Upgrading Operating Systems and GlusterFS

[anon314159]

Hello,

Is there documentation regarding upgrading a peers operating systems and importing the volume file/peer/uuid data from a backup?

For example, I would like to safely upgrade all of my trusted pools servers from RHEL7 to RHElL8 and the latest supported version of GlusterFS. All of the bricks for each peer reside on separate raid arrays independent of the OS. What (if any) process exists for safely upgrading a bricks OS?

[aravindavk]

Gluster configurations stored in /var/lib/glusterd. Backup this directory from every node (Data in all nodes are not same, each node will have different details wrt peers and local disks).

After the upgrade of OS, restore the respective node data (/var/lib/glusterd) from the backup. Once the data restored in all the nodes then start the Glusterd.

Make sure:

• Hostnames remains same after the upgrade
• Mount the Brick paths in the exact same way as earlier.

If the brick data is also needs backup up and restore, then copying the xattrs and preserving hard links are very important (Ex: -X and -H options for rsync)