Hello!
I'm using slurm-2.5.6 (and I cannot upgrade for serious reasons), and today I got very strange situation: I submit a job and it becomes running in some time (squeue shows it as R), but on target nodes there are no job processes (job steps). In slurm logs on nodes I cannot see any messages about these job steps start, but on master slurm node I see that:
... backfill: Started JobId=1226291 on node6-155-[10-11] How can I fix it or even diagnose?
