Soo... we obviously don't want Fedora CI to break (which uses tmt). But I don't understand why is a major update happening right now? Aren't they subject to a Fedora infra freeze as well? And even if not, don't they consider this a bad timing to be making major changes in the stack? Even if we push testcloud through the freeze, regressions might occur and this might affect our Beta release.

But I don't understand why is a major update happening right now?

Because they (tmt) have a monthly release schedule (which always have major changes) and testcloud/virtual provisioning changes weren't ready to merge before.

Aren't they subject to a Fedora infra freeze as well?

AFAIK no, please take a look at the infra SOP.

And even if not, don't they consider this a bad timing to be making major changes in the stack?

No, the development cycle of tmt/testcloud is not affected by the Fedora release cycle.

Even if we push testcloud through the freeze, regressions might occur and this might affect our Beta release.

Well, feel free to help with testing (of the important part of our tooling). I'd tested the way we use it for Validation Testing, works just fine there, as expected (as it is a very basic usage of testcloud).

And obviously,

BetaFE +1

Definitely a bad timing, but I, personally, see little risk for regressions. The code is tested on MRs to tmt. I would push it through just for sake of the tmt development, this is the only place where it will hurt the contributors (tests will be red for f39 only). Of course, in the worst case, we could workaround it in Packit with to pull in latest f39 build anyway from a copr repository, if this is a big problem.

https://cockpit-project.org/blog/tmt-cross-project-testing.html

This would make our life easier and as @frantisekz mentioned the changes are not disruptive. So unless it's a big problem I lean toward including it. If it would present too much troubles we can introduce workarounds on our side to suppress the red test jobs.

BetaFE +1

BetaFE +1

AGREED AcceptedBetaFE

"There is low risk of pulling this in as the new release contains additional code paths for tmt, while the cli we use still relies on the legacy code paths."

The following votes have been closed:

• Accepted BetaFreezeException (+3, 0, -0)
  • +1 by frantisekz, psss, lbrabec

Thanks for your responses, Miroslav and Petr. I'm fine with granting an FE, and I hope there will be no regressions in the process. But I'd like to suggest to take Fedora infra freeze into account next time, to limit the likelihood of breaking things during release periods. Thanks :-)

Release F39 is no longer tracked by BlockerBugs, closing this ticket.