qdel not killing all processes started under Intel MPI

qdel not killing all processes started under Intel MPI

Hi, when we run using Intel MPI with Hydra process manager (in a script submitted with qsub-- this is with OGS/GE 2011.11p1 on ROCKS 6.1 on a small blade cluster), qdel does not fully kill the job except when the the jobscript runs on the frontend. I have to kill the processes started by mpirun manually if the jobscript runs on a compute node. This is not a problem with OpenMPI.

Any ideas or suggestions on how to proceed with troubleshooting this would be much appreciated.
Thanks,
Noah

6 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.

Hi Noah,

I believe you're already engaged with one of our support engineers via the Intel Premier Support portal.  You're welcome to post here again once your issue has been resolved for the benefit of others.

Regards,
~Gergana

Gergana Slavova
Technical Consulting Engineer
Intel® Cluster Tools
E-mail: gergana.s.slavova_at_intel.com

Thanks Gergana, I intended to post back here when the issue is resolved.

ALTHOUGH, I've tried many times today to post my latest response to the support interface (even tried to open a new issue), but I keep getting the error:

The requested URL was rejected. Please consult with your administrator.
Your support ID is: XXXXXXXXXXXXXXX

Can you please help with this? I will be offline until tomorrow morning, and will try again then.

Thanks,
Noah

Hi Noah,

Could it be a firewall on your side?  I would try restarting your browser (cleaning up cookies, etc.) and trying again.

I was having some issues accessing Intel Premier Support earlier today as well although that's been fixed now.  If this doesn't work tomorrow morning either, post here and I'll get this escalated.

Thanks,
~Gergana

Gergana Slavova
Technical Consulting Engineer
Intel® Cluster Tools
E-mail: gergana.s.slavova_at_intel.com

Tried on laptop at home (different computer, different network). Even tried a fully "cleaned" browser. Same problem. Did several submissions successfully today before this error started appearing. Frustrating to have to get support to get support!

Thanks again for any help.

Noah

Hi Noah,

Thanks for trying this.  I'll get this issue escalated internally.  In the meantime, I'll get you in touch with Dmitry (the support engineer you were communicating with) directly.

Regards,
~Gergana

Gergana Slavova
Technical Consulting Engineer
Intel® Cluster Tools
E-mail: gergana.s.slavova_at_intel.com

Laisser un commentaire

Veuillez ouvrir une session pour ajouter un commentaire. Pas encore membre ? Rejoignez-nous dès aujourd’hui