sshare

List share information

$ sshare

$ sshare --parsable

$ sshare --json

$ sshare -o [format_string]

$ sshare -u [user1,user2]

$ sshare -A [account1,account2]

$ sshare -a

sshare [OPTIONS...]

sshare displays fair-share information for Slurm's priority/multifactor plugin. It shows how accounts and users are allocated shares, their actual usage, and resulting fair-share factors that influence job scheduling priority.
The data requires slurmdbd to be configured and operational. Fair-share scheduling ensures equitable resource distribution among users and projects based on their allocated shares and historical usage.

-A, --accounts accounts

Show data for specified accounts (comma-separated)

Display information for all users

Show data for specified users (comma-separated)

Print only user information (exclude ancestors)

Target specific clusters

Custom field selection

Extended output with normalized data

Omit header row

Pipe-delimited with trailing delimiter

Pipe-delimited without trailing delimiter

Output as JSON

Output as YAML

Show partition names

Verbose output

Print version

Account: Association account
User: User name
Raw Shares: Assigned share count
Norm Shares: Normalized shares (0.0-1.0)
Raw Usage: Raw usage value
Norm Usage: Normalized usage
Effectv Usage: Effective usage considering hierarchy
FairShare: Fair-share factor for scheduling

Requires slurmdbd and the multifactor priority plugin. Share values are relative within the hierarchy. Usage data updates with accounting polling intervals. Historical usage decay depends on PriorityDecayHalfLife configuration.

sshare is part of Slurm, developed at Lawrence Livermore National Laboratory starting in 2002. Fair-share scheduling was added to support equitable resource allocation in multi-user HPC environments. Slurm is now maintained by SchedMD.

sprio(1), sacctmgr(1), sreport(1), scontrol(1)