The general idea was, that with oVirt I'd get a little more automation and benefits
than with just using VirtualBox as a GUI for KVMs.
Boy did I underestimate the amount of intellectual investment for the first value return.
Turned out quite a bit bigger than vSphere, but much more intriguing, because after all:
The code was all there! I got hooked... beyond reasonble, perhaps.
With Openstack they jury has come back a long time ago: Unless you're willing to
sacrifice a double digit team for a year or three, don't go near it.
Three node HCI oVirt just looked like a tight little project with plenty of
expandability... sigh!