I found that I had to change SGE.pm, line 217 in RUM-Pipeline-v2.0.3_01
from: my $cmd = "qsub -V -cwd -j y $dir_opt @args 2>&1";
to: my $cmd = "qsub -S /bin/bash -V -cwd -j y $dir_opt @args 2>&1";
as the default shell in my SGE cluster was /bin/csh and the commands in outdir/rum*{proc,preproc}.sh are bash cmds that won't work in csh.
Please make this change to your code, as it could save others a long time to trace the root cause of all the errors I was seeing in the log files.
On a related note, please allow user to specify additional qsub options (as a string may be so any qsub option can be specified by changing above line to "qsub -S /bin/bash $option_string .."). I've several resource options like "-l h_vmem=6G' that are specific to my SGE environment which I would like to add to any qsub command that is submitted by RUM pipeline.
I found that I had to change SGE.pm, line 217 in RUM-Pipeline-v2.0.3_01
from: my $cmd = "qsub -V -cwd -j y $dir_opt @args 2>&1"; to: my $cmd = "qsub -S /bin/bash -V -cwd -j y $dir_opt @args 2>&1";
as the default shell in my SGE cluster was /bin/csh and the commands in outdir/rum*{proc,preproc}.sh are bash cmds that won't work in csh.