"ERROR: unable to receive from qsupelx - blank message"

[ashok]

Recently, we've been submitting batch jobs through the Qube Python api (Building a Job object and then qb.submit(job)). Every now and then, we receive this error:

"ERROR: unable to receive from qsupelx - blank message"

And the job is not submitted and crashes when trying to get the job ID. Resubmitting the same job usually works. We're submitting each job individually, which is not ideal, but unfortunately we cannot easily change our tools to submit all the jobs at the same time. Any ideas what causes this?

[jburk]

If you're submitting in a loop, can you put a sleep(0.5) in the loop to slow down the iterations?

If this works, you can experiment with the sleep interval to find the smallest value that prevents this from occurring.

This is why qube has job labels, so you can submit all the jobs in one go, and you don't need the job ID from an upstream job to set the dependency for a downstream job, you make both jobs at once, give the upstream job a label name, and use the label name in place of the jobID in the downstream job's callback trigger value.  Then you can submit both jobs in the same list in one operation.

Check out the docs for job labels: http://docs.pipelinefx.com/x/WYJ