[ovirt-users] Re: oVirt alternatives

I've been looking there, once I discovered they had axed their own original product (which used to be the only hypervisor officially sanctioned to get CPU partitioning good enough for core based licensing of their SQL servers) and gone with oVirt for their commercial offers. But the most outstanding evidence is that they never made the transition to 4.4, almost two years after 4.3 support stopped at oVirt. The only news since ages is a couple of videos in November: they are up the creek without a paddle, too, and that's the only aspect of this EOL that I find slightly amusing. oVirt is currently made up of so many components over which Redhat has exclusive control, that anyone who isn't Redhat's special friend would be crazy to take it on, because they couldn't keep things coordinated enough to create a product. Actually, my personal impression is that it's what killed oVirt, especially in the HCI variant even inside Redhat. I've just set up three XCP-ng nodes using nested virtualization on one of my home-lab workstations and I've been dumb-struck just how fast and painless it went. I then added a Xen-Orchestra appliance(the equivalent of the management engine), which again dumbfounded me by just how easy and quick it went (a single command grabs the appliance off the Internet and installs it on the node). Of course, then came the inevitable: the nagware! Every other button on the UI is nothing but a hint to upgrade to one of the many paid variants. At least one of those hidden away buttons actually allows you to upgrade the (freshly downloaded..?) nodes, so perhaps the free variant is minimally usable. The XOSAN HCI variant at €6000/year is definitely out of my home-lab range, but might compare favorably to RHV and certainly vSphere or Nutanix. I don't think they are quite in the same league, though, but I'll keep checking as much as I can. It's rather ironical that XCP-ng does support Gluster for HCI... One of the things I loved about oVirt was that you could follow what's going on. In parts of our business we have SLAs where outside help will always come too late. What I didn't like was that you had to dig deep far too often and that it was full of bugs in the setup phase, which had me doubt in its operational performance. In retrospect I have to conceed, that it didn't do so badly there even if I had plenty of scares, which is why I still have 4.3 running in the corporate labs: until EOL of CentOS7 do us part.