We have a batch job that’s being nomad run on command.
Subsequent runs return completed evaluations triggered by job-register with no placement failures, but no allocation is being made.
Why? How to debug? What docs to read on which evaluations result in allocation and which don’t?
Hi @ilibar-zpt,
What does the output of nomad job plan <jobfile> output?
If the job specification file has not changed since the last run, and the previous successful run of the job is still held within Nomad state, the scheduler will see the registration as a no-op as it will determine the job has run successfully. If you wish to trigger the same batch job multiple times, I would suggest looking into parameterized batch jobs, which would suit this use case.
Thanks,
jrasell and the Nomad team