Sorry for bumping this, but it seems like such a standard boilerplate step of installing a tool that someone muts know how.
-Joel
I installed the?bedtools tool from the public toolshed to my galaxy cluster. The jobs were failing because I hadn't installed the required binaries:
? ?An error occurred running this job: /opt/sge/default/spool/execd/ip-10-194-50-118/job_scripts/2: line 13: genomeCoverageBed: command not found
Taking a closer look at the toolshed's page, I noticed there were a handful of required binaries (genomeCoverageBed,?intersectBed, etc).
So I downloaded BEDTools and compiled all the binaries. I've put them in well organized directories using symlinks to specific versions, etc. The final executable directory containing symlinks to these binaries is "/mnt/galaxyTools/shed_tool_binaries/bin/".
I can make a snapshot of the tools volume so that these compiled binaries are always available to me when I bring up my cluster, but how do I integrate them into the PATH in a way that lets the bedtools galaxy tool see them, but also survive cluster shutdowns?
Thanks,
-Joel
On Mon, Nov 5, 2012 at 5:34 PM, Joel Rosenberg thisisjoel@hotmail.com wrote:
Sorry for bumping this, but it seems like such a standard boilerplate step of installing a tool that someone muts know how.
-Joel
Personally I've just been installing them on the $PATH, typically under $HOME/bin of the Galaxy user which is on our Galaxy user account's $PATH.
Peter
If you want to keep your tools separated by name and version (recommended) see the section "Managed Tool Dependencies" at http://wiki.g2.bx.psu.edu/Admin/Config/Tool%20Dependencies
-- James Taylor, Assistant Professor, Biology/CS, Emory University
On Mon, Nov 5, 2012 at 12:44 PM, Peter Cock p.j.a.cock@googlemail.com wrote:
On Mon, Nov 5, 2012 at 5:34 PM, Joel Rosenberg thisisjoel@hotmail.com wrote:
Sorry for bumping this, but it seems like such a standard boilerplate step of installing a tool that someone muts know how.
-Joel
Personally I've just been installing them on the $PATH, typically under $HOME/bin of the Galaxy user which is on our Galaxy user account's $PATH.
Peter ___________________________________________________________ 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:
Thanks Peter,
I tried doing something similar, but there wasn't a location in the galaxy user's $PATH that would persist across a cluster reboot or could be snapshotted (like the tools or data volumes). Adding /mnt/galaxyTools/my_dep_binaries or something to the $PATH would work temporarily, but not past a reboot as /home/galaxy/.bashrc, /etc/environment, etc are part of the AMI and would be lost.
In your cluster, wouldn't the galaxy user's ${HOME}/bin be transient? Are you running with a different EC2 setup or using a different AMI?
-Joel
----------------------------------------
Date: Mon, 5 Nov 2012 17:44:49 +0000 Subject: Re: [galaxy-dev] Where should I put a tool's required binaries? From: p.j.a.cock@googlemail.com To: thisisjoel@hotmail.com CC: galaxy-dev@lists.bx.psu.edu
On Mon, Nov 5, 2012 at 5:34 PM, Joel Rosenberg thisisjoel@hotmail.com wrote:
Sorry for bumping this, but it seems like such a standard boilerplate step of installing a tool that someone muts know how.
-Joel
Personally I've just been installing them on the $PATH, typically under $HOME/bin of the Galaxy user which is on our Galaxy user account's $PATH.
Peter
On Mon, Nov 5, 2012 at 6:09 PM, Joel Rosenberg thisisjoel@hotmail.com wrote:
Thanks Peter,
I tried doing something similar, but there wasn't a location in the galaxy user's $PATH that would persist across a cluster reboot or could be snapshotted (like the tools or data volumes). Adding /mnt/galaxyTools/my_dep_binaries or something to the $PATH would work temporarily, but not past a reboot as /home/galaxy/.bashrc, /etc/environment, etc are part of the AMI and would be lost.
In your cluster, wouldn't the galaxy user's ${HOME}/bin be transient? Are you running with a different EC2 setup or using a different AMI?
We're running a local cluster with SGE, and that means our $HOME/bin is stable. I hadn't realised this you mentioned EC2/AMI that you might be on a cloud cluster.
Peter
Thanks, James. That page is exactly what I was looking for.
galaxy-dev@lists.galaxyproject.org
-
James Taylor -
Joel Rosenberg -
Peter Cock