RE: Separate subsystem - Devices not being allocated to it.

[Buck Calabro <Buck.Calabro@xxxxxxxxxxxx>]

>One thing that's made me wonder is that claims are made that
>the _last_ subsystem to start is the one that succeeds in
>startup allocation. That seems to be not totally true though I
>can't prove it at the moment. I'd swear that it works more like this:
>
>  STRSBS QINTER
>  STRSBS OTHERSBS
>
>Although OTHERSBS started last, QINTER typically is processing
>a long list of potential workstations due to a *ALL entry
>while OTHERSBS only has a few. OTHERSBS finishes all its
>startup work quickly while QINTER is still chugging along. If
>the OTHERSBS workstations are towards the end of the QINTER
>list, those can still be allocated to QINTER unless *ENTER
>explicit handling is stated in a workstation entry.

My QINTER has been up since Sunday night (it's now Friday morning)
I have a *ALL workstation type in QINTER.
I start an emulation session, get to a sign-on display and it's QINTER,
QPADEV0024.
I create a new subsystem BUCK and ADDWSE SBSD(BUCK) WRKSTN(QPADEV0024).
STRSBS BUCK.
The emulation session now reads BUCK QPADEV0024.
ENDSBS BUCK.
The emulation session now reads QINTER QPADEV0024.

I don't think this is a timing issue in how long it takes QINTER to process
it's list of eligible devices.  I think it's a case of "the last subsystem
to ask for a device gets it."  I'm sure there's no guarantee, but I've been
starting my "special" subsystems last since S/38 days, and not had a problem
with devices in the wrong subsystem.

When I need a guarantee, I use the initial program.
  --buck