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

[HTCondor-users] Condor Submit file: big whole file or several submit files

Good day Condor Users!

Currently, I'm generating 2 kinds of condor submit file:

1.- One whole file, with several "Queue" commands, changing only
InitialDir and Arguments.

2.- Several submit files.

I like the solution (1) because :
    - Is *way faster* than launching several submit files (and I think
is better for Condor).
    - You only have one big submit file instead of several.
    - When using DAG, is way simpler (one Condor submit file, several
jobs related to only 1 node, the DAG file is even human-readable).

But the solution (2) has advantages:
    - if some job fails, I can relaunch it easily.
    - DAGMan can relaunch specific failed jobs (rescue).


- For DAGMan: How can I have one big submit file and get all the
"rescue" benefits? (i.e.: re-running failed jobs only).

- For normal condor_submit : is there an easy way to "rebuild" a
condor_submit entry? (currently I'm working on this, adding extra
"markups", so I can parse the big file and reproduce the failed job).