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

[Condor-users] condor_submit_dag not spotting dependency bug



Hi,

from one of our users I got the following report:

> I think I found a bug in condor (in the version installed on mauii).

... 6.7.3 for Alpha, running under Debian Sarge

> I have used my old code (by mistake) which had the following bug
> My dag submit file was something like:
> 
> Job 1 bla-bla
> Job 2 bla-bla
> Job 3 bla-bla
> Job 4 bla-bla
> Job 5 bla-bla
> 
> Parent 1 Child 2
> Parent 2 Child 3
> Parent 3 Child 3   <---- Here should be Parent 3 Child 4
> Parent 2 Child 5
> Parent 4 Child 5
> 
> condor_submit_dag should have spotted this error and aborted the 
> execution,
> but it didn't. It finished 1, 2 and gave code error for next jobs 
> ('unknown parameter').

Is this a known problem, and are there any plans to fix
condor_submit_dag? I agree that it might be a PEBKAC, but DAGs tend to
grow large and users would appreciate more assistance from Condor to
find such obvious dependency violations...

Cheers,
 Steffen

-- 
Steffen Grunewald * * * Merlin cluster admin (http://pandora.aei.mpg.de)
Albert-Einstein-Institut (MPI Gravitationsphysik, http://www.aei.mpg.de)
       Science Park Golm, Am Mühlenberg 1, 14476 Potsdam, Germany
e-mail: steffen.grunewald(*)aei.mpg.de * +49-331-567-{fon:7233,fax:7298}