I am getting similar errors as Brian reported back in March. (Note, we
appear to have the same last name, but no relation)

 An error occurred with this dataset: *Job output not returned from cluster*
*
*
- Running on Cloudman with 5-6 nodes. (xlarge)
- The error seems to occur consistently when I launch multiple workflows in
batch (using bioblend)
- Probably not relevant, but is failing on a BWA step.
- I am able to run successfully the same workflow against one of the
datasets that failed in batch.
- Change-set is from Feb 8, 2013. 8794:1c7174911392. Stable branch. Prior
to that, I was running different galaxy instances using changesets from
last year and never ran into this problem.
- I'm seeing errors like:
galaxy.jobs.runners.drmaa WARNING 2013-05-02 17:07:51,991 Job output not
returned from cluster: [Errno 2] No such file or directory\
: '/mnt/galaxyData/tmp/job_working_directory/002/2066/2066.drmec'
- In this example, the /mnt/galaxyData/tmp/job_working_directory/002/2066
folder and /mnt/galaxyData/tmp/job_working_directory/002/2066/2066.drmec
files do not exist.


Any suggestions? Seems like this might be some type of resource contention
issue, but I'm not sure where to investigate next.

Thanks in advance,
Dave



On Mon, Mar 11, 2013 at 9:04 AM, Brian Lin <brian....@tufts.edu> wrote:

> Hi guys, I'm running a galaxy cloudman instance and running the usual
> tophat->cufflinks->cuffdiff workflow from RNAseq data.
> I am using a m2.4xlarge as a master node, and autoscaling from 0-4 workers
> of the m2.xlarge type.
> I have gotten the error: An error occurred running this job: *Job output
> not returned from cluster*
> when running fasta groomer, tophat, and now cufflinks.
> Following up troubleshooting from other people in the mailing list, I have
> set a new line in universe_wsgi.ini of retry_job_output_collection=30
> Unfortunately, this does not seem to have fixed the problem.
> The stdout is blank, and stderr gives Job output not returned from cluster
>
> Under manage jobs in the admin panel, it lists 4 out of the 6 jobs as
> currently running. What is confusing is that of the 4 "running," one has
> already returned the error in the user dataset panel and yet is still
> listed as running.
>
> From the SGE log, I see these errors:
>
>
> 03/11/2013 14:32:52|worker|ip-10-159-47-223|W|job 42.1 failed on host 
> ip-10-30-130-84.ec2.internal before writing exit_status because: shepherd 
> exited with exit status 19: before writing exit_status
> 03/11/2013 14:32:52|worker|ip-10-159-47-223|W|job 43.1 failed on host 
> ip-10-30-130-84.ec2.internal before writing exit_status because: shepherd 
> exited with exit status 19: before writing exit_status
> 03/11/2013 14:39:41|worker|ip-10-159-47-223|E|adminhost 
> "ip-10-30-130-84.ec2.internal" already exists
> 03/11/2013 14:40:07|worker|ip-10-159-47-223|E|exechost 
> "ip-10-30-130-84.ec2.internal" already exists
> 03/11/2013 14:50:54|worker|ip-10-159-47-223|E|adminhost 
> "ip-10-30-130-84.ec2.internal" already exists
> 03/11/2013 14:50:55|worker|ip-10-159-47-223|E|exechost 
> "ip-10-30-130-84.ec2.internal" already exists
>
>
> Does anyone have any idea how to solve this error? It has removed my
> ability to use workflows completely and I still have not been able to run a
> single analysis to completion due to it.
>
> Thanks for any insight anyone provide!
>
> Brian
> --
> Brian Lin
> cont...@brian-lin.com
> brian....@tufts.edu
>
> ___________________________________________________________
> The Galaxy User list should be used for the discussion of
> Galaxy analysis and other features on the public server
> at usegalaxy.org.  Please keep all replies on the list by
> using "reply all" in your mail client.  For discussion of
> local Galaxy instances and the Galaxy source code, please
> use the Galaxy Development list:
>
>   http://lists.bx.psu.edu/listinfo/galaxy-dev
>
> 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/

To search Galaxy mailing lists use the unified search at:
  http://galaxyproject.org/search/mailinglists/

Reply via email to