[torqueusers] qdel jobs are not cleanup
samuel at unimelb.edu.au
Tue Sep 14 18:56:41 MDT 2010
-----BEGIN PGP SIGNED MESSAGE-----
On 15/09/10 01:38, Govind wrote:
> Sometime it is happening that a job is deleted(qdel) from
> submission host by user are not shown by qstat/showq.
> But job is still running on the client node.
This can happen if you are running MPI jobs and your
mpirun/mpiexec uses SSH to launch jobs across the cluster.
The solution then is to either use an MPI stack that
can be configured to use the Torque TM API to launch
jobs (such as Open-MPI) or, if you can't do that, try
the OSC mpiexec replacement which can be built with
support for various MPI stacks and knows about Torque.
Christopher Samuel - Senior Systems Administrator
VLSCI - Victorian Life Sciences Computational Initiative
Email: samuel at unimelb.edu.au Phone: +61 (0)3 903 55545
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
-----END PGP SIGNATURE-----
More information about the torqueusers