I've finally gotten around to writing my MONITOR job. Now that it is legacy
(has been in production for twenty-four hours and will not be removed), I'm
trying to figure out how to best restart it for each day. I have it loop
until its next loop would be after midnight, and then exit. My immediate
idea is to add
!IF JOBCNT('!HPJOBNAME,!HPUSER.!HPACCOUNT') = 1
!   STREAM HPFILE;AT=0:00
!ENDIF
after the ENDWHILE. But somethings nagging at me that this will bite back,
because there's something here I haven't thought of here. Comments,
suggestions?

For the record, the idea is to have something entirely independent of our
scheduler, which will monitor things such as that the scheduler is running,
and complain appropriately as needed. Its utility is in that the only shared
dependancy between it and regular scheduling and batch processing is the
hardware and OS. It will tell us if the scheduler fails, and the scheduler
will tell us if it fails (Whats Up will tell us if the system fails, and I'm
trying to figure out how to tell if Whats Up fails).

This job's loop executes a set of command files, which: check that either
our end of day job or our regular background jobs are running; list any
executing jobs other than our background jobs; checks that job limits minus
job count is zero or one; and monitors how many jobs are waiting now versus
during the last loop and notifies if both are non-zero. I'm trying to
determine what else is worth monitoring in this way. I'm sure this list will
slowly creep forward.

Greg Stigers

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *