QA » openQA Project » openQA Tests

coolo wrote:

Currently the xen virtualization tests boot into a wrong console and use ttyS1, which is not ending up in serial0 - so all wait_serials and script_run fail.

Only then we login into the system (blindly purely based on timeouts/sleeps as we have no working serial)

Well,we use screenshot, not blindly :(

to then fix the grub config and reboot again. This
is suboptimal and we should have a function in install_and_reboot that fixes the grub config for xen serial before the initial reboot.

Yes, this is suboptimal, will integrate it into install_and_reboot.
This task is with lower priority now, but will find time to do it.

(why are Xen installations broken anyway? isn't that a product bug?)

I'm afraid it is not a bug. Because xen serial console needs special setting, but during installation via pxe, we do not do it.

Well,we use screenshot, not blindly :(

The test is "blind" if it is not using assert_screen or assert_script_run

You use screenshot, but openQA is acting 'blind', and conducting it's steps regardless of what is going on. the Tests are not aware of the state of the system under test. That is coolos point.

This is an autogenerated message for openQA integration by the openqa_review script:

This bug is still referenced in a failing openQA test: gi-guest_sles11sp4-on-host_sles12sp3-xen
https://openqa.suse.de/tests/1019714