I have never done this , so you can simulate it on VMs before doint it :
1. Add the new host and put it in maintenance

2. You can backup your Hosted Engine VM and any critical VMs(just incase)
3. Create a new Gluster volume (either replica 3 or replica 2 arbiter  1)
4. Add the new volume via the UI as new storage
5. Do a storage migration from the distributed to the replica 3 volume.
6. Get rid of the old distributed volume and use it for something useful.
7. Remove the new hosts' maintenance and test live migration
8. Set global maintenance (via hosted-engine command) and poweroff the engine VM.
9. Manually power it up on a new host.
10. Remove global maintenance and shutdown a VM
11. Change the same VM to power up only on specific host and power it up.
12.  Remove the limitation of that test VM (from step 10 &  11)

And you are ready to go.

Best Regards,
Strahil Nikolov

On Dec 18, 2019 16:39, Ernest Clyde Chua <ernestclydeachua@gmail.com> wrote:
Hello currently i have a server running glusterfs on distributed 1 and a self hosted engine.
And planning to add two servers for hci.

Do i add new hosts and change gluster volume type to replicated or backup all the vms then start from scratch?

Is there any recommendation on this?