On 2026/02/18 01:56, Adam Novak via slurm-users wrote:
... Toil can't handle multiple copies of the same job running at once ... Is it possible to write an idempotent sbatch command, where it can be run any number of times but will only actually submit one copy of the job?
Could you not make use of the
--dependency=singleton
constraint, to achieve something close to what your meta-scheduler needs?
From the sbatch manpage:
singleton This job can begin execution after any previously launched jobs sharing the same job name and user have terminated. In other words, only one job by that name and owned by that user can be running or suspended at any point in time. In a federation, a singleton dependency must be fulfilled on all clusters unless DependencyParameters=disable_remote_singleton is used in slurm.conf.
You would still need to catch any queued dupe(s) that your meta-scheduler created but there wouldn't be two running at once.