Ok; it's quite weird. Perhaps a Galaxy guru could give you a better answer but I remembered that I had this kind of issues a while ago. In the meantime, you could take a look at these parameters:

$ grep retry galaxy.ini.sample 
# these instances, you can choose to retry setting it internally or leave it in
# a failed state (since retrying internally may cause the Galaxy process to be
# option to retry externally, or set metadata manually (when possible).
#retry_metadata_internally = True
# the job's stdout and stderr files when it completes, you can retry reading
# these files.  The job runner will retry the number of times specified below,
#retry_job_output_collection = 0


Best,

Remy 

2016-01-20 21:12 GMT+01:00 Eric Shell <eshell@soe.ucsc.edu>:
Thanks, Remy.  I went through the cluster documentation and our Rocks environment seems to be configured properly, after all.

It appears that my issue may be related to the UCSC Main table browser.  The jobs that Galaxy reports have failed are leaving the job_working_directory behind, with galaxy_#.e error files that contain "The remote data source application has not sent back a URL parameter in the request."

[root@campusrocks2 7045]# pwd
/campusdata/galaxy/galaxy/database/job_working_directory/007/7045
[root@campusrocks2 7045]# cat galaxy_7045.e
The remote data source application has not sent back a URL parameter in the request.

These errors correspond with empty dataset_#.dat files in /campusdata/galaxy/galaxy/database/files/011/:

[root@campusrocks2 7045]# ll /campusdata/galaxy/galaxy/database/files/011/
-rw-rw-r-- 1 galaxy galaxy        0 Jan 20 11:54 dataset_11387.dat

The job failures are intermittent.  Sometimes, a job requesting the exact same dataset will succeed moments before or after a failed job.  Is there perhaps a way to tell the table browser to retry when it fails to get the dataset it is requesting?  Is that even what's going on?

On Wed, Jan 20, 2016 at 7:05 AM, Rémy Dernat <remy.d1@gmail.com> wrote:
I forgot to point out the needs of sharing folders and checking the UID/GID of the galaxy user between your systems (and his access to SGE).

Remy

2016-01-20 16:00 GMT+01:00 Rémy Dernat <remy.d1@gmail.com>:
Hi Eric,

Here we use both solutions: Galaxy and RocksCluster. In Galaxy, you have to define your jobs in "config/job_conf.xml" and you should probably source a file (search for "environment" in your galaxy.ini) before the submit process. In fact, you could have to set a DRMAA_LIBRARY_PATH to load your drmaa library; see https://wiki.galaxyproject.org/Admin/Config/Performance/Cluster#DRMAA

Best,
Remy

2016-01-19 20:28 GMT+01:00 Eric Shell <eshell@soe.ucsc.edu>:
I am trying to get a Galaxy instance running on a Rocks cluster.  I am able to run jobs with the local runner at this point, but I am having an issue with the drmaa runner that I haven't been able to fix.  When I submit a job in Galaxy it is successfully submitted to the cluster and runs to completion according to qacct, but Galaxy just reports "failure running job".

Here's what is written to paster.log when I submit a job:

