How to remove a brick from a GlusterFS volume

Removing a brick from a GlusterFS volume shrinks capacity or retires a storage path while keeping the remaining bricks online for clients.

The gluster workflow for brick removal runs as a controlled migration where gluster volume remove-brick marks the target brick for decommissioning, triggers an automatic rebalance to move data away from it, and finalizes the new volume layout during the commit phase.

Brick removal can generate heavy disk and network I/O, and the remaining bricks must have enough free space to accept migrated data. Distributed-replicated and distributed-dispersed volumes must remove whole replica/disperse sets from the same sub-volume, and using force (or committing before migration completes) can make files disappear from the GlusterFS mount while still existing on the old brick path.

1. Display the current brick list for the volume.

   $ sudo gluster volume info volume1
   
   Volume Name: volume1
   Type: Distributed
   Status: Started
   Number of Bricks: 3
   Bricks:
   Brick1: node1:/srv/gluster/brick1
   Brick2: node2:/srv/gluster/brick1
   Brick3: node3:/srv/gluster/brick2

   For Distributed Replicate and Distributed Disperse volumes, remove the full replica/disperse set from the same sub-volume instead of removing a single brick.

2. Start the remove-brick migration for the brick scheduled for removal.

   $ sudo gluster volume remove-brick volume1 node3:/srv/gluster/brick2 start
   volume remove-brick start: success
   ID: fba0a488-21a4-42b7-8a41-b27ebaa8e5f4

   The start phase keeps the volume online while files are migrated onto the remaining bricks.

3. Check the remove-brick migration status.

   $ sudo gluster volume remove-brick volume1 node3:/srv/gluster/brick2 status
                                       Node  Rebalanced-files    size  scanned  failures  skipped       status
                                  ---------  ----------------  ------  -------  --------  -------  -----------
                                     node3               214   12.4GB      812         0        0  in progress

   Re-run the status command until the status field becomes completed, and abort an in-progress run with gluster volume remove-brick volume1 node3:/srv/gluster/brick2 stop if needed.

4. Commit the brick removal when the status shows completed.

   $ sudo gluster volume remove-brick volume1 node3:/srv/gluster/brick2 commit
   Removing brick(s) can result in data loss. Continue? (y/n) y
   volume remove-brick commit: success

   Committing before migration completes can strand files on the removed brick, and running remove-brick with force skips migration and can make data inaccessible from the mount point (treat force as the big red button).

5. Verify the volume no longer lists the removed brick.

   $ sudo gluster volume info volume1
   
   Volume Name: volume1
   Type: Distributed
   Status: Started
   Number of Bricks: 2
   Bricks:
   Brick1: node1:/srv/gluster/brick1
   Brick2: node2:/srv/gluster/brick1

   Check the removed brick path for leftover files before reusing the disk, and copy any remaining data back via a GlusterFS mount point.