Assuming grid engine is working outside of Galaxy and the Galaxy user
can say qsub files just fine - then you need to install the sun grid
engine libdrmaa package - or build from scratch. Assuming you have
done that and the resulting .so file is
/usr/lib/gridengine-drmaa/lib/libdrmaa.so. Then all you should have to
do is create a config/job_conf.xml file in your Galaxy root that looks
something like this:
<?xml version="1.0"?>
<job_conf>
<plugins workers="2">
<plugin id="gridengine" type="runner"
load="galaxy.jobs.runners.drmaa:DRMAAJobRunner">
<param
id="drmaa_library_path">/usr/lib/gridengine-drmaa/lib/libdrmaa.so</param>
</plugin>
</plugins>
<handlers default="main">
<handler id="main" />
</handlers>
<destinations default="cluster">
<destination id="cluster" runner="gridengine"/>
</destinations>
<tools>
</tools>
<limits>
<!--
<limit type="registered_user_concurrent_jobs">2</limit>
<limit type="unregistered_user_concurrent_jobs">1</limit>
<limit type="job_walltime">24:00:00</limit>
-->
</limits>
</job_conf>
A slightly more complicated setup with supervisor, uwsgi, multiple
handlers, etc... all on grid engine is covered in this tutorial that
Nate put together last year
(https://wiki.galaxyproject.org/Events/GCC2014/TrainingDay/AdminWalkthrough#Configure_Galaxy).
This job_conf.xml is a stripped down version of the one from that
tutorial.
In terms of rolling back and just getting jobs to start running again
- you need to remove $GALAXY_ROOT/config/job_conf.xml and
$GALAXY_ROOT/job_conf.xml and things should start working using the
local runner again. If this isn't working - there maybe a problem in
your galaxy.ini file - I would need to know what properties were being
set in there to know more though.
-John
> ___________________________________________________________
On Mon, Feb 2, 2015 at 12:53 PM, Ryan G <ngsbioinformatics@gmail.com> wrote:
> I tried setting up Galaxy to use Sun Grid Engine, but failed miserably. I
> removed my config/job_conf.xml hoping Galaxy would revert to its default
> state, but jobs are no longer being dispatched. I don't see any errors in
> the log.
>
> How can I get Galaxy to 1) use SGE, or 2) revert to its default state for
> running jobs. The page,
> https://wiki.galaxyproject.org/Admin/Config/Performance/Cluster isn't 100%
> clear on how to set up a simply configuration for SGE. I'd be happy to
> contribute mine as an example, if I can get it running.
>
>
> 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:
> https://lists.galaxyproject.org/
>
> To search Galaxy mailing lists use the unified search at:
> http://galaxyproject.org/search/mailinglists/