tallakahath
commented
5 years ago This was written as (more than a bit of) a terrible hack back-in-the-day. As a quick fix I'd suggest swapping the order - so few places use qsub these days and I can't think of any that use qsub and sbatch AND want qsub preferred. Longer term, a config file in ~/.pbs would fix this (assuming ~/.pbs is still in-use).
On Wed, Apr 10, 2019 at 2:15 PM Jonas Kaufman notifications@github.com wrote:
I was using casm/pbs on a cluster that uses slurm for scheduling but for some reason has both sbatch and qsub available as submission commands. Because getsoftware() first looks for qsub it determined that torque was the scheduler being used, but then pbs was unable to submit jobs properly.
Unless there is a more robust way to check which scheduler is being used, I suggest adding a case for if both qsub and sbatch are found and then prompting the user to give the correct one to use.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/prisms-center/pbs/issues/17, or mute the thread https://github.com/notifications/unsubscribe-auth/AGbfpH7I4e0yEzpUWkeKNRFQpDdUjA7Wks5vfipJgaJpZM4cn8_g .
bpuchala
I was using casm/pbs on a cluster that uses slurm for scheduling but for some reason has both sbatch and qsub available as submission commands. Because getsoftware() first looks for qsub it determined that torque was the scheduler being used, but then pbs was unable to submit jobs properly.
Unless there is a more robust way to check which scheduler is being used, I suggest adding a case for if both qsub and sbatch are found and then prompting the user to give the correct one to use.