[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] Condor not sending to other nodes



On 09/05/2011 01:47 PM, Matthew Farrellee wrote:
On 09/05/2011 12:12 PM, AL Olson wrote:
I have a cluster set up of 5 comps with 40 processors set up with ROCKS
5.4 and im trying to use
Condor to submit jobs to them but its only submitting them the first node
(compute-0-0). Ive done some research on Condor but im not sure if its
the
way i have ROCKS configured or if im not submitting it properly. I have
attached the condor submission script below. I can however ssh to all the
compute nodes if that means anything.

I would like to set my jobs so that i use processors on the head node as
well as processors on other nodes set up in the cluster. If this has
been discussed please point me to threads, im not sure what to search
for to find those. Thanks so much.


Universe = vanilla
Executable = /share/apps/2.1/cyana
Arguments = autorun.cya
Log = cyana.log
Output = autorun.out
Error = autorun.error
Queue

Chances are high you're being bitten by FILESYSTEM_DOMAIN and
should_transfer_files.

http://spinningmatt.wordpress.com/2011/07/04/getting-started-submittinag-jobs-to-condor/

http://spinningmatt.wordpress.com/2011/07/04/getting-started-submitting-jobs-to-condor/

^^ submittinag isn't a word, whoops

and,

should_transfer_files = if_needed
when_to_transfer_output = on_exit

Alternatively, you can setup FILESYSTEM_DOMAIN to be the same string
across nodes that have the same shared filesystem mounted.

BTW, condor_q -better should indicate FileSystemDomain as a problem for
the above to really be useful. 8o)

Best,


matt
_______________________________________________
Condor-users mailing list
To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/condor-users

The archives can be found at:
https://lists.cs.wisc.edu/archive/condor-users/