[GE users] Parallel jobs don't terminate

markhewitt mh613 at york.ac.uk
Thu Jul 30 15:21:42 BST 2009

> Neither is a simple switch, you job script and possibly you MPI
> installation needs to be tailored to it.

I'm using mvapich 1.1 over infiniband. The parallel environment 
configuration and job launch scripts are exactly as you would find in 
the example templates provided with SGE. e.g.

mpirun_rsh -rsh -np $NSLOTS -hostfile $TMPDIR/machines xhpl

It works fine launching the jobs, but as I've said when you qdel them, 
it does not kill the processes. I've looked through this in very close 
detail and I cannot find anything wrong with the configuration, and yet 
I'm still getting a system overrun by unterminated MPI jobs!



To unsubscribe from this discussion, e-mail: [users-unsubscribe at gridengine.sunsource.net].

More information about the gridengine-users mailing list