[slurm-dev] Help parsing squeue output

[Andreas Davour]

I have some jobs stuck in the PD state and the TIMELIMIT columns say:
4-00:00:00

and the reason in the NODELIST(REASON) column is:
(PartitionTimeLimit)

The last is understandable, I guess. Somebody have tried to run longer than
the limit on the partition. Now, this means I have two questions.

1. What does this "4-00:00:00" mean?

2. Shouldn't this case be an error message when trying to srun or sbatch? Why
PD and not FAILED?

/andreas
--
Systems Engineer
PDC Center for High Performance Computing
CSC School of Computer Science and Communication
KTH Royal Institute of Technology
SE-100 44 Stockholm, Sweden
Phone: 087906658
"A satellite, an earring, and a dust bunny are what made America great!"

[Jette, Moe]

1. This is from the squeue man page (similar information appears in most of the slurm
man pages): Time limit of the job or job step in days-hours:minutes:seconds.

2. We have partitions/queues in which the time limits or nodes in the partition change
at night and on weekends. You can easily modify this behavior with a simple patch to
src/slurmctld/job_mgr.c if desired. Something like this should do what you want:
(not I have not tested this):

Index: src/slurmctld/job_mgr.c
===================================================================
--- src/slurmctld/job_mgr.c (revision 20947)
+++ src/slurmctld/job_mgr.c (working copy)
@@ -3365,10 +3365,7 @@
fail_reason = WAIT_PART_TIME_LIMIT;
}
if (fail_reason != WAIT_NO_REASON) {
- error_code = ESLURM_REQUESTED_PART_CONFIG_UNAVAILABLE;
- job_ptr->priority = 1; /* Move to end of queue */
- job_ptr->state_reason = fail_reason;
- xfree(job_ptr->state_desc);
+ goto cleanup_fail;
}

cleanup:

________________________________________
From: owner-s...@lists.llnl.gov [owner-s...@lists.llnl.gov] On Behalf Of Andreas Davour [dav...@pdc.kth.se]
Sent: Tuesday, August 17, 2010 7:37 AM
To: slurm-dev
Subject: [slurm-dev] Help parsing squeue output