v7.1.1 multicluster support breaks lsf profile

[ASLeonard]

The changes made by #1459 break cluster execution using the "lsf" profile created by @mbhall88. I'm slightly guessing here, but I think the change now quoting the argument "{statuscmd} '{jobid}'".format( causes the "lsf" profile to try and parse the jobid as ' ' which is not a base10 number and causes an error.

Simple fix is to change the lsf_status.py script here and the line below to split the argv[1] and use the first for jobid and second for log. (Pull request here)

It is probably impossible to test-case every cluster profile, so not sure if this can be realistically avoided in the future.

[johanneskoester]

@mbhall88 any thoughts about this?

[mbhall88]

Best to continue the discussion on Snakemake-Profiles/lsf#46 I think

[ASLeonard]

Effectively resolved through Snakemake-Profiles/lsf#47

[jdblischak]

My apologies for inadvertently breaking the LSF profile! I didn't know that LSF passed the log file along with the job ID. Glad to see you were able to so quickly identify and fix the problem though.