On 09/27/2014 01:36 AM, Shane Ginnane wrote: > As a distraction from the shellshock sideshow, > > On Fri, 26 Sep 2014 15:54:05 +0000, Gary Snider wrote: > >> I am trying to understand how to utilize the RMF CPU report so that I can >> evaluate utilization of our CP and latent demand. > Tread with care - there be dragons. > RMF is a sampler - on such a lightly loaded system I'd consider any variation > as rounding error. There is no overt evidence that your PROD LPAR isn't > getting all the CPU it wants. > My first question would be "what made you go looking ?". If workload isn't > meeting (perceived) goals, that's a whole different kettle of fish. And I > wouldn't be blaming PR/SM. > If you think there is latent demand, you should be concentrating on the CPU > Activity portion of the report, not the Partition Data portion. > > Shane ... > > There is a reason "latent" demand is called "latent" (hidden) demand. There is no way RMF data is going to show latent demand that is represented by users choosing to do less work on the system and spending their time doing other things or doing their work by other means because they perceive system response as being too slow.
If you are running with LPARs capped, occasionally hitting the cap, and running with less than full machine capacity, then you actually have a way to measure short-term latent demand: raise the cap during a peak period and see how much more resource the LPAR uses. Then the question becomes whether satisfying the latent demand is sufficiently important to justify to the company the extra software costs likely involved. -- Joel C. Ewing, Bentonville, AR [email protected] ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
