Job finishes ok but does no work

chux · Post by **chux** » Thu Jul 12, 2007 7:48 pm

Hi Guys,

I'm seeing a strange intermittent error on Datastage 8 server jobs.

We have set up some sequences to run serveral jobs in a loop.

For the most part, everything works. However on occasion the jobs are 'started and then finished ok but no work actually happens. Ie the job does nothing.

When this happens in the log we see the

Starting job $foo
Environment variable settings
Finished job $foo

as opposed to various sequential file and transformer stages in between environment variable setting and job finishing.

Any ideas will be much appreciated.

chux · Post by **chux** » Thu Jul 12, 2007 7:51 pm

Additionally, the problem happens not on any particular job but seems to be random and at random interations of the sequence loop.

A restart of the Datastage server did not resolve this.

A restart of the machine on which DS was running temporarily seemed to resolve the issue for a few days.

DSguru2B · Post by **DSguru2B** » Fri Jul 13, 2007 6:18 am

Could very well be a bug. Are you sure that the job started at that particular time or it never got fired and the "Finished" status is from previous run?
Contact support and see what they have to say about it.

rleishman · Post by **rleishman** » Thu Oct 25, 2007 10:50 pm

I know it's an old thread, but I'll throw my support in as well.

This is happening to us on 8.0 as well. We have a number of small multi-instance administrative Server Jobs where this happens. We also have some multi-instance Parallel jobs, but we do not experience the same problem.

We don't have any looping jobs, so the experience is not identical.

We find that when a job starts to exhibit this problem, THAT INSTANCE of the job - when rerun - will continue to do nothing whereas other instances (ie. different invocation id) continue to work. We rind that a recompile fixes the problem temporarily.

jdmiceli · Post by **jdmiceli** » Mon Oct 20, 2008 1:39 pm

We are having a lot of this same issue right now, especially with my project that is all multi-instance jobs (about 1200+). We have applied a fix pack already and done some tweaks. Tonight another is going into place that is specific to this issue. After testing it tomorrow morning, I'll let you know if it worked