Fixed #1871327. Didn't tested #1872068 yet.

@adamw: the test realmd_join_sssd cannot work, I think. It's writing nameservers to /run/systemd/resolve/stub-resolv.conf, expecting something to happen.

Solves issue with machine stuck for 2 hours with abortd: https://usercontent.irccloud-cdn.com/file/IMWCgNNw/Screenshot%20from%202020-09-02%2021-11-05.png

@zbyszek for a test that can't possibly work it sure works fine ;)

the failure of the upgrade test is a known bug that happens on all F33 update tests, if you check the comments tab you'll see a link to the bug that flags this: https://bugzilla.redhat.com/show_bug.cgi?id=1871990

For the record, there's more change in this update than there really should be. There is a policy, though I've been lax in enforcing it lately, that updates during freeze should contain only as much change as is necessary to fix accepted blocker/FE bugs. This obviously changes a lot more, going by the description.

I guess we'll take it for now, but take this as advanced warning we might start being stricter about this again in future...

@adamwill This update changes how /etc/resolv.conf is handled, so it's more likely to be symlink, not a static file, and echo'ing stuff to /etc/resolv.conf, i.e. to /run/systemd/resolve/stub-resolv.conf is likely not the right thing [https://openqa.fedoraproject.org/tests/653019/modules/realmd_join_sssd/steps/1/src, line 25]. I assumed the test failed because of this, but maybe it was some other reason...

For the record, there's more change in this update than there really should be.

True, but it's still just cherry-picked patches with clear bugfixes.

Didn't spot any issues on default F33 installation

Do I need to do something to get this pushed? It's been waiting for a few days now.

no regressions noted