[torqueusers] parallel jobs left processes on computing nodes after qdel

wzlu wzlu at gate.sinica.edu.tw
Thu Mar 1 01:16:47 MST 2007


Dear ALL,

I am using torque-2.0.0p8 and maui-3.2.6p14 on RHEL 4 WS.
When delete a parallel job, the processes left on computing nodes and
kept running.

I refer the URL: http://bellatrix.pcl.ox.ac.uk/~ben/pbs/ to clean the
process:
1.put epilogue in /var/lib/torque/mom_priv
2.put mpicleanup in /usr/local/bin
But the processes still left on nodes.

Then I added "echo $JOBID $USER" >> /tmp/epilogue.log in epilogue
But /tmp/epilogue.log do not exist.
I think the epilogue do not execute.

Have any suggestion? Thanks a lot.


More information about the torqueusers mailing list