topic Re: LSF time event not firing after job exit in Administration and Deployment

LSF time event not firing after job exit

jklaverstijn — Fri, 24 Apr 2020 08:55:11 GMT

This is SAS 9.4M6 on AIX. When scheduling flows to LSF 10.1 we use the setting "Continue running the flow and when the flow is complete, change the flow state". This works as expected for flows that only contain jobs. However, we see that when there is a time event that is supposed to fire after a job fails (exit with rc>0) the event never fires. So this seems to disobey the "continue running" directive.

In this picture you see a job that fails at 9:10. The flow exits at that moment and the time event for 9:20 never fires and its dependent job never runs.

Are we seeing expected behaviour? If not I will contact tech support but I am unsure what to expect here.

Many thanks in advance for your input,

- Jan.

PS: there is mention of fixing similar issues by cleaning up all LSF and PM caches and work items and restarting. We have tried those to no avail.

Re: LSF time event not firing after job exit

Mark_sas — Mon, 27 Apr 2020 15:46:33 GMT

That's an interesting question - i.e., whether time events are honored after a job fails within a flow whose policy is set to continue running. I think you're going to want to engage Tech Support on this one as I'm not sure if you're getting expected behavior.

Re: LSF time event not firing after job exit

jklaverstijn — Wed, 29 Apr 2020 13:44:17 GMT

Thanks @Mark_sas ,

We are preparing files to send to tech support. I will let you know as soon as we have come to a conclusion.

Cheers Jan.