Page 1 of 1

Job schedules failing to run on schedule

Posted: Wed May 02, 2018 4:11 pm
by jackson.eyton
Hi, we have been scheduling jobs using the datastage director for some time now, it seems we had some issue over the weekend that caused one of the scheduled jobs PID to crash. This was a master controller sequence job that crashed, we recompiled it and everything it touches and confirmed it runs fine. Restarted the datastage servers and verified the schedules. Unfortunately none of the jobs are running on schedule as expected. I have a parallel job that I set a schedule on to test this and waited, but the job does not run on schedule. I can issue a job run from the director and that works fine. I re-entered the user credentials for scheduled jobs in DataStage administration for this project and verified the test passes. Does anyone have any suggestions on what else I can check?

Posted: Wed May 02, 2018 5:00 pm
by PaulVL
Define the behavior when you say they do not run on schedule.

Posted: Wed May 02, 2018 5:18 pm
by ray.wurlod
Check the setting of the system clock.

I know that sounds weird, but I did encounter it once back in the day - the O/S failed to update time from the network.

Posted: Thu May 03, 2018 4:49 am
by qt_ky
This may or may not be related, but for what it's worth, we had to install a client tier patch for Director in order to resolve some unusual rescheduling problems. We had obtained the patch by opening a case (PMR) with IBM Support.

JR53246 Fixed problem with rescheduling and encrypted job parameters

Posted: Fri May 04, 2018 1:33 am
by ray.wurlod
During a training class in version 1.0, we had a competition to design the fastest job. I won it by changing the system clock while the job was running, but then got disqualified on some technicality.

Posted: Fri May 04, 2018 8:20 am
by jackson.eyton
ray.wurlod thats funny! Your job was able to complete a task before it started?

As an update I received the follow update from IBM:
Please try this suggestion out and see if the scheduled job runs.



The setting in question is:

Local Security Policy -> Security Options -> User Account Control: Run all administrators in Admin Approval Mode



Setting this to Disabled seems to resolve the problem.



We also found that if we do not specify a username and password in the Administration client schedule page, then scheduling worked both with and without the setting.
...If the suggestion does not work, we have a patch that may help but I would need to confirm if it would work for 11.5. You would need to unschedule everything and reschedule the jobs though.
I was able to work with out AD specialist and, while he was not able to alter the policy which is set to enabled, he made some permissions changes for the scheduling user we have specified to the following directories and that resolved the issue in our dev environment:
\IBM\information server\projects

He's unsure why he had to do that as of yet however so we're still investigating. Definitely seems as if some permissions change was made to that user. I'll keep you all posted and thank you so much for the replies.