Atlas-connect-l mailing list

[David Lesny <ddl AT illinois.edu>]

Harinder,

The jobs which are going into hold state are the glidein jobs.
These jobs do not have a "user" attached to them at that moment.
The glidein is attempting to run a job on your node,
which in turn will create a slot into which a user job will run.

Since the glidein is going into the hold state on occasion
(but not all the time), I am suspecting something at your end.
But off hand I do not know what it might be.

I have submitted many jobs which cause the glideins to start,
but have not seen the problem, so whatever it is is just
transient or only on one particular node in your cluster.

When I get back from the LBNL meeting, I will see if I
can determine what might be causing this at your end.

Since the glidein cannot create a file in the home area,
I am guessing it is an NFS issue, but that is just a guess.

dave

On 8/18/2014 4:05 PM, Dr. Harinder Singh Bawa wrote:

Hello,

 I am looking for the best way to know the name of the user who submitted the job. Since you know that whatever the job that comes to any particular site e.g in Fresno, its labelled as Fresnoatlas. If I do:

[bawa@pt3wrk15 ~]$ condor_q -submitter fresnoatlas  or condor_q  -name t3head.atlas.csufresno.edu

It will show the list of jobs submitted by "fresnoatlas" since fresnoatlas is the queue name. It will show this for all the users who submitted jobs in "atlasconnect".

How to look for particular user name? Any suggestion?

Harinder

_______________________________________________
Atlas-connect-l mailing list
Atlas-connect-l AT lists.bnl.gov
https://lists.bnl.gov/mailman/listinfo/atlas-connect-l