GPFS Command Reference⚓︎
This page is a practical command reference for common IBM Storage Scale / GPFS administration tasks.
How To Use This Page⚓︎
Use this document as a quick operator reference. Commands are grouped by task so it is easier to find the right starting point during administration or troubleshooting.
Typical placeholders used below:
<filesystem><node><node-list><fileset><ces-node><ces-ip><export-path><client-subnet><user><group>
Cluster Basics⚓︎
Show cluster information⚓︎
Show the cluster manager node⚓︎
Show cluster configuration⚓︎
Change cluster configuration⚓︎
Show mounted GPFS file systems⚓︎
Show GPFS daemon state⚓︎
Show build and runtime diagnostics⚓︎
File System and Capacity Checks⚓︎
Show file system properties⚓︎
Show capacity and inode usage⚓︎
Show filesets⚓︎
Show node classes⚓︎
Health Monitoring⚓︎
Node-level health⚓︎
mmhealth node show
mmhealth node show -N all
mmhealth node eventlog
mmhealth node eventlog --hour --verbose
Cluster-level health⚓︎
mmhealth cluster show
mmhealth cluster show --verbose
mmhealth cluster show node
mmhealth cluster show filesystem
mmhealth cluster show gpfs
Show only unhealthy node output⚓︎
Inspect a specific event⚓︎
Resync health metadata if cluster state looks inconsistent⚓︎
Logs and Diagnostics⚓︎
Common log locations⚓︎
Dump GPFS waiters⚓︎
Review system logs⚓︎
CCR-related checks⚓︎
CES Administration⚓︎
Use these commands when working with Cluster Export Services.
Show CES cluster configuration⚓︎
Show CES service state⚓︎
Show CES addresses⚓︎
Show CES log level⚓︎
List CES events⚓︎
Move a CES IP manually⚓︎
Useful CES troubleshooting commands⚓︎
If CES status looks mismatched between mmhealth and mmces, a metadata resync may help:
NFS Export Management⚓︎
List current NFS exports⚓︎
Add a new export⚓︎
Example:
Add another client to an existing export⚓︎
Remove client access from an export⚓︎
Remove an export⚓︎
Load exports from a saved configuration⚓︎
This is useful when restoring many exports after a service failure or failover issue.
SMB Preparation⚓︎
These are generic Linux-side setup steps often used before exposing a directory through SMB.
Create the shared directory⚓︎
Assign owner and group⚓︎
chown <user> /gpfs/<filesystem>/smb/<share-name>
chgrp <group> /gpfs/<filesystem>/smb/<share-name>
Set permissions⚓︎
User and Access Preparation⚓︎
If a workflow requires a local service account, create it with your site’s standard UID and group conventions.
Create a local user⚓︎
Create a local group⚓︎
Add a user to a group⚓︎
Quick Investigation Set⚓︎
When you need a compact first-pass check, these are a good starting point:
mmlscluster
mmlsconfig
mmgetstate -a -L -v
mmhealth cluster show --verbose
mmhealth node eventlog --hour --verbose
mmdiag --all
mmlsfs <filesystem>
mmdf <filesystem>
mmccr lsnodes
Notes⚓︎
- For troubleshooting workflows, pair this page with gpfs-troubleshooting.md and gpfs-notes.md.