[ovirt-users] Re: How to create a backup in event of hardware failure of a single hosted engine?

On Wed, Nov 18, 2020 at 1:40 PM <ernestclydeachua(a)gmail.com&gt; wrote: Indeed, for production use, you need replica 3 gluster setup (either replica 3 or 2 + 1 arbiter). In your case, you may go temporarily with a replica 2 (two hosts) setup. In this case, it is best practise to dedicate a separate network for the gluster traffic. Lets assume gluster0 and gluster1 are the hostnames of each host at the gluster storage network and the bricks are at /gluster/engine/brick. Then you need to run from gluster0 host: gluster peer probe gluster1 gluster volume add-brick engine replica 2 gluster1:/gluster/engine/brick In that case though you need to adjust quorum: gluster volume set engine cluster.server-quorum-type none gluster volume set engine cluster.quorum-type fixed gluster volume set engine cluster.quorum-count 1 Repeat for each volume and wait for heal (sync to complete). Check heal status of each volume: gluster volume heal <volume name> info I do not include all details here. I assume the existing gluster volumes are already configured with the appropriate settings. As soon as you have the replica 2 setup in place, you then can proceed with a replica 3 setup with the same approach. Remember to enable quorum at each volume. Hope that helps.