Hi folks!
I wanted to put this out there so anyone who's interested is aware. As
you may know, openQA is intended to run a set of tests on all critical
path updates, and also on updates containing packages from a
supplementary whitelist. For the last couple of months or so, it has
been missing occasional updates - the tests are not run on an update
where they should be.
This is due to a Bodhi bug:
https://github.com/fedora-infra/bodhi/issues/3904
openQA relies on the messages published by Bodhi to schedule tests.
When Bodhi fails to publish a message that should trigger openQA
testing on an update, the tests don't run.
I haven't figured out quite how common this is yet; I'd guess it's
affecting something like 1-2% of updates that should have tests run,
but it really is only a guess.
If you do follow openQA test results for your updates and you notice a
case where you think the tests should have been run but weren't, please
poke me or lruzicka on IRC, or mail the test@ list, or file a ticket on
fedora-qa, or something like that, and we can trigger them manually.
Bodhi folks have identified a possible fix for this and we're hoping it
can get out into production in the next couple of weeks.
Thanks!
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net
http://www.happyassassin.net