QA (public) » openQA Project (public) » openQA Tests (public) » qe-yam

In the scenario where the bootloader screen is stuck, the worker will detect stall and enter post_fail_hook. During post_fail_hook, it has occurred that select_console('root') induced the emergency shell. In this case, there is no check that will redirect any logs to the serial console, like handle_emergency function.

We cannot add the check in activate_console , as the function is called repeatedly in many tests.

A solution would be to find out which step of the activate_console function is triggering the emergency shell and see if we can implement a check for stuck bootloader screen in wait_boot_past_bootloader or post_fail_hook, try to trigger emergency shell and if successful, collect logs.

Failure example:
https://openqa.suse.de/tests/4922078

It may not be easy to reproduce the scenario, as it is sporadic for SUTs with RAID0 root partition, during firstboot.
https://bugzilla.suse.com/show_bug.cgi?id=1178515