Below are the steps to take to upgrade the cluster from version 3.12 to 4.1.7 which will align with the staging cluster. This version has long term support from the GlusterFS
You want to perform these updates one node at a time, this way you can keep the cluster up while each node is updated one by one.
1. Stop the gluster and CTDB services
systemctl stop glusterd; systemctl stop ctdb
2. Remove the gluster 3.12 package
yum remove centos-release-gluster312
3. Install the gluster 4.1 package
yum install centos-release-gluster41
4. Update all of the glusterfs packages Note: there is python2-gluster and samba-vfs-glusterfs, they should get updated because they are dependencies.
yum update glusterfs*
5. Check to see that all packages are now running version 4.1.7.
rpm -qa | grep gluster
6. Back up the smb.conf file as well as the /etc/sysconfig/ctdb file before we update the samba and ctdb packages. In our experience, updating will overwrite these files.
cp /etc/samba/smb.conf /etc/samba/smb.conf.back; cp /etc/sysconfig/ctdb /etc/sysconfig/ctdb.back
7. Update to the new version of samba, samba-vfs-glusterfs and ctdb (4.8.3-4.el7.0.1)
yum update --enablerepo=fasttrack samba*
8. Check /etc/samba/smb.conf and /etc/sysconfig/ctdb to see if they were overwritten, if so, copy the backups over to the file
cp /etc/samba/smb.conf.back /etc/samba/smb.conf; cp /etc/sysconfig/ctdb.back /etc/sysconfig/ctdb
9. Restart the gluster and CTDB services
systemctl start glusterd; systemctl start ctdb
10. Check that the CTDB returns to “OK” on that particular node. If not check /var/log/log.ctdb as it will let you know why it is staying Unhealthy or failing to start the service.
Watch -n1 "ctdb status"
11. Once all is good, move onto the next node and follow the same process.