I think the distinction between system call return codes and software exit codes was correctly observed - it may be that exit 12 means something different for
the software you’re trying to run. Maybe there’s some documentation for the code, or if it’s a script you could “grep ‘exit 12’ test.sh” and see if there’s anything there. If you’ve got a shell script wrapper around a binary that’s being run, you could put a “sleep” statement in an exit trap function in the scripot that would allow
you to condor_ssh_to_job to it after it terminates but before the slot is released so you could have a look around in /proc/self and see if anything looks unusual. Michael V. Pelletier From: HTCondor-users [mailto:htcondor-users-bounces@xxxxxxxxxxx]
On Behalf Of Jeffrey Tsai via HTCondor-users Thank you for the suggestion! I tried increasing the request_memory value to 32GB, but I am still receiving the same output in my error log. Is there a way I can see what the hard limit on memory usage is? Or alternatively, check how much memory my software is trying to allocate? Again, thank you very much for the help! |