[Condor-users] out of order condor-g failures in log file

[Peter Doherty <doherty@xxxxxxxxxxxxxxxxxxx>]

I've got a large DAG running with Condor-G on Open Science Grid.
All the nodes are using one log file, and I'm seeing a lot of errors  
like this:

> 000 (887642.000.000) 04/09 17:04:23 Job submitted from host:
> <10.0.10.39:54286>
>    DAG Node: 3cb5-00257
>
> 018 (887642.000.000) 04/09 17:04:55 Globus job submission failed!
>    Reason: 22 the job manager failed to create an internal script
> argument file
>
> 017 (887642.000.000) 04/09 17:05:14 Job submitted to Globus
>    RM-Contact: gridgk01.racf.bnl.gov/jobmanager-condor
>    JM-Contact: https://gridgk01.racf.bnl.gov:20908/26140/1270847110/
>    Can-Restart-JM: 1
> ...
> 027 (887642.000.000) 04/09 17:05:14 Job submitted to grid resource
>    GridResource: gt2 gridgk01.racf.bnl.gov/jobmanager-condor
>    GridJobId: gt2 gridgk01.racf.bnl.gov/jobmanager-condor
> https://gridgk01.racf.bnl.gov:20908/26140/1270847110/



Shouldn't code 018 for a Globus submit failure have to happen after  
the job is submitted to Globus?

Some of these jobs then manage to execute and complete, although I  
have to look further to see if they are running successfully.

Best,
Peter