As stated, it will require building lago for ppc64.
It's difficult to evaluate the amount of time that is needed for this task,
but for sure it's not something immediate.

I remember this wasn't possible due to missing pkgs on power, but I don't remember any details.

Is it possible given the time and resources?

What about starting with a functional test that only installs hosts like VDSM check-merged is doing?

I guess that will be the same as simply running VDSM's check-merged on PPC64LE. That would indeed be easier then running the full OST, but we will still need to start building Lago for PPC64LE.

What about starting with a functional test that only installs hosts like VDSM check-merged is doing?

Let's track them one by one and block this issue.
We need to end up with a full arch coverage CI at some point and we need to start somewhere.

Unless we're going to either allow running x86_64 VMs on PPC hosts or build engine for PPC, our current infrastructure cannot run OST on PPC.

A PPC Lago port would be needed, and if no PPC build of engine will be done, I guess we will need to add multi-arch capabilities to Lago (Can PPC qemu/kvm emulate x86_64 machines?).

Another option to consider is creating some kind of "Lago swarm" where Lago would be able to run an engine VM on an x86_64 host and a host VM on a PPC host, and have them talk to each other. Perhaps we could use oVirt itself as the substrate to do this once oVirt gains a simple an reliable overlay networking support.

Bottom line - several new Lago and possibly oVirt features would be required to accomplish this.