setup watches before creating the buildconfig #12702
Conversation
|
@csrwng ptal |
bparees
force-pushed
the
controller_flake
branch
4 times, most recently
from
60b08f4 to
adb9252
Compare
|
[test] |
|
Passing the wrong resource version to the watch is bad. But regarding the ordering, I don't see how a build would be created by just creating the buildconfig. An imagestreamtag has to be created for a build to be created, and that was getting created after the watch started. LGTM |
|
yeah you're right about the triggering part, I made a quick assumption about what was causing the build to get created that was incorrect. oh well, hopefully it was the resource version issue that was causing the flakes. [merge] |
bparees
force-pushed
the
controller_flake
branch
3 times, most recently
from
84f26bf to
a24a394
Compare
bparees
force-pushed
the
controller_flake
branch
from
a24a394 to
95c082e
Compare
|
Evaluated for origin merge up to 95c082e |
|
Evaluated for origin test up to 95c082e |
|
continuous-integration/openshift-jenkins/test SUCCESS (https://ci.openshift.redhat.com/jenkins/job/test_pr_origin/13402/) (Base Commit: 33c36f4) |
|
continuous-integration/openshift-jenkins/merge SUCCESS (https://ci.openshift.redhat.com/jenkins/job/test_pr_origin/13405/) (Base Commit: 5ee0737) (Image: devenv-rhel7_5808) |
fixes #12017
between the fact that we were watching on builds based on the buildconfig resource version, and that we were starting the watches after the build was potentially created, there was probably a race condition here if the build got created really quickly.