Hello Deborah,
the issue is that the run command did not set a jobid entry in the properties.ini file.
This needs to be fixed in simfactory in this file: lib/simrestart.py which needs to be modified such that the userRun command stores a jobid in the same way that the submit command does in line 765.
This would be a bit of a hack since self.userRun is analogous to self.userSubmit and self.userSubmit does not store the jobid (sself.ubmit does, but run cannot store the self.jobid since it exeutes from *within* the job). To make it worse, self.userRun does not even know the jobid since the jobid is returned by the submit command, the submit entry in the machine.ini file, which is normally qsub/sbatch/whatnot but configured to by something else that just starts the SubmitScript as a background process and returns the process ID for machines without schedulers.
So something like:
'jobid': os.getpid()
in line 945 (just after 'pbsSimulationName': pbsSimulationName) probably does the trick. It's a hack though since it makes assumptions what the job id will be (namely my process ID) and is dangerous in case process IDs are re-used (there's only 65k of them so this does happen).
Yours, Roland
Periodically when continuing a run, I get the error
Error: job id is negative
Aborting Simfactory
It is fixed by deleting the symlink to the previous active directory and then resubmitting.
Although that fix works, when I want to submit a job multiple times at once (so that when one output is finished running, the next one starts automatically), this error prevents me from doing that.
Anyone have any ideas on what might be the underlying cause of this issue and how to actually fix it?
Thanks!
Deborah
Deborah Ferguson
Graduate Student
Georgia Institute of Technology
Center for Relativistic Astrophysics
Boggs 1-66