SQUEUE(1)                       Slurm Commands                       SQUEUE(1)

NAME
       squeue - view information about jobs located in the Slurm
       scheduling queue.

SYNOPSIS
       squeue [OPTIONS...]

DESCRIPTION
       squeue reports the jobs the controller knows about.  By default it
       shows pending, running, completing and suspended jobs of all users,
       pending jobs first in priority order.  For a pending job the last
       column shows, in parentheses, why it is not running yet; for a
       running job it shows the allocated nodes.

OPTIONS
       -h, --noheader
              Do not print the header line.

       -j, --jobs=<job_id_list>
              Show only these jobs (comma separated), in any state.

       -l, --long
              Show more information, including the full job state and the
              time limit, below a timestamp line.

       --me   Show only the caller's jobs.

       -o, --format=<output_format>
              Choose the columns.  Each field is written %[.][size]type; a
              "." right-justifies the value and size sets the width.
              Useful types include:

              %i  job ID                    %P  partition
              %j  job name                  %u  user
              %t  state, short code         %T  state, full name
              %M  time used                 %l  time limit
              %D  number of nodes           %R  nodes or (reason)
              %r  reason only               %b  TRES per node (GPUs)
              %q  QOS                       %a  account
              %Q  priority                  %S  (expected) start time
              %V  submit time               %E  dependency
              %C  CPUs                      %y  nice value

              The default is "%.18i %.9P %.8j %.8u %.2t %.10M %.6D %R".

       -p, --partition=<part_list>
              Show only jobs in these partitions.

       --start
              Show pending jobs with their expected start time.

       -t, --states=<state_list>
              Show only jobs in these states, given as short codes or full
              names (PD, R, CG, CD, F, CA, ...), or "all".

       -u, --user=<user_list>
              Show only jobs of these users.

       -w, --nodelist=<hostlist>
              Show only jobs that use any of these nodes.

       -V, --version
              Print the Slurm version.

JOB STATE CODES
       PD PENDING    R RUNNING     CG COMPLETING   CD COMPLETED
       F  FAILED     CA CANCELLED  TO TIMEOUT      S  SUSPENDED

JOB REASON CODES
       Priority          Higher-priority jobs are ahead of this one.
       Resources         Waiting for resources to become free.
       Dependency        Waiting for a job it depends on.
       DependencyNeverSatisfied
                         The dependency can no longer be met.
       JobHeldUser       Held by its owner (scontrol release to free it).
       JobHeldAdmin      Held by an administrator.
       QOS...            A limit of the job's QOS would be exceeded
                         (QOSMaxJobsPerUserLimit, QOSGrpGRES, ...).
       ReqNodeNotAvail   A required node is down, drained or reserved.
       PartitionDown     The partition is not accepting new starts.
       PartitionTimeLimit
                         The time limit exceeds the partition's MaxTime.

EXAMPLES
       List pending jobs with the reason they are waiting:

              $ squeue -t PD -o "%.10i %.9P %.10u %.8Q %R"

       Show your own jobs:

              $ squeue --me

       Show jobs running on two particular nodes, without a header:

              $ squeue -h -w node[01-02]

SEE ALSO
       sinfo(1), sbatch(1), scancel(1), sprio(1), scontrol(1)

Slurm                             GreenRay AIO                       SQUEUE(1)
