none
"Microsoft.Hpc.Scheduler.Session.SessionException: Failed to get job owner from the scheduler." error? RRS feed

  • Question

  • Running our processes today we received the following: "Microsoft.Hpc.Scheduler.Session.SessionException: Failed to get job owner from the scheduler." from a couple of the jobs we submitted (within 2 seconds of each other) and was hoping to better understand what was going on and hopefully fix the underlying issue if we can.  Re-running the process it worked fine so seems some sort of transient issue.

    I googled the error and found 1 reference to an older forum post but can't seem to get any information on what this error might actually mean.  We do not use the cached credentials option, and pass a user/pass each run.  As mentioned in the other post, we are currently initializing multiple DurableSession/ClientBrokers, so wondering if it could be another manifestation of too much load on either the client or head machine.  We currently do not have any dedicated broker nodes, but can add those if you think might be helpful.

    Thanks!

    -Jason


    Wednesday, September 2, 2015 5:02 PM

Answers

  • This Session Exception while creating a SOA session is because the broker failed to obtain the user's SID from the job for access authorization. This could happen when the scheduler is temporarily unavailable or under high stress. If the Session Exception happens, the SOA client should catch it and try to create the SOA session again. Meanwhile please check the number of SOA sessions/jobs or other batch jobs running concurrently on the cluster and reduce the stress on the cluster if possible.

    BR,

    Yutong Sun

    • Marked as answer by Jason Lee 1234 Thursday, September 17, 2015 3:25 PM
    Sunday, September 6, 2015 3:44 AM
    Moderator