Scheduled Downtime
On Tuesday 24 October 2023 @ 5pm MT the forums will be in read only mode in preparation for the downtime. On Wednesday 25 October 2023 @ 5am MT, this website will be down for maintenance and expected to return online later in the morning.
Normal Operations
The forums are back online with normal operations. If you notice any issues or errors related to the forums, please reach out to help@ucar.edu

error with ./case.submit in CLM5

msulis

Mauro
New Member
I am getting the following error when submitting a CLM5 run:

ERROR: Command: 'sbatch .case.run --resubmit' failed with error 'b'usage: sbatch [--ecverbose] [--ecdummy] [-a ARRAY] [-A ACCOUNT] [--bb BB]\n [--bbf BBF] [-b BEGIN] [--bxi1] [--bxi2] [--checkpoint-restart]\n [--comment COMMENT] [--cpu-freq CPU_FREQ] [-c CPUS_PER_TASK]\n [-d DEPENDENCY] [--deadline DEADLINE] [--delay-boot DELAY_BOOT]\n [-D CHDIR] [-e ERROR] [--export EXPORT]\n [--export-file EXPORT_FILE] [--get-user-env] [--gid GID]\n [--gres GRES] [--gres-flags GRES_FLAGS] [-H] [--ignore-pbs]\n [-i INPUT] [-J JOB_NAME] [-k] [-L LICENSES] [-M CLUSTERS]\n [-m DISTRIBUTION] [--mail-type MAIL_TYPE]\n [--mail-user MAIL_USER] [--mcs-label MCS_LABEL] [-n NTASKS]\n [--nice NICE] [--no-requeue] [--ntasks-per-node NTASKS_PER_NODE]\n [-N NODES] [-o OUTPUT] [-O] [-p PARTITION] [--parsable]\n [--power POWER] [--priority PRIORITY] [--profile PROFILE]\n [--propagate PROPAGATE] [-q QOS] [-Q] [--reboot] [--requeue]\n [-s] [-S CORE_SPEC] [--signal SIGNAL] [--spread-job]\n [--switches SWITCHES] [--thread-spec THREAD_SPEC] [-t TIME]\n [--time-min TIME_MIN] [--uid UID] [--use-min-nodes] [-v] [-W]\n [--wckey WCKEY] [--wrap WRAP]\n [--cluster-constraint CLUSTER_CONSTRAINT] [--contiguous]\n [-C CONSTRAINT] [-F NODEFILE] [--mem MEM] [--mincpus MINCPUS]\n [--reservation RESERVATION] [--tmp TMP] [-w NODELIST]\n [-x EXCLUDE] [--exclusive EXCLUSIVE] [--mem-per-cpu MEM_PER_CPU]\n [-B EXTRA_NODE_INFO] [--sockets-per-node SOCKETS_PER_NODE]\n [--cores-per-socket CORES_PER_SOCKET]\n [--threads-per-core THREADS_PER_CORE]\n [--ntasks-per-core NTASKS_PER_CORE]\n [--ntasks-per-socket NTASKS_PER_SOCKET] [--hint HINT]\n [--mem-bind MEM_BIND] [--cpus-per-gpu CPUS_PER_GPU] [-G GPUS]\n [--gpu-bind GPU_BIND] [--gpu-freq GPU_FREQ]\n [--gpus-per-node GPUS_PER_NODE]\n [--gpus-per-socket GPUS_PER_SOCKET]\n [--gpus-per-task GPUS_PER_TASK] [--mem-per-gpu MEM_PER_GPU]\n [--kill-on-invalid-dep KILL_ON_INVALID_DEP] [-h] [--usage] [-V]\n [positional [positional ...]]\nsbatch: error: unrecognized arguments: --resubmit'' from dir '/lus/h2resw01/scratch/lums/clm5_runs/FR-Pue'

It seems the argument "--resubmit" is generating some issues to sbatch command but I did not manage to find out where to fix the problem. Thank you in advance, Mauro.
 

slevis

Moderator
@msulis does this mean that you submitted and ran successfully before trying to resubmit? If so, then you probably already ran "./manage_externals/checkout_externals"? This has to be run when you first checkout code and also when you update code versions. I'm mentioning just in case, because that was my first thought. Otherwise I will let others make suggestions.
 

msulis

Mauro
New Member
Thanks for the feedback, Sam! No, it is a cold start. This problem arises when porting a single-point test case to a new machine. Is there any way to remove the "--resubmit" option when submitting the job? Thank you!
 

msulis

Mauro
New Member
I found out the issue. It happens because the wrapper around the sbatch command doesn't handle passing of arguments to the user script correctly and so is passing them to the real sbatch command instead. A workaround is to use a double-dash to tell the sbatch wrapper to stop parsing arguments.
 

msulis

Mauro
New Member
Unfortunately, the double-dash option is interfering with the submission of the dependent run for archiving the results. Is there any way to use different "sbatch" arguments for the two types of submission? Thank you in advance!
 
Top