QA » openQA Project » openQA Tests

Oliver could you please check? (as you are maintainer).
It seems that

    wait_screen_change { type_string $realname };
    send_key 'tab';      # Select password field
    send_key 'tab';
    $self->type_password_and_verification;

Thank you.

I doubt that the test module does anything wrong here. I am not aware of any problem within "user_settings" within SLE validation tests for multiple months and in the described scenario it is very likely to happen. https://openqa.suse.de/tests/2114659#step/user_settings/4 shows characters missing when typing the username. So there is not much different we can do in the test except for typing slower but why should we?

Looking at the history in more details it seems that the tests were looking better at https://openqa.suse.de/tests/2033951#next_previous from 22 days I can not find user_settings failing in 25 tests.

I suspect it can only be one problem: The questionable approach of adding the huge amount of addon repositories over the GUI which is very far from what a sane human would do is making the yast installer hit which might be supported by the observation within https://openqa.suse.de/tests/2110537/file/user_settings-ps_auxf with a huge list of defunct zombie processes. I guess I can not help much there. We could try to type super slow or wait until the installer recovers from whatever it is busy with for the time being if that will help at all but this also feels so wrong. So WDYT?

Sorry, don't know what to do better. I guess we should work on #41921 first?

Main maintenance tests have been converted to being based on autoyast. The scenario as in https://openqa.suse.de/tests/latest?distri=sle&test=qam-allpatterns%2Baddons&flavor=Server-DVD-Updates&arch=x86_64&version=12-SP3&machine=64bit#next_previous and jobs in its history show that it is really stable now.