Closed choosehappy closed 4 months ago
As this is only present in snakemake_interface_executor_plugins
and code making use of it, not in the snakemake
core, I too wonder how this can be tuned.
Meanwhile, I too, find this value a bit high: Particularly when teaching, but also upon development (when many jobs fail) such a default hinders progress.
So, is there a way to set this without tinkering with the code? Would a default of 5 seconds (or even lower) be more sensible? If a slurmdb, as suggested in the code comment, can take half a minute to be queried, there is something wrong with the cluster, and we might reasonably be catching issues when inquiring the job status, right?
Amazing, thank you!!
wow - you are fast. I wanted to write to you, only to find this issue already closed.
I am sorry, it took so long.
Yea no problem -- really ... amazingly, its coming in right on time : )
I nicely have everything up and running with the latest versions of snakemake and snakemake-executor-plugin-slurm pulled from github
Thanks for all your hard work!
One quick question that i can't suss out from the available documentation - how do modify the CommonSettings parameters, and in particular this one:
https://github.com/snakemake/snakemake-executor-plugin-slurm/blob/58af422542aa4a4e503d9b61123d52c01b41893a/snakemake_executor_plugin_slurm/__init__.py#L43
my current command line looks like this, but no modifications seem to be able to impact that variable:
I manually changed the variable from 40 down to 2, and in my test situation the computation time goes from 4 minutes down to 40 seconds -- most of it is clearly spent in waiting to begin querying for jobs which are already completed
Thanks again!