So for the sake of figuring out what's going on, I un-deployed the hosted-engine from
3 of the 6 hosts, and then removed those 3 from the cluster. I left only 2 VMs up (HE and
another one), and the remaining 3 hosts have the capability of running the hosted-engine
with an Active Score of 3400, so no split-brain scenario could happen.
I simply unplugged the network from one of the hosts, that also happened not to be hosting
any VMs so as to simulate an unexpected shutdown/failure, and in around 30 seconds the 2
VMs (including the HE) were down and stayed down until the failed host was brought back
up.
What I'm trying to say here is that HE and any other highly available VM fail if a
single host (just one) fails. Is there anything I could debug or look into that I
might've missed when setting up oVirt? Is there a misconfiguration somewhere?
Thank you and best regards