a question for an expert?
Please add a title for your question
Get answers from a TechTarget expert on whatever's puzzling you.