Weak +1 I suppose. I am still not sure doing workstation first is a good plan, but we can give it a go...

Hmm, after reading the discussion on the mailing list, it feels like it might be too early. And I think that starting with Workstation might not be the best idea — I think a non-blocking deliverable or two would be better. We're already on a very tight schedule, and likely to be late with F39, and introducing such a huge and untested (*) change for our main deliverable is very high risk.

(*) "Untested" mostly in the sense of external tooling not being integrated, and people not being accustomed to the new workflow. I know ImageBuilder itself is tested quite well and used. But lack of integration with the wider ecosystem is likely to result in various small changes being needed here and there.

I think we should make a checklist of items that need to be implemented before we switch this on:
- human-readable logs
- better lists in koji?
- openQA conversion to ImageBuilder?

Let's discuss this at the next meeting.

This was discussed during today's meeting:
AGREED: Image Builder-based Workstation images do not replace the current image build process, but are added as non-blocking additional deliverables to allow iterating to a state where we can consider replacing livemedia-creator with osbuild-composer (+5, 0, 0)

This Change seems to be at risk wrt the schedule too. Unless I missed something, the releng ticket has never been acted on and I don't see that we are actually building such an image for F39 currently: the only Workstation images I see in current composes are the 'regular' ppc64le and x86_64 lives (built with livemedia-creator) and the aarch64 disk image (not a live).

Hey @adamwill, sorry I was out of office for a bit. I'll pick up the necessary changes for the pungi configuration tomorrow. Which will then be linked in the releng ticket.