[qe-core][functional][epic] rework postfail hooks
From time to time, we end up having post fail hooks that take a long time, see kontact on live cd for instance:
[0m[1;34m[2020-06-19T07:53:36.894 CEST] [debug] ||| starting kontact tests/x11/kontact.pm ... [0m[37m[2020-06-19T08:30:32.577 CEST] [debug] post_fail_hook failed: command 'curl --form upload=@problem_detection_logs.tar.xz --form upname=kontact-problem_detection_logs.tar.xz http://10.0.2.2:20103/H_2GxrpPuX6zL3oK/uploadlog/problem_detection_logs.tar.xz' timed out at /var/lib/openqa/cache/openqa1-opensuse/tests/opensuse/lib/opensusebasetest.pm line 212. [0m[37m[2020-06-19T08:30:32.580 CEST] [debug] Loading a VM snapshot lastgood
If my math is right, that's 30 minutes spent on collecting logs, that for one instance could be greatly reduced by switching to serial terminal, and using scripts rather than typing all of the commands.
Other improvements would be needed, including making openQA display the time spent in a test, even if it fails, which we currently don't have.
#5 Updated by szarate about 1 year ago
See for the reason of tracker change: http://mailman.suse.de/mailman/private/qa-sle/2020-October/002722.html