Hi all,
We've been running a galaxy server for the last few years successfully, using pbs_python for connection to our HPC. Recently, we've purchased a new system, and we're trying to install galaxy on it. However, we are having some issues.
When we try to submit jobs, galaxy returns the following error:
galaxy.jobs.runners.pbs WARNING 2017-02-15 16:55:22,178 (20) pbs_submit failed (try 1/5), PBS error 15025: Bad UID for job execution
On the pbs_server, we see:
02/16/2017 09:14:43.631;128;PBS_Server.7156;Req;req_reject;Reject reply code=15025(Bad UID for job execution MSG=ruserok failed validating galaxyuser/galaxyuser from galaxy.uza.be), aux=0, type=QueueJob, from galaxyuser@galaxy.uza.be
However, when we submit the generated qsub script manually, using the same user, queue and pbs_account (-A) settings, the job runs just fine. In our opinion this indicats an issue with galaxy somewhere. Running jobs locally also works.
The docs state that we need to scramble a specific pbs_python version for our recent pbs version, but the mentioned files are missing in galaxy 16_10.
Some information: =============
pbs_server version : 6.02
galaxy: v16_10
deprecated procedures: https://wiki.galaxyproject.org/Admin/Config/Performance/Cluster (in PBS : egss.ini and scripts/scramble.py are missing)
Has anybody seen this before? Are there newer docs available on pbs_python ?
Best,
Geert