wiki:SLURMHints

Useful SLURM commands

Bill Wichser

Here is a list of some of my favorite commands for Slurm.

slurmtop

sinfo

squeue

seff

sdiag

sshare -l

slist

snodeacct

sacctmgr list qos format=name,grpcpus,maxnodes,maxcpusperuser,maxjobsperuser,maxnodesperuser

squeue -o "%.18i %.9P %.8j %.8u %.2t %.10M %.6C %q %R"

squeue -o "%.18i %.9q %.8j %.8u %.2t %.10M %.6C %R"

squeue -o "%.18i %.9q %.8j %.8u %.10a %.2t %.10M %.10L %.6C %R"

squeue -o "%.18i %.9q %.8j %.8u %.10a %.2t %.10M %.10L %.6C %R" -t r -S -e # sort by time ending

squeue -o "%.18i %.9q %.8j %.8u %.10a %.2t %.10M %.10L %.6C %R" -t r -S S # sort by time started

sinfo --format="%12P %.8p %.16N %.10n %.5T %.5c %14C %.7m %.6G"

sinfo --format "%10P %5a %.10l %16F %16C"

sacct --state r

sacct -j <jobid> --format jobname,NTasks,nodelist,MaxRSS,MaxVMSize,CPUTime,SystemCPU,UserCPU

squeue --start

squeue --start --format="%.7i %.7Q %.7q %.15j %.12u %.10a %.20S %.6D %.5C %R" --sort=S --states=PENDING

watch -n 30 -d 'squeue --start --format="%.7i %.7Q %.7q %.15j %.12u %.10a %.20S %.6D %.5C %R" --sort=S --states=PENDING | egrep -v "N/A" | head -20'

squeue --sort=S -t r

sjstat

squeue -w <nodename> -- checknode <nodename> sinfo -R

watch -n 15 -d "squeue -t completed"

sacctmgr modify account molbio set qos+=tiger-vshort where cluster=tiger sacctmgr -p list user name=bill withassoc where cluster=tiger

sacctmgr -p show association where cluster=della

sacctmgr -p show association where cluster=della slurmd -C (on node)

sacctmgr list events -- show bad things on nodes

snacct della-r1c1n1 snodes slist

Last modified 22 months ago Last modified on Dec 7, 2015 2:25:25 PM