oodt-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mattmann, Chris A (3980)" <chris.a.mattm...@jpl.nasa.gov>
Subject Re: OODT and Torque
Date Tue, 17 Jun 2014 05:23:46 GMT
Hi Jeff,

Sorry for the delay in replying - yes a post condition wait is
certainly one way to solve this. Another method would be to create
a Resource manger BatchMgr plugin that submits Jobs to Torque and
then a Torque/QSub monitor to monitor the job status.

This would keep the WM-RM binding that already exists including
the interface, and isolate the plugin to just the Resource manager.

Let me know what you think, CC'ing the dev list where others besides
me can respond more actively :)

Cheers,
Chris


++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Chris Mattmann, Ph.D.
Chief Architect
Instrument Software and Science Data Systems Section (398)
NASA Jet Propulsion Laboratory Pasadena, CA 91109 USA
Office: 168-519, Mailstop: 168-527
Email: chris.a.mattmann@nasa.gov
WWW:  http://sunset.usc.edu/~mattmann/
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Adjunct Associate Professor, Computer Science Department
University of Southern California, Los Angeles, CA 90089 USA
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++






-----Original Message-----
From: Jeff Kern <jkern@nrao.edu>
Date: Tuesday, June 3, 2014 2:26 PM
To: Chris Mattmann <Chris.A.Mattmann@jpl.nasa.gov>
Cc: Lindsey Davis <ldavis@nrao.edu>, 'Rick Lively' <rlively@nrao.edu>
Subject: OODT and Torque

>Hi Chris,
>     We're back to working with OODT and integrating it with Torque for
>the production pipeline.  I'm just writing to make sure we are not
>re-inventing the wheel.  The issue we have is that Torque is a batch
>system, so basically when we want to run one of the PGE's on the cluster
>we build the command and then execute a queue submission, which returns
>immediately.  From OODT's perspective it thinks the PGE has completed
>and moves to the next stage in the workflow.
>     What we've discussed doing is to use the post-conditional, so that
>after the job has been submitted we block on the conditional waiting for
>a message before moving on.
>     Has this problem already been solved in the workflow manager? Is
>there a better way to implement this?
>     Thanks,
>         Jeff
>
>-- 
>-------------------------------------
>Dr. Jeffrey S. Kern
>CASA Group Lead
>ALMA Software North American Deputy
>National Radio Astronomy Observatory (NRAO)
>Ph: (575) 835-7391
>E-mail: jkern@nrao.edu
>-------------------------------------
>


Mime
View raw message