On Tue, Aug 16, 2011 at 12:38 PM, Nate Coraor
<nate@bx.psu.edu> wrote:
Hi Luobin,
Sorry for the delay in response. I would suggest moving this discussion
to the galaxy-dev mailing list since it does not contain any private
data. The wide audience on galaxy-dev may be able to come up with
additional ideas than what we on the Galaxy Team come up with.
Please see my responses inline below.
Luobin Yang wrote:
> Hi,
>
> Thanks for Martin Dahlo's excellent blog (
>
http://mdahlo.blogspot.com/2011/06/galaxy-on-uppmax.html) on making SLURM
> work with GALAXY, I am able to run galaxy jobs on a cluster, but I've got a
> couple of issues:
>
> 1. Sometimes a job is in waiting state and it won't start to run until I
> restart galaxy.
(PBS, SGE, something else?)?
When this happens, the job is stuck in Galaxy, it doesn't submit the job to SLURM queueing system.
In your config file (universe_wsgi.ini), is track_jobs_in_database =
True or False? Are you using SQLite or another database? If you watch
I am using PostgreSQL database system.
track_jobs_in_database is False in my config file. Which option is better when I have a queueing system?
the Galaxy log, are any exceptions or other errors logged when a job
becomes stuck?
Galaxy log doesn't show any exception or errors.
> 2. Sometimes a job is in running state even though it is already finished
> and restarting galaxy can make the job's state change from running to
> finished.
Most likely, Galaxy is setting metadata on the job's outputs. You can
probably speed up this process by setting:
set_metadata_externally = True
in the config file.
Thanks,
--nate
>
> I am wondering what's causing those issues and how they can be fixed.
>
> Thanks,
> Luobin
>