[condor-users] Globus Universe Job not running

[Hari Pulapaka <pulapaka@xxxxxxxxxxxxxxxx>]

Hi,

I am running a test job under globus universe and I find that it is 
sitting idle in the queue.

bash-2.05a$ cat test.sub
universe = globus
globusscheduler = ldas-grid.ligo.caltech.edu/jobmanager-condor
executable = /bin/hostname
output = host.$(Process).out
error = host.$(Process).err
log = host.$(Process).log
transfer_executable = false
queue

Inside the SchedLog I see the following message. Looks like it exited 
with error code 256?

[root@ldas-grid log]# tail SchedLog
12/23 10:46:27 Send_Signal: ERROR Connect to <10.14.0.12:56590> 
failed.12/23 10:46:27 condor_write(): Socket closed when trying to write 
buffer
12/23 10:46:27 Buf::write(): condor_write() failed
12/23 10:46:27 SECMAN: Error sending response classad!
12/23 10:46:27 condor_gridmanager exited pid=1782 status=256 
owner=lsc-cit/DC=org/DC=doegrids/OU=People/CN=Hari_Pulapaka_561239/CN=proxy/CN=proxy
12/23 10:47:47 DaemonCore: Command received via TCP from host 
<10.14.0.12:56616>
12/23 10:47:47 DaemonCore: received command 421 (RESCHEDULE), calling 
handler (reschedule_negotiator)
12/23 10:47:47 Sent ad to central manager for lsc-cit@ldas-cit
12/23 10:47:47 Started condor_gmanager for owner lsc-cit pid=1803
12/23 10:47:47 Called reschedule_negotiator()
12/23 10:47:47 condor_gridmanager exited pid=1803 status=256 
owner=lsc-cit/DC=org/DC=doegrids/OU=People/CN=Hari_Pulapaka_561239/CN=proxy/CN=proxy

Any help would be appreciated.

I am running condor version 6.5.5 which was provided along with VDT 1.1.10.

I remember seeing a message earlier where there was a bug in the grid 
manager, but I dont remember exactly what it was, I wonder if this is 
related?

Thanks,

Hari



Condor Support Information:
http://www.cs.wisc.edu/condor/condor-support/
To Unsubscribe, send mail to majordomo@xxxxxxxxxxx with
unsubscribe condor-users <your_email_address>