sprio
Slurm job scheduling priority viewer
TLDR
SYNOPSIS
sprio [OPTIONS...]
DESCRIPTION
sprio displays the components that determine job scheduling priority when Slurm's multi-factor priority plugin is active. It shows how factors like age, fairshare, job size, partition, and QOS contribute to each job's overall priority score.
By default, it displays information for all pending jobs. The tool helps administrators and users understand why certain jobs are scheduled before others and troubleshoot scheduling issues.
PARAMETERS
-j, --jobs jobids_
Show priority for specific job IDs (comma-separated)-u, --user users
Filter by user names or IDs (comma-separated)-p, --partition partitions
Restrict to specific partitions-l, --long
Display extended output information-n, --norm
Show normalized priority factors (0.0-1.0)-w, --weights
Display configured priority weight values-o, --format format
Customize output format-S, --sort fields
Sort results by specified fields-h, --noheader
Suppress column headers-M, --clusters names
Target specific clusters--federation
Display jobs from all federated clusters--local
Show only local cluster jobs-v, --verbose
Verbose output-V, --version
Display version
FORMAT SPECIFIERS
%i: Job ID
%Y: Priority
%A: Age factor
%F: Fairshare factor
%J: Job size factor
%P: Partition factor
%Q: QOS factor
%u: User name
CAVEATS
Only works when the multi-factor priority plugin is configured in Slurm. Read-only utility; cannot modify priorities. Priority values are relative and depend on cluster configuration. Some factors may be zero if not configured.
HISTORY
sprio is part of Slurm (Simple Linux Utility for Resource Management), developed at Lawrence Livermore National Laboratory starting in 2002. The multi-factor priority plugin was added to provide fair and configurable job scheduling in HPC environments. Slurm is now maintained by SchedMD.
