Adapt post_fail_hook to an incomplete openQA setup #149
Conversation
kalikiana
changed the title
kalikiana
force-pushed
the
instl_webui_nofailhook
branch
from
87f124c
to
d07f53e
Compare
|
What I find strange is that the jobs triggered by the PR all are missing the install/* tests (and thus failing): |
| @@ -14,7 +14,7 @@ sub post_fail_hook { | |||
| if (check_var('OPENQA_FROM_GIT', 1)) { | |||
| send_key 'ctrl-c'; # Stop current command, if any | |||
| assert_script_run 'cd /root/openQA'; | |||
| enter_cmd 'script/openqa-cli api jobs'; | |||
| enter_cmd '[ -e script/openqa-cli ] && script/openqa-cli api jobs'; | |||
There was a problem hiding this comment.
This will help if the git clone failed, but not if the cpanm following that fails.
I guess it's really not that easy.
I'm wondering if there is a general way to make it more obvious that a test is in the post_fail_hook stage.
There was a problem hiding this comment.
well, we can just split each test module to be in more specific steps. Then for each module we can define more specific post_fail_hooks
Still not sure but what I see is that https://github.com/kalikiana/os-autoinst-distri-openQA/tree/instl_webui_nofailhook says |
|
@kalikiana I would have still been interested in investigating the CI problems by a simple rebase. But I guess we will see in the next PR if it works or not. |
|
ok, #151 passed, so I assume the reason was that the branch here was based on such an old commit. |
See: https://progress.opensuse.org/issues/138320