search

CoBrALab / qbatch

The Unlicense
28 stars 13 forks source link

Investigate heuristics of job re-running #94

Closed gdevenyi closed 8 years ago

gdevenyi commented 8 years ago

How are dependencies handled here, do dependencies fail? Do job numbers change?

gdevenyi commented 8 years ago

Job numbers don't change.

pipitone commented 8 years ago

Can you elaborate? I'm not sure I understand the issue.

gdevenyi commented 8 years ago

SGE

       -r y[es]|n[o]
              Available for qsub and qalter only.

              Specifies whether a job can be rerun or not.  If the value of -r is 'yes', the job will be rerun if it gets aborted without leaving a consistent exit state.  (This is typically the case if the node on which the job is running crashes).  If -r is 'no'
              (the default), the job will not be rerun under such circumstances.  It will still be rerun if it finishes with exit code 99 unless FORBID_RESCHEDULE is set in qmaster_params in sge_conf(5).

PBS:

       -r y|n  Declares whether the job is rerunable.  See the qrerun command.  The option argument is a single character, either y or n.

               If the argument is "y", the job is rerunable.  If the argument is "n", the job is not rerunable.  The default value is 'y', rerunable.

I'm ruminating about what happens if you have a node crash and the job re-runs (as I have configured here for the default). I was wondering about the impact on dependencies.

Thinking about this more, this isn't really a qbatch issue regardless so I think I'll close this.