×
The internal search function is temporarily non-functional. The current search engine is no longer viable and we are researching alternatives.
As a stop gap measure, we are using Google's custom search engine service.
If you know of an easy to use, open source, search engine ... please contact support@midrange.com.
rob wrote:
I wouldn't blanket say you must leave any job queues associated with QCTL
or QINTER running anymore than I would say you need to leave QBATCH
running. They need to be studied. Now, I could never see a reason to
hold QLPINSTALL. And doing a huge "hold 'em all" before a upgrade or ptf
install can slap you silly. But even for that situation it would probably
be save to hold the job queues associated with QCTL and QINTER.
Not everything goes through a queue.
Yeah, I mostly agree. Personally, I wasn't exactly worried about
"ALL" jobqs as much I was about 'many' jobqs that may be used by the
system as it starts TCP/IP and all the various servers, or possibly
other considerations.
I've never actually tried to start a system with "ALL" jobqs held. I
might have to try it just to see what happens (or doesn't happen).
I was trying to picture the two sides of it -- the HLDJOBQ [*ALL]
side to bring things to a stop, as well as the possible eventual
RLSJOBQ [*ALL] sometime after the system comes back up. Maybe the
restart side isn't as important as long as someone is available to
do it manually.
On either side, it still seemed reasonable that there could always
be jobqs that shouldn't be held or shouldn't be released. And that
had me wondering if "ALL" might really be a good idea under the
circumstances.
Unfortunately, ENDSBS doesn't seem appropriate either. The issue of
jobs starting when they shouldn't be run after the system comes back
up is significant. One related possibility is CHGIPLA CLRJOBQ(*YES)
in the UPS monitor program. I don't like that idea much, but it
_might_ be appropriate. It might be nice if the parameter was
HLDJOBQ(*USER) and jobqs had an attribute that kept system queues
separated.
Every really useful idea I can think of involves creating a list of
jobqs and processing the list. That's either by DSPOBJD *ALL/*ALL
OBJTYPE(*JOBQ) to an outfile or by a call to QUSLOBJ. Then, apply
possible filters to the list in order to handle exclusions. The list
could be re-used after the system restarted.
But that approach was pretty explicitly out of the question.
Tom Liotta
midrange.com
