The Question is:
I have a problem that one batch job will re-schedule itself's twice, so 2 batch
jobs run the second night --- occassional another batch job will be spawned
from these and so on.
I have other batch jobs that reschedule themselves with no problem.
Recommendations highly welcome.
Swisslog Software (UK)
The Answer is :
You will need to find the source of the second (re)submission.
It is possible that a sufficiently skewed set of system times within
an OpenVMS Cluster configuration can permit a batch job to be started
when its expected time arrives, and to be invoked on a cluster member
node with an earlier time (and as the job then resubmits itself with
an /AFTER=TOMORROW from the perspective of the system with the slower
system time) and the resubmitted job is then immediately triggered (by
the system with the later time).
One simple technique to avoid this is to issue a wait in the procedure
for an interval that is longer than the largest expected time skew, then
the procedure can resubmit itself for the next run.
Alternatively, use DTSS or NTP or similar tool to maintain the cluster
time -- see the OpenVMS FAQ for details.
If the above discussion does not address your environment -- if this
is not an OpenVMS Cluster configuration -- then please contact the
Compaq Customer Support Center with additional details of the
configuration and the errant procedure resubmissions. Expect to be
asked for details of the DCL procedure and the resubmission, as well
as information from the system accounting logs.