We are running CTS 2.2. In CPSM we have defined a Wlmspec as described in the manual Managing Workloads chapter 9 and RTA is not defined.There are three TORs and four AORs and there is a problem with the distribution of the transactions. With low transaction rates it is quite good but with increasing transaction rates the distribution gets very uneven. It seems like when CPSM has selected one of the AORs to receive transactions, it stays with it for 10-15 seconds before selecting another one. During this period, none of the other AORs have work to do. There are no error messages. Have you got any suggestions on how to make CPSM distribute the load more evenly in a near round robin manner?
I must confess that I do not claim to be a CPSM expert but several opinions seem relevant to this question.
My first observation is that CPSM does not do round-robin function. It tries always to select the best place for execution and that best place is not necessarily the next available AOR. My second observation is to ask why you are worried about these AORs not doing any work? If the routing accomplishes the goals and meets response time requirements, then why worry about where the execution proceeds within?
You do not say whether or not your TORs and AORs are in the same LPAR of in different ones. However, the 15-second observation leads me to think that it is an LPAR environment. I think this because when CPSM is monitoring across LPARs, various usage and statistical info flows every 15 seconds. Thus, one would expect a different routing decision to occur roughly every 15 seconds as that's when the data for the determination of the best place across the LPAR is updated.
However, this may not be the case as you also mention that things seem to route widely at low transaction rates but unevenly at moderate or high transaction rates. I do not believe the high rate situation as that would ensure what you think of as a proper distribution. So, I'm left with moderate transaction rates for your unease.
Again, I think I appeal to the "who cares" argument. As things are working to satisfy your requirements, so what? That's rather unhelpful but perhaps you have got your goal and queue WLM definitions imprecisely setup. You could try tinkering with them but I really don't recommend it. You seem to have a configuration with spare capacity and performance satisfies your users. I'd be very inclined to leave things alone.
I'd recommend that if you are still concerned about this, and I think you shouldn't, then your next step might be to look at the various Share/Guide presentations on CPSM. There was an especially good series of articles in recent issues of Xephon's CICS Update by Mr. CPSM himself.
CICS Technical Strategist -- CICS expert at Search390.com
Editor's note: Do you agree with this expert's response? If you have more to share, post it in one of our .VO7aaqqaAFk.0@/search390>discussion forums.
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.