Skip to content

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⚓︎

Bash
mmlscluster

Show the cluster manager node⚓︎

Bash
mmlsmgr

Show cluster configuration⚓︎

Bash
mmlsconfig

Change cluster configuration⚓︎

Bash
mmchconfig

Show mounted GPFS file systems⚓︎

Bash
mmlsmount all -L

Show GPFS daemon state⚓︎

Bash
mmgetstate
mmgetstate -a -L -v
mmgetstate -s

Show build and runtime diagnostics⚓︎

Bash
mmdiag
mmdiag --all

File System and Capacity Checks⚓︎

Show file system properties⚓︎

Bash
mmlsfs <filesystem>

Show capacity and inode usage⚓︎

Bash
mmdf <filesystem>

Show filesets⚓︎

Bash
mmlsfileset <filesystem>

Show node classes⚓︎

Bash
mmlsnodeclass

Health Monitoring⚓︎

Node-level health⚓︎

Bash
mmhealth node show
mmhealth node show -N all
mmhealth node eventlog
mmhealth node eventlog --hour --verbose

Cluster-level health⚓︎

Bash
mmhealth cluster show
mmhealth cluster show --verbose
mmhealth cluster show node
mmhealth cluster show filesystem
mmhealth cluster show gpfs

Show only unhealthy node output⚓︎

Bash
mmhealth cluster show node | grep -v HEALTHY

Inspect a specific event⚓︎

Bash
mmhealth event show <event-id>
mmhealth event show quorum_down

Resync health metadata if cluster state looks inconsistent⚓︎

Bash
mmhealth node show --resync

Logs and Diagnostics⚓︎

Common log locations⚓︎

Text Only
/var/adm/ras/mmfs.log.latest
/var/log/messages

Dump GPFS waiters⚓︎

Bash
mmfsadm dump waiters

Review system logs⚓︎

Bash
journalctl -p 4
grep "mmfs:" /var/log/messages
Bash
mmccr lsnodes
mmccr check -e -Y

CES Administration⚓︎

Use these commands when working with Cluster Export Services.

Show CES cluster configuration⚓︎

Bash
mmlscluster --ces

Show CES service state⚓︎

Bash
mmces service list -a
mmces state show
mmhealth cluster show ces

Show CES addresses⚓︎

Bash
mmces address list --full-list

Show CES log level⚓︎

Bash
mmces log level

List CES events⚓︎

Bash
mmces events list -a

Move a CES IP manually⚓︎

Bash
mmces address move --ces-ip <ces-ip> --ces-node <ces-node>

Useful CES troubleshooting commands⚓︎

Bash
mmhealth node show
mmhealth cluster show ces
mmces state cluster ces

If CES status looks mismatched between mmhealth and mmces, a metadata resync may help:

Bash
mmhealth node show --resync

NFS Export Management⚓︎

List current NFS exports⚓︎

Bash
mmnfs export list

Add a new export⚓︎

Bash
mmnfs export add <export-path> --client "<client-subnet>(Access_Type=RW)"

Example:

Bash
mmnfs export add /gpfs/<filesystem>/<fileset> --client "10.0.0.0/24(Access_Type=RW)"

Add another client to an existing export⚓︎

Bash
mmnfs export change <export-path> --nfsadd "<client-ip-or-subnet>(Access_Type=RW)"

Remove client access from an export⚓︎

Bash
mmnfs export change <export-path> --nfsremove "<client-ip-or-subnet>"

Remove an export⚓︎

Bash
mmnfs export remove <export-path>

Load exports from a saved configuration⚓︎

Bash
mmnfs export load /path/to/gpfs.ganesha.exports.conf

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⚓︎

Bash
mkdir /gpfs/<filesystem>/smb/<share-name>

Assign owner and group⚓︎

Bash
chown <user> /gpfs/<filesystem>/smb/<share-name>
chgrp <group> /gpfs/<filesystem>/smb/<share-name>

Set permissions⚓︎

Bash
chmod o-rx /gpfs/<filesystem>/smb/<share-name>
chmod g+ws /gpfs/<filesystem>/smb/<share-name>

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⚓︎

Bash
useradd -u <uid> <user>

Create a local group⚓︎

Bash
groupadd -g <gid> <group>

Add a user to a group⚓︎

Bash
usermod -aG <group> <user>

Quick Investigation Set⚓︎

When you need a compact first-pass check, these are a good starting point:

Bash
mmlscluster
mmlsconfig
mmgetstate -a -L -v
mmhealth cluster show --verbose
mmhealth node eventlog --hour --verbose
mmdiag --all
mmlsfs <filesystem>
mmdf <filesystem>
mmccr lsnodes

Notes⚓︎