I can submit a job using a script, but if I try and run an interactive job, it just hangs. The system is free, no other jobs are running, so it is not a matter of resources. Just hangs and never produces a prompt. What should I check for? Thanks.
Please check:
- DNS issues (forward and reverse address resolution)
- firewall
- SELinux
Please browse this discussion: Interactive Job errors out with 'apparently deleted'
Well, a co-worker discovered the problem and offlined compute-0-0. Doesn’t explain why we could submit a job and it would run, but not perform an interactive job. Thanks.
1 Like
I am experiencing interactive job submission.
qsub -I
Submit the jobs, then deletes it. Everything was working until a power failure Friday.
We brought everything back up. I can submit a batch job, but not an interactive job.
I also found this in the server log, I removed the ip info:
req_reject;Reject reply code=15139, aux=0, type=8,
Please check
- Firewall
- Please check DNS for forward and reverse address resolution
- Please check this debug method
Interactive Job errors out with 'apparently deleted' - Use the full path of the qsub , if in case you have multiple WLM’s
/opt/pbs/bin/qsub -I (capital I for ice ) - Share the tracejob of the interactive job id