With some tweaks to the test code (especially not touching resolv.conf) the replica tests do pass, so changing my feedback to positive. Let's just push this stable and pretend everything's fine!
Filed https://bugzilla.redhat.com/show_bug.cgi?id=1886205 for that problem. Assuming it's accepted as a blocker (as it should be) I'd suggest we should edit this update so it's marked as fixing that bug but not fixing the resolved bug, and go ahead and push it.
No. It's not an upgrade test. It's a fresh deployment test. This is exactly how it failed before you tried to deal with resolved. I know the failure is on the master, I think I mentioned that initially in the bug report.
Note, we have an ancillary problem now - upgrade from F32 to F33 doesn't work any more because F32's freeipa is now newer than F33's. We may have to push this anyway just to fix that problem.