What is the overall CPU usage? If the REAL system is "overloaded", then
splitting work between LPAR's won't help. There is only 100% of the
real CPU.
Is your JES MAS across all LPARs? That was the last "LPAR" sharing
environment I used, but it was across multiple physical processors.
We run almost all of our batch jobs in the same class but use Thruput
Manager agents and limits to manage work and assign WLM service class.
If I opened up the same class on a MAS'ed LPAR and set the TM limits
appropriately, work would move to that MAS as JES selected it.
For example, I allow the following job limits on one LPAR:
10 "express" (WLM Express)
20 "Normal" (WLM Normal)
10 "slack" (WLM Discretionary)
13 "big cpu"
But only 25 "total" jobs, but some "system" jobs such as db log dumps
are exempt from that limit.
When we had a second system in the MAS, I could set TM limits for 10
express jobs and 2 slack jobs for the 2nd system. Jobs in those
categories could run on either system. Of course, other "resources"
dictate otherwise, a job bound to IMSPROD would be limited to that LPAR.
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
