Re: [Condor-users] Problems with Condor-C in 7.0.4 [Sec=Unclassified]

[Jaime Frey <jfrey@xxxxxxxxxxx>]

On Jul 31, 2008, at 8:24 PM, Troy Robertson wrote:

> Thought I would give this another spin as I had no takers last time.
> Can anyone else at least confirm or deny my problem with Condor-C on
> 7.0.4
> between Windows submit and Linux central-manager/execute
>
> -------------------------------------------
> Hi all,
>
> I having been having problems with Condor-C for a while now.
>
> With 7.0.3 it was that if I submitted more than one job then the jobs
> would execute but then sit on the central manager of the remote pool  
> and
> not be returned until I removed the first job. This was was found to  
> be
> a bug and fixed in 7.0.4
>
> I have now upgraded to 7.0.4 but now if I submit a job from a Windows
> Personal Pool to the remote linux pool,a new remote job is created by
> the central manager every negotiation cycle but is not run.  They stay
> on hold and just accumulate, one new job for every cycle until there  
> are
> heaps of them for the one submitted job.
>
> Both local GridManager logs and remote central manager logs do not  
> seem
> to indicate why this might be the case.
>
> If I try 'Release' one of the held remote jobs its goes back on hold.
> Condor_q -analyse then indicates that it 'Cannot access initial  
> working
> directory' on the local submit machine but I suspect this is a  
> furphy as
> a straight Condor job submitted from the same machine runs fine.
>
> Can anyone please help me run this down?


Ouch. Can you send me the gridmanager log from your submit machine?

Thanks and regards,
Jaime Frey
UW-Madison Condor Team