[GE users] Shutting down sgeexecd does not kill running job

Reuti reuti at staff.uni-marburg.de
Sat Nov 10 15:13:27 GMT 2007


Hi,

Am 30.10.2007 um 21:10 schrieb Orion Poplawski:

> Running 6.1u2 on Fedora.  If I shutdown sgeexecd on the execution  
> host (service sgeexecd stop), the jobs continues to run.  shutdown  
> output is:
>
> configuration wind.cora.nwra.com not defined
>    Shutting down Grid Engine execution daemon
>    Shutting down Grid Engine shepherd of job 174.1
>
> messages contains:
> 10/30/2007 14:03:33|execd|wind|I|controlled shutdown 6.1u2
>
> The job still appears in the running stat with qstat:
>
>     174 0.55500 job.csh    orion        Rr    10/30/2007 14:01:13  
> admin.q at wind.cora.nwra.com         1 1
>
>
> Is this expected?

yes, it's like "qconf -k <node>". But you can use:

qconf -kej <node>

to kill the job(s) there in addtition. As they will not be  
rescheduled in this case, maybe using qresub before is necessary.

-- Reuti


> If I restart sgeexecd, messages contains:
>
> 10/30/2007 14:07:45|execd|wind|W|local configuration  
> wind.cora.nwra.com not defined - using global configuration
> 10/30/2007 14:07:45|execd|wind|I|starting up GE 6.1u2 (lx26-x86)
> 10/30/2007 14:07:45|execd|wind|E|abnormal termination of shepherd  
> for job 174.1: "exit_status" file is empty
> 10/30/2007 14:07:45|execd|wind|E|can't open usage file "active_jobs/ 
> 174.1/usage" for job 174.1: No such file or directory
> 10/30/2007 14:07:45|execd|wind|E|shepherd exited with exit status 19
>
> the job goes back into the waiting queue and gets restarted, but  
> now i have two copies running.
>
> TIA,
>
>  Orion
>
> -- 
> Orion Poplawski
> Technical Manager                     303-415-9701 x222
> NWRA/CoRA Division                    FAX: 303-415-9702
> 3380 Mitchell Lane                  orion at cora.nwra.com
> Boulder, CO 80301              http://www.cora.nwra.com
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
> For additional commands, e-mail: users-help at gridengine.sunsource.net

---------------------------------------------------------------------
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