QA (public) » openQA Project (public) » openQA Tests (public)

These tickets really read in a weird way when there is some text above the first header referencing an openQA test result that is also only mentioned afterwards.

@lkocman please don't add text before "Observation" but as part of it - like I did now.

The failed step https://openqa.opensuse.org/tests/1180871/modules/x11/steps/23 you referenced is in the post_fail_hook though which can not be the root cause of the problem. You already mentioned potential performance regressions in e.g. plymouth that could explain it. That's good. Still, what I see is a lot of false-positive matches which at least disturb the test flow, e.g. https://openqa.opensuse.org/tests/1180871#step/x11/1 showing a match of "screensaver" when instead there is a logged in text console seen. This I am handling in #63346 so I will track this ticket as "Blocked" and see what issue might be left after I deleted the faulty needle.

A retriggered test now fails as expected but in a way more explicit way: https://openqa.opensuse.org/tests/1182632# shows that simply the selection of x11 session doesn't yield a valid desktop in time in https://openqa.opensuse.org/tests/1182632#step/x11/1

After checking in an interactive session in https://openqa.opensuse.org/tests/1182650 I am pretty sure that the issue was triggered by a performance regression in a recent gnome submission to Leap 15.2 but the test had a problem since long which is tracked in #48110 : Simply the wrong tty is selected. tests/x11/toolkit/x11.pm calls select_console 'x11' coming from an X11 session but what we do is ask gnome to open another session on tty7 instead of re-using the existing one on tty2.

@mgriessmeier I suggest to prioritize #48110 to have the issue resolved properly soon but you might prefer to handle this ticket instead as a quick fix one.

can be picked up

This has been covered by ggardet_arm and me in #63643, tests green again: https://openqa.opensuse.org/tests/1191353