Today i had plan to start upgrading my installation to the latest 4.2. As i was already on a 4.2.x version i have figured it will be easy .. well.guess what i have having trouble re-activating one of the freshly upgraded host. So here is the resume.

My setup have 3 hosts that are also part of a gluster pool. Those are the original host that was use to setup the HA engine with glusterfs. One of the 3 host is only used for arbiter and can also run the engine. But nothing else can run on it because it is a fairly small machine.

I have also 6 others hosts for running vms that are not part of the gluster pool.

So first thing i did to upgrade is set the global maintenance. Update the Engine/restart the engine and remove global maintenance.. that went without a glitch.

Next i choose the arbiter for the first machine to upgrade. Then set the machine to maintenance run the upgrade and reboot the host. That went also without a glicth.

Now here is the glitchy part. When i try to reactivate the host i got a strange error with a gluster peer command that failed on a host that is not in the gluster peer list.

Here is the two errors i had.

"Gluster command [gluster peer probe ovhost3] failed on server sfmo5002ov74."

"Add Host failed error: ovhost3 is either already part of another cluster of having volumes configured return code: 3"

One thing i do not understand is the sfmo5002ov74 is not part of the gluster pool. It does not provide any bricks to the system.

Now i am not sure what to look for. I have stop the upgrade process until i figure out the issue.

Any hints would be appreciated.

Thanks

Carl