Actually Job runs once in a month and this month It got abended because
Not able to recall the dataset.
after some time when I resubmit the job by changing msgclass parameter
it ran fine.
someone suggested me to add one more step in the beginning of the job so that I recall the dataset first and later executes remaining step.
Joined: 20 Oct 2006 Posts: 6970 Location: porcelain throne
on our machine, they migrate ds after 3 days - which makes mondays a lot of fun.
I wrote a rexx script that generates an iefbr14 for all of my datasets,
with a disp=shr.. before submitting, I cull those ds's that I will probably not use.
i made the mistake of submitting the job with a CLASS = 20 second jobs,
the recalls required nearly an hour (wall clock), though the job did not abend due to timeout, did tie up a job class much too long.
since then I submit under a CLASS = nolimit.
never have experienced a job timing out due to migration recall.
because job fails to recall migrated dataset
I think the TS's was timed out due to another reason.
i base that on the ts's comment about msgclass having an effect,
and not knowing enough about JCL to create his own step to effectively 'recall' a migrated ds.
Joined: 14 Mar 2007 Posts: 8657 Location: Back in jolly old England
Dick, if you do need to do it, it would probably be better to generate the HRECALL cards with NOWAIT as this will queue the recall requests in HSM, and HSM is clever enough to group the recalls by tape so that each ML2 volume is mounted only once.
there are also links within this website that provide the same info.
e.g. HRECALL to recall all migrated datasets
again, thx for the tip expat .... always looking for ways to do things,
and not attract the attention of operators.