69.181.235.240 - - [19/Jan/2016:11:24:31 -0700] "GET /api/histories/fb86c918c0d3d33b/contents?dataset_details=bae154fe2294752e%2C6fe732485990d2ac%2C604c4e6e60e997bc%2Cf015f1cb819ec50e%2C9f6f4b3cb6cf43eb%2C3d13d598882b6eb8%2C551006fddcb290ae%2C10b9bbc646c48387%2C7670dfdf35146bc5%2Ce0ec2cf59f1fc79e%2Cee30922e5e4854db%2C9e7a0ba216194210 HTTP/1.1" 200 - "https://galaxy.soe.ucsc.edu/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/47.0.2526.111 Safari/537.36"
69.181.235.240 - - [19/Jan/2016:11:24:38 -0700] "GET /tool_runner/data_source_redirect?tool_id=ucsc_table_direct1 HTTP/1.1" 302 - "https://galaxy.soe.ucsc.edu/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/47.0.2526.111 Safari/537.36"
galaxy.tools.actions.__init__ INFO 2016-01-19 11:24:42,801 Handled output (327.778 ms)
galaxy.tools.actions.__init__ INFO 2016-01-19 11:24:43,236 Verified access to datasets (0.023 ms)
galaxy.tools.execute DEBUG 2016-01-19 11:24:43,343 Tool [ucsc_table_direct1] created job [7019] (919.481 ms)
69.181.235.240 - - [19/Jan/2016:11:24:42 -0700] "POST /tool_runner HTTP/1.1" 200 - "https://genome.ucsc.edu/cgi-bin/hgTables" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/47.0.2526.111 Safari/537.36"
galaxy.jobs DEBUG 2016-01-19 11:24:44,056 (7019) Working directory for job is: /campusdata/galaxy/galaxy/database/job_working_directory/007/7019
galaxy.jobs.handler DEBUG 2016-01-19 11:24:44,070 (7019) Dispatching to sge runner
galaxy.jobs DEBUG 2016-01-19 11:24:44,378 (7019) Persisting job destination (destination id: sge_default)
galaxy.jobs.runners DEBUG 2016-01-19 11:24:44,403 Job [7019] queued (332.423 ms)
galaxy.jobs.handler INFO 2016-01-19 11:24:44,444 (7019) Job dispatched
69.181.235.240 - - [19/Jan/2016:11:24:44 -0700] "GET /api/genomes HTTP/1.1" 200 - "https://galaxy.soe.ucsc.edu/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/47.0.2526.111 Safari/537.36"
69.181.235.240 - - [19/Jan/2016:11:24:44 -0700] "GET /api/datatypes?extension_only=False& HTTP/1.1" 200 - "https://galaxy.soe.ucsc.edu/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/47.0.2526.111 Safari/537.36"
69.181.235.240 - - [19/Jan/2016:11:24:44 -0700] "GET /history/current_history_json HTTP/1.1" 200 - "https://galaxy.soe.ucsc.edu/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/47.0.2526.111 Safari/537.36"
galaxy.jobs.runners DEBUG 2016-01-19 11:24:46,399 (7019) command is: python /campusdata/galaxy/galaxy/tools/data_source/data_source.py /campusdata/galaxy/galaxy/database/files/011/dataset_11361.dat 0; return_code=$?; python "/campusdata/galaxy/galaxy/database/job_working_directory/007/7019/set_metadata_IaPURP.py" "/campusdata/galaxy/galaxy/database/tmp/tmp9Qt0cv" "/campusdata/galaxy/galaxy/database/job_working_directory/007/7019/galaxy.json" "/campusdata/galaxy/galaxy/database/job_working_directory/007/7019/metadata_in_HistoryDatasetAssociation_13512_oucw5s,/campusdata/galaxy/galaxy/database/job_working_directory/007/7019/metadata_kwds_HistoryDatasetAssociation_13512_ZrUbrF,/campusdata/galaxy/galaxy/database/job_working_directory/007/7019/metadata_out_HistoryDatasetAssociation_13512_twCvq7,/campusdata/galaxy/galaxy/database/job_working_directory/007/7019/metadata_results_HistoryDatasetAssociation_13512_FO1cy9,/campusdata/galaxy/galaxy/database/files/011/dataset_11361.dat,/campusdata/galaxy/galaxy/database/job_working_directory/007/7019/metadata_override_HistoryDatasetAssociation_13512_Z_cUTF" 5242880; sh -c "exit $return_code"
galaxy.jobs.runners.drmaa DEBUG 2016-01-19 11:24:46,787 (7019) submitting file /campusdata/galaxy/galaxy/database/job_working_directory/007/7019/galaxy_7019.sh
galaxy.jobs.runners.drmaa DEBUG 2016-01-19 11:24:46,808 (7019) native specification is: -R y -pe mpi 8 -q small.q
galaxy.jobs DEBUG 2016-01-19 11:24:46,828 (7019) Changing ownership of working directory with: /usr/bin/sudo -E scripts/external_chown_script.py /campusdata/galaxy/galaxy/database/job_working_directory/007/7019 eshell 100000
galaxy.jobs.runners.drmaa DEBUG 2016-01-19 11:24:47,020 (7019) submitting with credentials: eshell [uid: 38559]
galaxy.jobs.runners.drmaa DEBUG 2016-01-19 11:24:47,129 (7019) Job script for external submission is: /campusdata/galaxy/galaxy/database/pbs/7019.jt_json
galaxy.jobs.runners.drmaa INFO 2016-01-19 11:24:47,130 Running command ['/usr/bin/sudo', '-E', 'scripts/drmaa_external_runner.py', '38559', '/campusdata/galaxy/galaxy/database/pbs/7019.jt_json']
galaxy.jobs.runners.drmaa INFO 2016-01-19 11:24:47,981 (7019) queued as 116563
galaxy.jobs DEBUG 2016-01-19 11:24:48,198 (7019) Persisting job destination (destination id: sge_default)
galaxy.jobs.runners.drmaa DEBUG 2016-01-19 11:24:48,823 (7019/116563) state change: job is queued and active
69.181.235.240 - - [19/Jan/2016:11:24:45 -0700] "GET /api/histories/fb86c918c0d3d33b/contents?dataset_details=bae154fe2294752e%2C6fe732485990d2ac%2C604c4e6e60e997bc%2Cf015f1cb819ec50e%2C9f6f4b3cb6cf43eb%2C3d13d598882b6eb8%2C551006fddcb290ae%2C10b9bbc646c48387%2C7670dfdf35146bc5%2Ce0ec2cf59f1fc79e%2Cee30922e5e4854db%2C9e7a0ba216194210 HTTP/1.1" 200 - "https://galaxy.soe.ucsc.edu/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/47.0.2526.111 Safari/537.36"
galaxy.jobs.runners.drmaa DEBUG 2016-01-19 11:24:53,532 (7019/116563) state change: job is running
galaxy.jobs WARNING 2016-01-19 11:24:53,922 (7019) Ignoring state change from 'error' to 'running' for job that is already terminal
69.181.235.240 - - [19/Jan/2016:11:24:54 -0700] "GET /api/histories/fb86c918c0d3d33b/contents HTTP/1.1" 200 - "https://galaxy.soe.ucsc.edu/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/47.0.2526.111 Safari/537.36"
69.181.235.240 - - [19/Jan/2016:11:24:55 -0700] "GET /api/histories/fb86c918c0d3d33b HTTP/1.1" 200 - "https://galaxy.soe.ucsc.edu/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/47.0.2526.111 Safari/537.36"
galaxy.jobs.runners.drmaa INFO 2016-01-19 11:25:06,240 (7019/116563) job left DRM queue with following message: code 18: The job specified by the 'jobid' does not exist.
galaxy.jobs DEBUG 2016-01-19 11:25:06,412 (7019) Changing ownership of working directory with: /usr/bin/sudo -E scripts/external_chown_script.py /campusdata/galaxy/galaxy/database/job_working_directory/007/7019 galaxy 59997
galaxy.jobs DEBUG 2016-01-19 11:25:06,622 (7019) Changing ownership of working directory with: /usr/bin/sudo -E scripts/external_chown_script.py /campusdata/galaxy/galaxy/database/job_working_directory/007/7019 galaxy 59997

