Re: Monitor for existing job.

[Evan Harris <spanner@xxxxxxxxxx>]

Rob

a better approach depending on your situation might be to hold all the job
queues at the start of the backup job so that no new jobs can start and
extend the grace/wait period.

I use a control file that contains details of all the job queues and hold
them before the backup, then release them afterwards. A nice side effect is
that I can hold specified queues during business hours but release them
after business hours to assist interactive performance (This is where all
my daytime queries are forced to go)

Hope this helps
Evan Harris



>--
>[ Picked text/plain from multipart/alternative ]
>One problem I've had with many of these techniques is the following
>scenario:
>
>I have a job running which does an activity, sleeps for a set amount of
>time, checks a data area or something to see if it should shut itself
>down, if not it repeats the loop, otherwise - endpgm.
>
>Sounds nice and orderly.  Backup starts and sets the dataarea to shutdown
>the other program, waits until it ends, and then continues.  Don't want to
>do an endjob because it might be in a critical update.  Problem arises
>when, somehow or another, there is a job waiting on the job queue.  First
>one ends, second one kicks in, and then the backup is waiting again.
>Probably what I should do is have the backup update the dataarea again
>with the shutdown command every time it loops.  And after x loops, notify
>someone.
>
>Rob Berendt