[torqueusers] Dependencies

Garrick Staples garrick at clusterresources.com
Mon Dec 11 18:16:40 MST 2006


On Mon, Dec 11, 2006 at 04:56:01PM -0800, Sam Rash alleged:
> There is some way to submit forward dependencies, too.  It's more or less
> the same thing.
> 
> It would be quite nice if the system dealt with jobs completing and
> submitting dependencies on them.  Sure it's not too hard to wrap the proper
> submission functionality, but if it's such a common problem, why not make
> torque's dependencies robust enough to do it?  

I spent a few days trying to solve the problem with 'Completed' jobs.
Turns out it is pretty hard.  I'll have to revisit it at some later
date.


> 
> -sr
> 
> 
> Sam Rash
> srash at yahoo-inc.com
> 408-349-7312
> vertigosr37
> 
> -----Original Message-----
> From: torqueusers-bounces at supercluster.org
> [mailto:torqueusers-bounces at supercluster.org] On Behalf Of Ronny T. Lampert
> Sent: Saturday, December 09, 2006 9:42 AM
> To: Krause, Roland; torqueusers at supercluster.org
> Subject: Re: [torqueusers] Dependencies
> 
> > in case step1 is already gone, step2 is on hold forever. Can I change
> > this behaviour? Can I force
> > torque to keep information about finnished jobs for a certain amount of
> > time?
> 
> I thought in that case, step2 will be killed because the dependency cannot 
> be satisfied?
> Out of the box afaik this can't be done. You would end up with hacking the 
> source. BUT!
> 
> What we do is: submit the first job in the dependency chain "on hold", then 
> submit all following, then un-hold the jobs, and then be happy :)
> 
> Cheers,
> Ronny
> 
> _______________________________________________
> torqueusers mailing list
> torqueusers at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torqueusers
> 
> 
> _______________________________________________
> torqueusers mailing list
> torqueusers at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torqueusers


More information about the torqueusers mailing list