https://review.openstack.org/#/c/51793/ is a good instance of a behavior I've seen a lot of recently, where someone approves a patch that last ran CI on it a month ago (the last Jenkins pass on this patch was Nov 19th).
If you come across a patch like that, as a core reviewer, please
"recheck no bug" to make sure it actually passes.
This patches unit tests don't. And what happens then is it becomes a
wrecking ball.
It fails, and gets pulled to the side, causing a reset for anything
behind it. In this case with unit tests failing that means a 20 - 30
minute delay to everything behind it. If anything in front of it fails,
zuul puts it back into rotation, because the change in front of it that
failed *might have been the problem*. Then it fails again, resets the
queue behind it. Another 20 - 30 minute delay.
If there are lots of other races in the gate, and it's a long queue, a
change like this could add *hours* of gate delay.
So please look for recent passes before +Aing anything.
-Sean
--
Sean Dague
Samsung Research America
s...@dague.net / sean.da...@samsung.com
http://dague.net
signature.asc
Description: OpenPGP digital signature
_______________________________________________ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
