Open natefoo opened 9 years ago
Well this is tricky - probably need to share some new code between Pulsar and Galaxy. I think the right solution to this would also let Pulsar jobs work with the job resubmit code in Galaxy.
Note I am not entirely sure this is correct.
I think in general we should maybe merge a lot of the job handling code wherever possible?
I think in general we should maybe merge a lot of the job handling code wherever possible?
Solid goal for sure.
Galaxy currently does this for walltime and memory limits in slurm. Right now, if a Pulsar job hits the walltime, I believe the most likely result in Galaxy is the "failed to communicate with remote job runner" message.