RE: Why separate pools?

[Chris Bipes <chris.bipes@xxxxxxxxxxxxxxx>]

We have several shared and private pools on our systems.  We have critical
server job, our real time on-line authorization system for instance that we
give a private memory pool to.  We want these jobs to have sub-second
responses and I give the subsystem/pool enough resources to do so.  Printing
goes to it's own pool with limits set so the printers don't rob all the
memory for large AFP/IPDS to PCL print jobs.  Qinter runs in *interact,
server jobs run in *sharepool1 and batch jobs run in *sharepool2.  Both
share pools have an upper limit and lower limit set so they have what they
need and do not hog all the memory.  Oh programmers get a private pool so I
can keep them from taking over the system.

For those who don't know, when you define a subsystem, the subsystem monitor
itself runs in the first pool defined.  We define all of our subsystems with
pool 1 equal to *base and pool 2 something else.  I then set the routing
entries to use pool 2.  Jobs within the subsystem then run in pool 2 and the
subsystem monitor runs in pool 1, which I always set to *base.

Chris Bipes