We have determined the batch job resources are all coming from the pool for our
batch processing.
I was thinking under the covers some parts might run in QSPL, but not that we
have seen.
Nathan - either your system or ours has the QSPL subsystem modified - in ours,
*SPOOL is pool 1 and *BASE is 2..
I'm in the process of setting up a monitor (data queue) on the originating outq
and dividing up the processing. We not only convert the original spool, but
sometimes merge additional spools together, generate a .tif from the afpds
spool, then direct the afpds to various outqs around the country.
We have 2nd processor in the hardware, but not activated - it's in the budget
but they are slow to sign.
Thanks to all,
Jim
________________________________
From: Nathan Andelin <nandelin@xxxxxxxxx>
To: Midrange Systems Technical Discussion <midrange-l@xxxxxxxxxxxx>
Sent: Wed, January 30, 2013 3:45:34 PM
Subject: Re: Spool APIs performance and memory pools
Is the pool *SPOOL used when batch programs are calling QUSLSPL,
QUSRSPLA ...
The *SPOOL memory pool is associated with a subsystem, not an API. What
subsystem is your "batch process" running in? Only the QSPL subsystem on my
servers use *SPOOL, and even that one has *BASE memory pool in position "1" and
*SPOOL in position "2".
I doubt your performance issues have anything to do with memory constriction in
the *SPOOL memory pool, but check to see which subsystem your "batch process" is
running in, and the memory pool(s) associated with it.
If you can't throw more hardware and memory at the problem, you may need to
split your "batch process" into multiple instances that run concurrently and can
take advantage of multiple CPUs. That probably entails some redesign.
-Nathan
midrange.com
