Running galaxy in a cluster as if it was local
Hi, We're trying to setup a galaxy server, but where we want some of the tools to run on a cluster. This cluster is supporting mosix so at the moment we cannot use any of the drmaa stuff available (or can we?). The way we're setting it up at the moment is to run the command we want to run on the cluster through a remote mosix run. This works (almost) fine if we do it for specific command line tools (e.g. samtools). But now I've tried to run the whole python wrapper instead of only the command-line tool. In this case the problem we have is that the libraries on the end of the compute node are not set in the environment, since the environment is set only on the machine running the frontend. Is there any way I can set up the PYTHONPATH or other environment variables so that this would work in the compute node? Thanks, Daniel
Is there any way I can set up the PYTHONPATH or other environment variables so that this would work in the compute node?
Try using your galaxy_user's ~/.sge_request file. See http://wiki.g2.bx.psu.edu/Admin/Config/Tool%20Dependencies for more information. Good luck, J.
participants (2)
-
Daniel Sobral
-
Jeremy Goecks