Allow investigation to schedule modules regardless PUBLISH_ #299
Conversation
b10n1k
force-pushed
the
investigate_publish_workaround
branch
from
f59d553 to
97b883d
Compare
There was a problem hiding this comment.
According to https://progress.opensuse.org/issues/156394#note-14
Jobs would still fail in the same way in before. And how does os-autoinst and openQA handle the jobs trying to publish a non-existant asset "none"?
I am looking at the second part now. Which maybe it would be good idea to do first |
That's actually a good thing. With that I meant that it ran the whole schedule, so the investigation job could reproduce the original failure. But I guess we should set this to not-ready until changes are done on the openQA side. |
Due to conditions in main_common PUBLISH_* variable cant be empty. Set to `None` to allow the scheduling. https://progress.opensuse.org/issues/156394 Signed-off-by: ybonatakis <[email protected]>
b10n1k
force-pushed
the
investigate_publish_workaround
branch
from
97b883d to
9e196eb
Compare
|
@b10n1k can you reply to #299 (comment) please? |
Sorry missing to reply earlier. However we were looking on this with @kalikiana . I remember running a job (with an older commit of os-autoinst os-autoinst/os-autoinst#2470) and observe to observe the behavior. the job finishes without uploading hdd and without any further problem. If there is something else to be done in openQA i havent seen it. openQA seems to handle the rest properly without any change. |
Due to conditions in main_common PUBLISH_* variable cant be empty. Set to
Noneto allow the scheduling.https://progress.opensuse.org/issues/156394