Closed cmeesters closed 3 months ago
when specifying the cluster in sbatch the output changes to Submitted batch job 61317320 on cluster wice
. This causes line 162 to mistakenly pick up the name of the cluster instead of the jobid.
@meliamne Thank you, your suggestion went into my last commit.
@johanneskoester does the current state find your approval?
I dropped
self.__clusters
as an attribute, insteadsacct
and scancel
use --clusters=all
, as suggested.@johanneskoester we do not have a test, yet (due to faulty multicluster setup in the test-engine). Should we merge regardless, considering that the test will take considerable time and effort to implement?
@johanneskoester not sure, isn't the requested change already in there, resp. the issue solved?
@johanneskoester please have a look, after all this rebasing and fixing, I am a little puzzled. It's embarrassing, but it took too long.
We (@johanneskoester ) and me discussed that we do not have a solution for a multicluster test at hand for the time being. Hence, this PR is merged without a dedicated test.
I am sorry, I could merge this sooner. Too many things kept coming up, then holidays ....
Thanks for this merge! I think there is a mistake in the documentation though:
The snakemake resources seems to be clusters
(with an s), and not cluster
:
I also made a comment regarding the way jobid is now parsed: https://github.com/snakemake/snakemake-executor-plugin-slurm/issues/121#issuecomment-2270905243
Cheers,
Ah, thanks. Will fix this asap (= not today).
Putative fix for issue #53