[torqueusers] Re: Job stucks forever

Abhishek Gupta abhig at Princeton.EDU
Wed Jan 14 07:31:55 MST 2009


Hi All,
Just to add what I observed yesterday. The job runs for single node but 
even if I enter nodes=2, it do not start the job for the first time. Can 
someone tell me what parameters are restricting it?
Thanks,
Abhishek.

Abhishek Gupta wrote:
> Hi,
> I am trying to submit a job. Once the job is submitted, it shows that 
> the status is running while it is not doing anything. When I re-submit 
> the job keeping the previously submitted job, the newer one runs fine 
> and the old one still stays there as it is.
> Could someone help?
> The PBS script I am using is:
>
> #!/bin/bash
> ############################################################
> ## 'man qsub' for additional information.
> ############################################################
>
> #$ -v LD_LIBRARY_PATH=/usr/lib
> #PBS -N MCRX
> #PBS -l nodes=20
> #PBS -l cput=36:00:00
> #PBS -r n
>
> export NPROCS=`wc -l $PBS_NODEFILE |gawk '//{print $1}'`
>
> ### The following statements dump some diagnostic information to
> ### the batch job's standard output.
> echo The master node of this job is `hostname`
> echo The working directory is `echo $PBS_O_WORKDIR`
> echo The node file is $PBS_NODEFILE
> echo "=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-"
> echo This job runs on the following nodes:
> echo `cat $PBS_NODEFILE`
> echo "=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-"
> echo This job has allocated $NPROCS nodes
>
> echo "=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-"
>
> ### Change to the working directory of the qsub command.
> ###cd $PBS_O_WORKDIR
>
> cd $HOME/an
>
> ### the full pathname as is done here.
> mpd &
> mpiexec -n 20 $HOME/an/mcmcRX
> #mpiexec  $HOME/an/mcmcR
>
> Thanks,
> Abhishek.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.supercluster.org/pipermail/torqueusers/attachments/20090114/1b109e4d/attachment.html


More information about the torqueusers mailing list