[torqueusers] Job not running due to features when node name specified

Ken Nielson knielson at adaptivecomputing.com
Fri Aug 3 11:23:49 MDT 2012


On Mon, Jul 30, 2012 at 10:23 AM, Andrus, Brian Contractor <bdandrus at nps.edu
> wrote:

>
> I am a bit confused as to how to troubleshoot and understand why this is.
>
> Running torque 2.5.12 and moab 6.1.6
> I submit a job with a specific node:
>         qsub -l nodes=compute-3-1
> It queues up fine, but never runs.
> qshow shows it as eligible and queued.
> When I run checkjob -v on it it says it shows:
>         compute-3-1              rejected: Features
>
> ??? Um ok...
> qstat shows:
>     Resource_List.nodect = 1
>     Resource_List.nodes = compute-3-1
>     Resource_List.pmem = 1gb
>
> But I can force it with 'qrun 839' and it does run on the node requested.
>
> One thing I would REALLY like to know is how to determine the specific
> features a job is being rejected for on a particular node.
> And why would a job be rejected due to features when it clearly can and
> should run?
>
> FWIW, compute-3-1 has no other jobs on it at the time.
>
>
> Brian Andrus
> ITACS/Research Computing
> Naval Postgraduate School
> Monterey, California
> voice: 831-656-6238
>
>
>
Brian,

if compute-3-1 is the name of the host then this is strange indeed. We run
jobs with -l nodes=<host> all the time. Do you have some Moab logs and
TORQUE server logs when this happens?

Ken
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.supercluster.org/pipermail/torqueusers/attachments/20120803/3aa5b604/attachment-0001.html 


More information about the torqueusers mailing list