Move send_key 'ret' above softfail check #6697
Conversation
The softfail check was one line too early, after moving send_key 'ret'; above it, the softfail gets recorded and the workaround works as intendet. Verification run: <http://agraul-vm.qa.suse.de/tests/94#step/yast_virtualization/45>
| @@ -54,13 +54,13 @@ sub run { | |||
| systemctl 'start libvirtd', timeout => 60; | |||
| wait_screen_change { send_key 'ret' }; | |||
| systemctl 'status libvirtd', timeout => 60; | |||
| send_key 'ret'; | |||
There was a problem hiding this comment.
I don't understand what we need this 'ret' for at all. Can you show a job that failed?
There was a problem hiding this comment.
I can try without 'ret', but e.g. in https://openqa.opensuse.org/tests/843249#step/yast_virtualization/42 the softfail was not triggered even though the return code of the check should have been zero (judging from the screenshot, grep did find the Active: no line). I don't really know why that is the case and why it worked when I moved the 'ret' before the check 😕
There was a problem hiding this comment.
Well, taking a look into the latest job in the current Tumbleweed tests seem that your workaround is just fine, see https://openqa.opensuse.org/tests/843910#step/yast_virtualization/45 . Your hypothesis that the 'ret' key press helps seems like a stretch ;)
There was a problem hiding this comment.
Created a new needle and retriggered the production test https://openqa.opensuse.org/tests/844239
There was a problem hiding this comment.
Well, taking a look into the latest job in the current Tumbleweed tests seem that your workaround is just fine, see https://openqa.opensuse.org/tests/843910#step/yast_virtualization/45
ahh, thanks! I tried it on my openQA instance before that job was run and it failed - but that seems to be unrelated, I have some instability problems right now.
Created a new needle and retriggered the production test https://openqa.opensuse.org/tests/844239
If your new needle looks like this one, virt_install should pass ;)
I'll close the PR as the change is not needed.
The softfail check was one line too early, after moving send_key 'ret';
above it, the softfail gets recorded and the workaround works as
intendet.