On Fri, Aug 27, 2021 at 7:57 PM Gianluca Cecchi <gianluca.cecchi@gmail.com> wrote:


Next step will be to try removing one of the two hosts while still all env in 4.3.10 and then take backup of engine and then install the second as 4.4.8 and see if it goes ok.
I'm going to revert the 4.3.10 snapshot consistent env and try...


Actually this step is the same as what performed (new host, because I pre-remove the existing one...).
And in fact I remember at the beginning of July I made a similar test on the same test env with ovirt 4.4.7 async 2 node iso and didn't have this kind of problem.

In the weekend I had to do similar steps with two different SHE environments with RHV and was able to do as expected, without the fingerprint error and using the same hostname for the first host I redeployed in 4.4.
Passed from latest RHV 4.3 to latest 4.4, that currently is iso 4.4.7.4-0.20210804 and is based on 4.4.7 async 2 (Hypervisor Image for RHV 4.4.z batch#6 (oVirt-4.4.7-2) Async #1)

So I think there was something wrong with my system or probably a regression on this in 4.4.8.

I see these lines in ansible steps of deploy of RHV 4.3 -> 4.4

[ INFO  ] TASK [ovirt.ovirt.hosted_engine_setup : Remove host used to redeploy]
[ INFO  ] changed: [localhost -> 192.168.222.170]

possibly this step should remove the host that I'm reinstalling...?

I will redo the same again on ovirt and eventually open bugzilla with all log files generated.

Gianluca