One job blocking all servers?...

One job blocking all servers?...

Portrait de Peter Sels

Hello IntelMTLteam / Mike,

Thanks to your great support, I used to be able to submit jobs to the
queue to completion. But now, I submitted one to machine 2 and one to machine 3.
But I see:
Job id Name User Time Use S Queue
---------------- ---------------- ---------------- -------- - -----
12768.acaad01 fraktal_pauliny clls-s02 335:35:4 R workq
12778.acaad01 platform2log sels 0 Q workq
12779.acaad01 platform2log sels 0 Q workq
and this does not change at all, over days. So it seems the fraktal_pauliny is taking a lot of time, ok.
But does it block all machines? It should be running on a specific
machine and the
others should still be accessible, right?

Since I need to specify a specific machine (for my license key
reasons), I tried 02 and 03:

[sels@acano01 Ocapi]$ more queuePlatform2log.sh
qsub -lncpus=4 ./platform2log.sh

[sels@acano01 Ocapi]$ more platform2log.sh
#!/bin/sh
#PBS -N platform2log
#PBS -j oe
#PBS -l host=acano03
#PBS -l walltime=20:00:00
export OMP_NUM_THREADS=4
time /home/sels/projects/KUL/Ocapi/Debug/ocapi 0

once for acano03 and once for acano02.
These are the queued jobs 12778 and 12779 mentioned above.

Would you know what is going on please? thanks and best regards,

Peter

3 posts / 0 nouveau(x)
Dernière contribution
Reportez-vous à notre Notice d'optimisation pour plus d'informations sur les choix et l'optimisation des performances dans les produits logiciels Intel.
Portrait de Mike Pearce (Intel)

The reason for the jobs staying in the queued state, is a result of aPBSlicense issue, that we believe will be resolved shortly.

Portrait de Mike Pearce (Intel)

Fixed we believe - let us know if you still see issues with the batch system.

Connectez-vous pour laisser un commentaire.