slurm command does not use MB, NPROC from config file

[hughesdj]

We successfully ran the small 200kb test case through fc_run and fc_unzip, under slurm using srun, with falcon-kit 1.4.4 & pypeflow 2.3.0.

But a larger data set fails with "out of memory" when the first srun command uses --mem-per-cpu=4000M --cpus-per-task=1 , and not 'NPROC': '6', 'MB': '30000' as specified in the cfg file.

How do we make the first call of srun use 30000MB? There is no mention of 4000MB anywhere in the cfg file.

"job.defaults": {
"JOB_QUEUE": "standard",
"MB": "30000",
"NPROC": "6",
"job_type": "slurm",
"njobs": "8",
"pwatcher_type": "blocking",
"submit": "srun --wait=0 -p ${JOB_QUEUE} -J ${JOB_NAME} -o ${JOB_STDOUT} -e ${JOB_STDERR} --mem-per-cpu=${MB}M --cpus-per-task=${NPROC} ${JOB_SCRIPT}",
"use_tmpdir": false
},
"job.step.asm": {},
"job.step.cns": {},
"job.step.da": {},
"job.step.dust": {},
"job.step.la": {},
"job.step.pda": {},
"job.step.pla": {}
}
[INFO]In simple_pwatcher_bridge, pwatcher_impl=<module 'pwatcher.blocking' from '/home/data/bioinf_resources/programming_tools/miniconda3/envs/denovo_asm5/lib/python3.7/site-packages/pwatcher/blocking.py'>
[INFO]job_type='slurm', (default)job_defaults={'job_type': 'slurm', 'pwatcher_type': 'blocking', 'JOB_QUEUE': 'standard', 'njobs': '8', 'NPROC': '6', 'MB': '30000', 'submit': 'srun --wait=0 -p ${JOB_QUEUE} -J ${JOB_NAME} -o ${JOB_STDOUT} -e ${JOB_STDERR} --mem-per-cpu=${MB}M --cpus-per-task=${NPROC} ${JOB_SCRIPT}', 'use_tmpdir': False}, use_tmpdir=False, squash=False, job_name_style=0
[INFO]Setting max_jobs to 8; was None
[INFO]Num unsatisfied: 2, graph: 2
[INFO]About to submit: Node(0-rawreads/build)
[INFO]Popen: 'srun --wait=0 -p standard -J P26a7bf2afdd410 -o /home/data/pest_genomics/DH_test/falcon_example5/out/0-rawreads/build/run-P26a7bf2afdd410.bash.stdout -e /home/data/pest_genomics/DH_test/falcon_example5/out/0-rawreads/build/run-P26a7bf2afdd410.bash.stderr --mem-per-cpu=4000M --cpus-per-task=1 /home/data/bioinf_resources/programming_tools/miniconda3/envs/denovo_asm5/lib/python3.7/site-packages/pwatcher/mains/job_start.sh'
[INFO](slept for another 0.0s -- another 1 loop iterations)
[INFO](slept for another 0.30000000000000004s -- another 2 loop iterations)
[...]
[...]
[INFO](slept for another 180.0s -- another 18 loop iterations)
[INFO](slept for another 190.0s -- another 19 loop iterations)
[INFO](slept for another 200.0s -- another 20 loop iterations)
srun: error: rothhpc402: task 0: Out Of Memory