Re: [Condor-users] about condor-g to globus to condor problem


Date: Fri, 11 Feb 2005 16:24:43 -0500
From: Jaime Frey <jfrey@xxxxxxxxxxx>
Subject: Re: [Condor-users] about condor-g to globus to condor problem
On Feb 11, 2005, at 5:03 AM, Carson Hung wrote:

I have tried removing the globusrsl command, but after submission, the job
seems to state there without running. There is no special error msg for
checking GridmanagerLog.kshung file.

Does the job appear in the remote condor pool? If so, you should look there to see if the job isn't running because of an error or because it's taking a while to find an available resource. There is where condor_q -analyze will be useful.


Actually I have tried by sending another submit file like that (i.e. file6
is an executable file output by using condor_compile), the job is being
held.


executable = file6
universe = grid
grid_type=gt2
globusscheduler = pc-0242.eee.hku.hk/jobmanager-condor
globusRSL = (universe =  standard)
output = file6.out
log = file6.log
queue

The previous syntax you were using was close in this case. What you want is:


globusRSL = (condor_submit=(universe standard))

The following will also work for this one case (standard universe):

globusRSL = (jobtype=condor)

+----------------------------------+---------------------------------+
|            Jaime Frey            |  Public Split on Whether        |
|        jfrey@xxxxxxxxxxx         |  Bush Is a Divider              |
|  http://www.cs.wisc.edu/~jfrey/  |         -- CNN Scrolling Banner |
+----------------------------------+---------------------------------+


[← Prev in Thread] Current Thread [Next in Thread→]