[GE users] Problem with commd communications

Craig Tierney ctierney at hpti.com
Tue Jun 15 01:32:49 BST 2004


On Mon, 2004-06-14 at 18:27, Ron Chen wrote:
> For the first issue, we need some kind of timeout, and
> then basically do something similar to qstat to check
> for the existence of the job(s) that qevent is waiting
> for.

I hope that the check is a qstat 'lite'.  We just need
to use the job number, try and get a job-record, and if
it says 'no such job, quit'.  We shouldn't have to
call all of qstat, that puts overhead on the server
that we are trying to avoid.

> 
> I don't fully understand the 2nd issue, since I think
> we are supposed to get the info from qacct.

Since we are now getting information directly from
SGE instead of scripts calling qstat, there must
be an easy way to get the value before it gets to the
file.  The accounting logs are ascii.  The are big
and on my system slow.  Plus, for qacct to work you
have to have SGE installed in an NFS share directory,
which we don't.  When JOB_DEL is called, I figure that
the exit code has already been determined and that
must be around in memory still.

Craig



> 
>  -Ron
> 
> 
> --- Craig Tierney <ctierney at hpti.com> wrote:
> > I have added a few extra features to your patch to
> > qevent,
> > but there are two things I haven't figured out yet.
> > 
> > 1) How do you make sure that if you supply the wrong
> > job number
> > for a non-existent job that the code will not sit
> > indefinitely
> > but exits immedately?
> > 
> > 2) Does the event pass back enough information
> > through the
> > sgeE_JOB_DEL event so that I can get at the
> > exit_status of
> > the job (the one that gets written to the accounting
> > logs)?
> > 
> > Thanks,
> > Craig
> 
> 
> 
> 	
> 		
> __________________________________
> Do you Yahoo!?
> Friends.  Fun.  Try the all-new Yahoo! Messenger.
> http://messenger.yahoo.com/
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
> For additional commands, e-mail: users-help at gridengine.sunsource.net
> 


---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
For additional commands, e-mail: users-help at gridengine.sunsource.net




More information about the gridengine-users mailing list