As long as CentOS was downstream of RHEL, it was a base so solid it might have been better
than the oVirt node image, even if that was theoretically going through some full stack QA
testing.
But with CentOS [Up]Stream you get beta quality for the base and then the various acquired
parts that make up the oVirt house of cards on top.
If then the oVirt node OS were to go through full-stack QA, that could be quite the better
choice.
But I'm more and more inclined to believe that the if is a false and any testing is
just unit testing.
Around Christmas VDO got dropped from a kernel update of CentOS8 (still non-stream) and my
4.4 oVirt HCI farm dropped dead, until I found out what happend.
To me that was the final straw: I will phase oVirt out with CentOS7.
IBM may have a few more paying customers, but oVirt will lose the edge, the only place
outside the cloud where RedHat can grow without being taxed by the cloud service
providers.