What steps will reproduce the problem?
1. Start running jobs using the `shellcmd` backend
2. Interrupt GC3Pie while the jobs are still in RUNNING state
3. Repeat from step 1. until all available slots are consumed (`max_cores`)
What is the expected output? What do you see instead?
Execution slots are allocated but never claimed back. It seems that
the `shellcmd` backend only checks that a PID file is stale when the
job status is updated; if the job is not part of the current session,
it is never updated...
We should probably run a check for stale PIDs when the backend is
initialized. I am not sure however, what happens if a resource file
pertaining to another session is deleted.
Original issue reported on code.google.com by riccardo.murri@gmail.com on 12 Jun 2015 at 10:40
Original issue reported on code.google.com by
riccardo.murri@gmail.com
on 12 Jun 2015 at 10:40