[GE users] File Staging (local)

Vijay Avarachen vavarachen at gmail.com
Thu Mar 17 22:27:49 GMT 2005


    [ The following text is in the "ISO-8859-1" character set. ]
    [ Your display is set for the "ISO-8859-10" character set.  ]
    [ Some special characters may be displayed incorrectly. ]

I did have prolog/epilog defined, but I didn't think they would make a
difference if I am not using the functionality.  Anyways I disabled
it, and tried to run the job again but same problems.  This time
however I actually got a error output from abaqus saying:
ABAQUS Error: Unable to create a temporary directory
"/tmp/115.1.all.q/root_2d-4wave-rigid-dies_19790" to be used for
scratch files.

Also job.sh.o115 file was created by contains no information.  On the
compute nodes message file it still says:
03/17/2005 09:30:03|execd|node2|E|shepherd of job 109.1 exited with
exit status = 7
03/17/2005 09:30:03|execd|node2|W|reaping job "109" ptf complains: Job
does not exist

Here is my job script:
#!/bin/bash
#Make sure that the .e and .o file arrive in the working dir
#$ -cwd
#Mege the standard out and standard error to one file
#$ -j y
#
#$ -S /bin/bash
#
cd /working/2d-4wave-rigid-dies
/usr/bin/abq651 job=2d-4wave-rigid-dies cpus=2

I didn't get much time to work on this today (had to deal with a
lvm/backup issue).

Thanks,
Vijay Avarachen



On Thu, 17 Mar 2005 21:02:54 +0100, Reuti <reuti at staff.uni-marburg.de> wrote:
> Do you still have the prolog defined, which is now no longere necessary (as the
> file is already accessible via NFS)? Do you have any error file
> sge-date-run.e106? - Reuti
> 
> Quoting Vijay Avarachen <vavarachen at gmail.com>:
> 
> > Ok for the now I make /working on master a nfs share.  Mounted it rw
> > on all nodes and tried to run the abaqus job.  Still no luck.
> >
> > 03/17/2005 09:01:32|execd|node1|E|shepherd of job 106.1 exited with
> > exit status = 7
> > 03/17/2005 09:01:32|execd|node1|W|reaping job "106" ptf complains: Job
> > does not exist
> >
> > Worst part is that the error output file that it created
> > sge-date-run.o106 was empty so I could not tell why it did not run.
> >
> > Vijay Avarachen
> >
> > On Wed, 16 Mar 2005 23:58:44 +0100, Reuti <reuti at staff.uni-marburg.de>
> > wrote:
> > > Quoting Vijay Avarachen <vavarachen at gmail.com>:
> > >
> > > > > question: you can make a passwordless rcp from an exec host to the
> > head
> > > > node
> > > > > where your directory with the files is?
> > > >
> > > > [root at spduslisclust01 working]# rcp changecase.sh node1:/working
> > > > [root at spduslisclust01 working]# ssh node1 ls -l /working | grep
> > changecase
> > > > -rw-r--r--    1 root     root          821 Mar 16 16:10 changecase.sh
> > >
> > > Little misunderstanding: the other way round is in question. Login to a
> > node,
> > > and on the node issue the rcp to copy something from the master
> > > (spduslisclust01) to the local /tmp on the node for a test. - Reuti
> > >
> > > ---------------------------------------------------------------------
> > > 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
> >
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
> For additional commands, e-mail: users-help at gridengine.sunsource.net
> 
> 


-- 
"Knowledge is the only wealth that grows as you spend it, and
diminishes as you save it."
-- ancient Sanskrit saying

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