David,
For instances that run jobs under real user ids I think the job_name variable
in drmaa.py does not need the username component in this case anymore as the
batch system takes care of that and the resulting job name looks cleaner in the
qstat output. I use
@@ -184,10 +183,12 @@
ofile = %s.drmout % os.path.join(job_wrapper.working_directory,
job_wrapper.get_id_tag())
efile = %s.drmerr % os.path.join(job_wrapper.working_directory,
job_wrapper.get_id_tag())
ecfile = %s.drmec % os.path.join(job_wrapper.working_directory,
job_wrapper.get_id_tag())
-job_name = g%s_%s_%s % ( job_wrapper.job_id, job_wrapper.tool.id,
job_wrapper.user )
+job_name = g%s_%s % ( job_wrapper.job_id, job_wrapper.tool.id )
The above could be made conditional depending on whether the instance runs jobs
under real user ids or the galaxy user if Nate decides to put something like
this into the upstream codebase.
Regards,
Alex
On Oct 3, 2012, at 11:51 AM, Nate Coraor n...@bx.psu.edu wrote:
On Oct 2, 2012, at 3:56 PM, David Hoover wrote:
When I use the real user method for running drmaa jobs as the real user, the
jobName value gets lost. It's kind of annoying, but not a show stopper.
The other values (outputPath, errorPath, remoteCommand, nativeSpecification,
etc.) are all there, but not jobName.
This doesn't happen when the drmaa jobs are run using the default Galaxy
user. Weird.
Hi David,
This has been fixed in changeset 7823:62fbfc687692. Thanks for letting us
know.
--nate
David Hoover
Helix Systems Staff
___
Please keep all replies on the list by using reply all
in your mail client. To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
http://lists.bx.psu.edu/
___
Please keep all replies on the list by using reply all
in your mail client. To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
http://lists.bx.psu.edu/
___
Please keep all replies on the list by using reply all
in your mail client. To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
http://lists.bx.psu.edu/
