Open Aariq opened 2 years ago
Ok, so I think the problem on my end is an incorrect <qsub.jobid>
, but I still think this should fail fast if the job ID is NULL
So to clarify, here's what I think should happen if the job ID is NULL
for any reason (e.g. an incorrect pattern in ): 1) print a message like "Job ID is NULL. Jobs are running but won't be automatically retrieved from host. Check hopst$qsub.jobid in settings" 2)
remote.copy.from()` does NOT get called and the function exits.
This issue is stale because it has been open 365 days with no activity.
Bug Description
runModule_start_model_runs()
should, I think, copy therun
directory over to the HPC, then launch jobs on the HPC, then wait for them to be done (withqsub_run_finished()
), then copy theout
directory back (withremote.copy.from()
). This isn't working for me because the jobid is NULL. Here's a section of the output fromrunModule_start_model_runs()
:I'm not entirely sure what the fix is.
To Reproduce
I think I'd need some guidance on how to reproduce this.
Expected behavior
I'd expect the R session to wait until the HPC runs were finished and results were copied back over. Additionally, it should fail fast in the case the the job ID is not valid (e.g. if it's NULL).
Machine (please complete the following information):
I think @dlebauer and @KristinaRiemer are aware of this bug as well, but couldn't find an open issue.