I'm afraid this is a kludgy fix that will break if/when the drmaa.py file gets updated. It'll work for now, but there has got to be a better way.
On 2/18/11 1:30 PM, Zhibin Lu wrote:
Hi, I had same problem since we do not install software in the system directory. I had to add PATH into ./lib/galaxy/jobs/runners/drmaa.py manually. I do not know if anyone has better solution. Zhibin Lu Bioinformatics Support Ontario Institute for Cancer Research MaRS Centre, South Tower 101 College Street, Suite 800 Toronto, Ontario, Canada M5G 0A3 www.oicr.on.ca > Date: Fri, 18 Feb 2011 13:17:26 -0500 > From: n...@bx.psu.edu > To: golha...@umdnj.edu > CC: firstname.lastname@example.org > Subject: Re: [galaxy-dev] cluster path question > > Ryan Golhar wrote: > > > > >Ryan Golhar wrote: > > >>I'm setting up Galaxy to run on my cluster using drmaa. I can see > > >>jobs getting submitted to Torque however the jobs keep failing, > > >>because the tools are in the default system path. > > >> > > >>How do I tell galaxy to append a directory to the path used by jobs? > > > > > >Hi Ryan, > > > > > >Whatever user is running the jobs will need to modify the $PATH in its > > >shell startup files. If you're using bash, this would probably be > > >~/.bash_profile > > > > It already is in .bashrc (which is called by .bash_profile) for the > > galaxy user. I can submit the script from the shell and it runs > > okay so something is up with either how the galaxy app is submitting > > the job or the environment being used by the galaxy job. > > When you say submit via the shell, do you mean with qsub, or by logging > in to a node and running the wrapper script? > > I haven't tested Torque with the drmaa runner so perhaps it's submitting > in a way which starts the shell in non-interactive, non-login mode. Let > me know if you can determine for certain that no shell startup files are > being read, and I'll try to replicate this in my environment. > > > I'd like to modify the job script to print out some debugging > > information to determine what environment variables are set when the > > job gets executed. Actually, this would be a useful "tool" within > > Galaxy to test the cluster. > > That's probably the simplest method, to just create a tool that outputs > (either to stdout or the output file) what you'd like to see. > > > > > > begin:vcard > > fn:Ryan Golhar, Ph.D. > > n:Golhar;Ryan > > org:The Cancer Institute of NJ;Cancer Informatics Core/Bioinformatics > > adr:5th floor;;120 Albany St;New Brunswick;NJ;08901;USA > > email;internet:golha...@umdnj.edu > > title:NGS Bioinformatics Specialist > > tel;work:(732) 235-6613 > > tel;fax:(732) 235-6267 > > tel;cell:(732) 236-1176 > > x-mozilla-html:FALSE > > url:http://www.cinj.org > > version:2.1 > > end:vcard > > > > > _______________________________________________ > > To manage your subscriptions to this and other Galaxy lists, please use the interface at: > > > > http://lists.bx.psu.edu/ > > _______________________________________________ > To manage your subscriptions to this and other Galaxy lists, please use the interface at: > > http://lists.bx.psu.edu/ _______________________________________________ To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/
--CONFIDENTIALITY NOTICE: This email communication may contain private, confidential, or legally privileged information intended for the sole use of the designated and/or duly authorized recipient(s). If you are not the intended recipient or have received this email in error, please notify the sender immediately by email and permanently delete all copies of this email including all attachments without reading them. If you are the intended recipient, secure the contents in a manner that conforms to all applicable state and/or federal requirements related to privacy and confidentiality of such information.
_______________________________________________ To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/