On Tue, 2019-06-25 at 17:45 +0000, Fedora compose checker wrote:

Missing expected images:

Atomichost raw-xz x86_64 Atomichost qcow2 x86_64

Compose FAILS proposed Rawhide gating check! 3 of 47 required tests failed, 4 results missing openQA tests matching unsatisfied gating requirements shown with **GATING** below

So, quite a few of these failures were down to some appearance changes in GNOME; I've updated the needles and am re-running those tests in prod now.

Dusty, Colin, and atomic@ : you got this mail because fedfind / check- compose considered the Atomic Host images to be "missing", and I just implemented a thing in check-compose yesterday to send you reports when there is an Atomic-related failure (like we used to do with the separate Fedora-Atomic composes):

https://pagure.io/fedora-qa/check-compose/issue/2

there is of course a bug there - fedfind shouldn't consider Atomic Host to be expected for Rawhide any more. I noticed that problem for F30+ updates composes and fixed it prior to rolling out the new check- compose, but didn't notice it also needed fixing for Rawhide and Branched. I will fix that now and send a new fedfind out.

Here are some notes on the real failures:

Failed openQA tests: 18/137 (x86_64), 3/23 (i386), 1/2 (arm)

ID: 415492 Test: x86_64 Server-dvd-iso modularity_tests URL: https://openqa.fedoraproject.org/tests/415492

This fails every day because of https://bugzilla.redhat.com/show_bug.cgi?id=1691030 . I wonder if we should adjust the test not to expect it at this point, since it's been broken for three months.

ID: 415528 Test: x86_64 KDE-live-iso desktop_update_graphical URL: https://openqa.fedoraproject.org/tests/415528

This one is a test issue, KDE just keeps changing how updates and notifications work and it's driving me nuts. https://bugzilla.redhat.com/show_bug.cgi?id=1716005 does not help either. I'll keep trying to make this reliable.

ID: 415532 Test: x86_64 KDE-live-iso apps_startstop URL: https://openqa.fedoraproject.org/tests/415532

Both failures here are test issues (one not waiting long enough before closing the app, one got sandbagged by an 'updates available' notification popping up over the button it was trying to press, thanks again KDE...)

ID: 415534 Test: arm Minimal-raw_xz-raw.xz install_arm_image_deployment_upload URL: https://openqa.fedoraproject.org/tests/415534

This has just been failing forever, it's probably just emulation issues. We should probably take this test out as I never have time to maintain it. Ideally we want to run 32-bit ARM tests on aarch64 hosts, but haven't got around to that yet either.

ID: 415591 Test: x86_64 universal upgrade_server_domain_controller URL: https://openqa.fedoraproject.org/tests/415591 ID: 415608 Test: x86_64 universal upgrade_realmd_client URL: https://openqa.fedoraproject.org/tests/415608

These two go together...I need to look into it further but it looks like perhaps DNS stops working on the server after the upgrade, as the client's attempt to refresh repos when it goes to run its own upgrade fails.

ID: 415601 Test: x86_64 universal install_kickstart_firewall_disabled **GATING** URL: https://openqa.fedoraproject.org/tests/415601

This is https://bugzilla.redhat.com/show_bug.cgi?id=1722979 , mkolman has a fix for it under review ATM.