'qacct -j -o eshell' shows that job 7019 completed, though:

qname        all.q
hostname     campusrocks2-0-4.local
group        users
owner        eshell
project      NONE
department   defaultdepartment
jobname      g7019_ucsc_table_direct1_eshell_ucsc_edu
jobnumber    116563
taskid       undefined
account      sge
priority     0
qsub_time    Tue Jan 19 11:24:47 2016
start_time   Tue Jan 19 11:24:53 2016
end_time     Tue Jan 19 11:25:05 2016
granted_pe   mpi
slots        8
failed       0
exit_status  0
ru_wallclock 12
ru_utime     9.285
ru_stime     0.908
ru_maxrss    98384
ru_ixrss     0
ru_ismrss    0
ru_idrss     0
ru_isrss     0
ru_minflt    81778
ru_majflt    2
ru_nswap     0
ru_inblock   6728
ru_oublock   184
ru_msgsnd    0
ru_msgrcv    0
ru_nsignals  0
ru_nvcsw     13952
ru_nivcsw    301
cpu          10.192
mem          2.326
io           0.150
iow          0.000
maxvmem      448.820M
arid         undefined

Why does Galaxy not see the job after it has been submitted to the cluster?

Thanks in advance for your help!

--
Eric Shell
UNIX Software & Google Apps Administrator
Baskin School of Engineering
UC Santa Cruz
831 459 4919

___________________________________________________________
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/





--
Eric Shell
UNIX Software & Google Apps Administrator
Baskin School of Engineering
UC Santa Cruz
831 459 4919