[GE users] MPICH2 GridEngine (6.1 u2)

Reuti reuti at staff.uni-marburg.de
Tue Nov 20 15:49:56 GMT 2007


Hi,

Am 20.11.2007 um 13:40 schrieb Spiro Stathakis:

> Just got a tightly integrated mpich2  (smpd method) working with grid
> engine by following this how to:
>
> http://gridengine.sunsource.net/howto/mpich2-integration/mpich2- 
> integration.html
>
> The job runs fine however qstat still shows the job running when the
> computation has finished. If i query the processes on the nodes the
> executable has indeed finished.
>
> This grid engine process is still running
> /opt/mpich2_smpd/pathscale/bin/smpd -port 21324 -d 0
>
> I have to qdel the job id to free up the for the next job. Does anyone
> know why SGE is not cleaning up ?

this looks like the smpd wasn't started by a Tight Integration. Can  
you please post:

ps -e f -o pid,ppid,pgrp,command --cols=80

(f w/o -) to get a complete process tree of a node with a running job.

-- Reuti

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
For additional commands, e-mail: users-help at gridengine.sunsource.net




More information about the gridengine-users mailing list