Re: Can the subsystem a job is running in make it run slowly?

[Jim Oberholtzer <midrangel@xxxxxxxxxx>]

If that graphic is correct, part of the problem is all the jobs in QPGMR 
are running *BASE and dollars to donuts there is not enough memory in 
there and the activity level is not set correctly.  This system needs 
someone who knows work management to tune it.  Also it's a reasonable 
guess that most everything else is running in *BASE as well.

Step one:  create at least one shared pools, (wrkshrpool) and set one up 
for this batch, use *INTERACT for the interactive subsystem
    CHGSHRPOOL POOL(*SHRPOOL15)
           SIZE(1500)   (pick a number here that makes sense)
           ACTLVL(18)
           PAGING(*CALC)
I picked pool 15 because its most likely not used.  If shared pool 1 is 
not in use, change that to 1.

Step two:  Change the subsystem QPGMR with the following command:
                CHGSBSD SBSD(QPGMR) POOLS((2 *SHRPOOL15))

Step three: (the one most people miss)  Change the routing step to move 
the jobs into the second subsystem memory pool
  CHGRTGE SBSD(QPGMR)
          SEQNBR(9999)
          POOLID(2)

Now all the new jobs will run in *SHRPOOL15.  You can set the memory as 
needed.

Jim Oberholtzer
CEO/Chief Technical Architect
Agile Technology Architects, LLC


On 12/10/2010 4:17 PM, James H. H. Lampert wrote:
> sjl wrote:
> > >  Yes, if it is thrashing because too much work is being done in the
> > >  subsystem - If the number of activity levels is too low for the amount
> > >  memory and and the number of jobs allowed to run simultaneously in a given
> > >  memory pool, thrashing will occur.
> It's not a very active subsystem: it's one of twelve jobs in a subsystem
> that says it will take a maximum of 17.
>
>   From the "Work With Subsystems" display:
>
> Opt  Subsystem     Storage (M)   1   2
>        QPGMR                 .00   2   2