I'm not a WLM expert, but I do understand the basics and how it works with
forked OMVS (BPXAS initiator pooled) address spaces.

First: this redbook says that it was updated in 2010, but I don't see
evidence of that in the actual PDF which still says (sic) "*Forth Edition
(March 2008)" *

I've also reviewed again this SHARE presentation on WLM and USS, which is
pretty good but dated.

We have a client who is having trouble managing OMVS AS work for
forked-from-batch job work.
They are running (JES) batch jobs with our product which fork /bin/ssh
(which won't local-spawn due to extattr settings) in a separate OMVS
address space.  /bin/ssh actually forks another short-lived OMVS address
space when it starts up, but that goes away after authentication.

So the problem that one of our new clients reports that when they start up
a bunch of jobs at once that their CPU "spikes" and it causes problems for
other workloads.  This have never been a problem for us before, since our
clients usually have more WLM expertise than we do :-)

My assumption is that they need to make sure that they are properly
classifying both the batch job and the OMVS AS workloads in WLM and then
configure the correct goals for each that so that this isn't a problem for

Q1: Anything that I'm missing at a high level?

Q2: Any other suggestions or advice for best practices?
If I can develop some succinct advice for this general problem, we'll
document it somewhere on our web site, like we have for "Managing temp
filesystems on z/OS".

Q3: Just curious: are there any knobs for the pool of BPXAS initiators?
 They just seem to be started as needed and then kept for up to 30 minutes
of non-use.   No max-in-use, max-avail, min-avail settings?

Thanks in advance.

Kirk Wolf
Dovetailed Technologies

For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Reply via email to