[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Condor-users] condor & MPI part 3



Fixed earlier issues with config file syntax when trying to run MPI jobs under condor. Now, my jobs get scheduled and try to run, but I get the following error in the StarterLog on the execute machine:

1/26 10:39:17 ******************************************************
1/26 10:39:17 ** condor_starter (CONDOR_STARTER) STARTING UP
1/26 10:39:17 ** /astro/net/condor/sbin/condor_starter
1/26 10:39:17 ** $CondorVersion: 6.6.7 Oct 11 2004 $
1/26 10:39:17 ** $CondorPlatform: I386-LINUX_RH9 $
1/26 10:39:17 ** PID = 26527
1/26 10:39:17 ******************************************************
1/26 10:39:17 Using config file: /users/condor/condor_config
1/26 10:39:17 Using local config files: /net/condor/etc/astrolab18.local
1/26 10:39:17 DaemonCore: Command Socket at <128.95.99.182:54145>
1/26 10:39:17 Done setting resource limits
1/26 10:39:17 Starter communicating with condor_shadow <128.95.98.82:54801>
1/26 10:39:17 Submitting machine is "carrion98.astro.washington.edu"
1/26 10:39:17 Starting a MPI universe job with ID: 3685.0
1/26 10:39:17 Can't find MPI_CONDOR_RSH_PATH in config file! Aborting!
1/26 10:39:17 Failed to start job, exiting
1/26 10:39:17 ShutdownFast all jobs.
1/26 10:39:17 **** condor_starter (condor_STARTER) EXITING WITH STATUS 0


I don't know what the MPI_CONDOR_RSH_PATH variable is, or what it should be set to... thanks for any/all suggestions!


-----------------------------------------
Rok Roskar
University of Washington
Department of Astronomy