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

RE: [Condor-users] Trouble with a schedd getting out-of-sync withreality



> are you sure this isn't what the system reports due to 
> retiring claims...
> 
> so one job is running but retiring (even though not for a 
> very long time ;) and the other is the preempting claim to 
> the machine but has not yet taken over - thus the schedd 
> considers both 'running' but the startd only lists the one it 
> is really running at that point...

This could definitly be the case. I haven not observed the mis-count go
beyound 20 or so, never into the ridiculous realm of say 80 jobs listed
as running on 14 machines.

I wasn't aware that waiting claims were reported as running. I thought
they had to actually start running on the machine before they took on
the R state in the condor_q output.

- Ian