Do you think it would add significant value to your use of oVirt if
- single node HCI could easily promote to 3-node HCI?
- single increments of HCI nodes worked with "sensible solution of quota
issues"?
- extra HCI nodes (say beyond 6) could easily transition into erasure coding for good
quota management, distinguishable by volumes?
- oVirt clusters supported easy transition between HCI and SAN/NFS storage as initial 1 or
3 node HCI "succeed" into a broader deployment with role differentiation?
- it was validated on "edgy hardware" like Atoms, which support 32GB RAM these
days, nested virtualization with affordable 100% passive hardware?
- oVirt node images were made only from fully validated vertical stacks, including all
standard deployment variants (SAN/NFS/Gluster 1/3/6/9 node HCI) including VDO and all
life-cycle operations (updates)?
- import and export of OVA were fully supported/validated standard operations against
oVirt, VMware and VirtualBox?
- oVirt, Docker, Podman (and OKD) could work side-by-side on hosts, recognizing each
other's resource allocations and networks instead of each assuming it owned the host?
- RealTek drivers, both for onboard and USB3 2.5Gbit were included in the oVirt node
images and actually worked properly across warm reboots?
- nested virtualization was fully supported with oVirt on oVirt for fully testing
migration and expansion scenarios before applying them on the physical hardware?
- Ansible was just 10000x faster?
- oVirt 4.3 could upgrade to 4.4 automagically and with a secure fail-back at any point?
(ok, I know this is getting madly out of hand...)