[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] Flocking drawback



I am pretty sure that the requirement expression is correct since if his
jobs had really started they would return an error since they can access the
input files. But that doesn't happen, they just hang around until a machine
in pool A is free.
As to the other possibility (another job) it is impossible since there were
no other jobs in the queues at the time, and anyway condor_q -global -run
showed me that his jobs were "starting" on the machines that I saw being
claimed at the same time. 
Plus, as you can see from the other messages on this subject, Michael Rusch
wrote that he faced the same problem. I guess there must be a bug somewhere.

The solution of tampering with the priorities doesn't work for me since I
don't have the necessary permission on pool B.

Thomas

Cyclotron Institute, Texas A&M university
ZIP 77843-3366
(979)-845-1411 ext. 258
Mobile: (979)-571-9782
homepage: http://demon.ulb.ac.be/yeehaa/yeehaa.html  

> -----Original Message-----
> From: condor-users-bounces@xxxxxxxxxxx 
> [mailto:condor-users-bounces@xxxxxxxxxxx] On Behalf Of Erik Paulson
> Sent: Tuesday, September 27, 2005 21:09
> To: Condor-Users Mail List
> Subject: Re: [Condor-users] Flocking drawback
> 
> On Tue, Sep 27, 2005 at 08:49:29PM -0500, Thomas Materna wrote:
> > What do you mean by a mistake requirement? The job requirements is 
> > just
> > (Machine==machina1-in-poolA) || (Machine==machina2-in-poolA) || ...
> > It has no reason to start on a machine in pool B, and actually it 
> > doesn't really start but the machines are claimed for a 
> brief moment.
> 
> If those machines were claimed by a job, even for a brief 
> moment, then the requirements expression were satisifed at 
> the matchmaker. If the requirements expression is correct, 
> and _ONLY_ matches with machines in pool A, then it is a 
> different job matching with it.
> 
> I'm certainly not going to rule out a bug in Condor, but the 
> matching code is pretty solid.
> 
> Any hope of getting similar jobs submitted again so we can 
> see what's going on?
> 
> -Erik
> 
> _______________________________________________
> Condor-users mailing list
> Condor-users@xxxxxxxxxxx
> https://lists.cs.wisc.edu/mailman/listinfo/condor-users
>