I'll presume you didn't fully backup your hosts root file systems on the host which was fried.

It may be easier to replace with a new hostname/IP.
I would focus on the gluster config first, since it was hyperconverged.

I don't know which way engine UI is using to detect gluster mount on missing host and decides not to remove the old host.
You probably also have the storage domain "mounted in the data-center" with backup volume servers pointing at the old host details
The remaining gluster peers also notice the outage, and it could be detecting that?

I would try to make gluster changes, so maybe the engine UI will allow you to remove old hyperconverged host entry.
(The Engine UI is really trying to protect your gluster data).
I'd try changing the mount options and there is a way to tell gluster to only use two hosts and stop trying to connect to the
third, but I don't remember the details.










On Thu, Jun 6, 2019 at 4:32 PM <adrianquintero@gmail.com> wrote:
Definitely is a challenge trying to replace a bad host.

So let me tell you what I see and have done so far:

1.-I have a host that went bad due to HW issues.
2.-This bad host is still showing in the compute --> hosts section.
3.-This host was part of a hyperconverged setup with Gluster.
4.-The gluster bricks for this server show up with a "?" mark inside the volumes under Storage ---> Volumes ---> Myvolume ---> bricks
5.-Under Compute ---> Hosts --> mybadhost.mydomain.com the host  is in maintenance mode.
6.-When I try to remove that host (with "Force REmove" ticked) I keep getting:
    Operation Canceled
             Error while executing action:
    mybadhost.mydomain.com
            - Cannot remove Host. Server having Gluster volume.
Note: I have also confirmed "host has been rebooted"

Since the bad host was not recoverable (it was fried), I took a brand new server with the same specs and installed oVirt 4.3.3 on it and have it ready to add it to the cluster with the same hostname and IP but I cant do this until I remove the old entries on the WEB UI of the Hosted Engine VM.

If this is not possible would I really need to add this new host with a different name and IP? 
What would be the correct and best procedure to fix this?

Note that my setup is a 9 node setup with hyperconverged and replica 3  bricks and  in a  distributed replicated volume scenario.

Thanks
_______________________________________________
Users mailing list -- users@ovirt.org
To unsubscribe send an email to users-leave@ovirt.org
Privacy Statement: https://www.ovirt.org/site/privacy-policy/
oVirt Code of Conduct: https://www.ovirt.org/community/about/community-guidelines/
List Archives: https://lists.ovirt.org/archives/list/users@ovirt.org/message/N4HFTCWNFTOJJ34VSBHY5NKK5ZQAEDB7/