First of all, awesome job implementing this on the SGE.
in misopy/cluster_utils.py, the function run_SGEarray_cluster writes a cluster script with qsub parameters. This function writes as one of those parameters -p -1023.
I am unsure if the -p parameter is even necessary (the default is 0) for this script. I may be wrong, but it seems eliminating the hardcoded -p -1023 (or make it an option) may solve this issue.
Hi,
First of all, awesome job implementing this on the SGE.
in
misopy/cluster_utils.py
, the functionrun_SGEarray_cluster
writes a cluster script with qsub parameters. This function writes as one of those parameters-p -1023
.The line is here: https://github.com/yarden/MISO/blob/25abf51b2197d0ccaf42b5a60ecd131eaa099335/misopy/cluster_utils.py#L123
In SGE-6.0u8, it seems to throw the error:
Unable to run job: invalid "job priority" value in job request.
Apparently, it's a known bug that in SGE-6.0u8 qsub throws errors when priority is negative.
https://arc.liv.ac.uk/pipermail/gridengine-users/2006-June/010337.html
I am unsure if the -p parameter is even necessary (the default is 0) for this script. I may be wrong, but it seems eliminating the hardcoded -p -1023 (or make it an option) may solve this issue.