[GE users] Running SGE jobs sequentially

Richard Bohn rxbeee at rit.edu
Thu Feb 22 16:43:07 GMT 2007


Thanks Chris!

I looked through some of the docs but I evidently didn't connect all the
dots to get the big picture.

Rick

-----Original Message-----
From: Chris Dagdigian [mailto:dag at sonsorol.org] 
Sent: Thursday, February 22, 2007 11:31 AM
To: users at gridengine.sunsource.net
Subject: Re: [GE users] Running SGE jobs sequentially

Hi Rick,

You can name a job with a special pattern that would make sense to  
your workflow using the "-N" switch at the command line or embedded  
into a job script:

  #$ -N job001

... then you can submit a follow on job like so:

  $  qsub -hold_jid job001  ./job002_script.sh

In this case, job002_script.sh would hold in the pending queue in  
state "qw" until job001 exits.

Basically look into the SGE docs for information about job naming,  
"job dependencies" and the hold_jid option.

Regards,
Chris



On Feb 22, 2007, at 11:27 AM, Richard Bohn wrote:

> Hello,
>
>
>
> I have a user that needs to run a number of jobs sequentially  
> rather than concurrently what is the best way to handle this in  
> SGE? They want to use all the available compute nodes for each job  
> before moving on the next one.
>
>
>
> They could write a script that submits the first job and then  
> checks for it to finish before doing the next qsub but this seems a  
> little klugy.
>
>
>
> Thanks in advance for any help.
>
>
>
> Rick (SGE newbie)
>
>
>
>

---------------------------------------------------------------------
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