Register SDN informers synchronously#15354
Register SDN informers synchronously#15354liggitt merged 2 commits intoopenshift:release-3.6from liggitt:sdn-register-order-3.6
Conversation
|
continuous-integration/openshift-jenkins/merge Waiting: You are in the build queue at position: 14 |
|
Evaluated for origin merge up to 07c92dd |
|
[Test]ing while waiting on the merge queue |
|
continuous-integration/openshift-jenkins/test FAILURE (https://ci.openshift.redhat.com/jenkins/job/test_pull_request_origin/3303/) (Base Commit: 12575c5) (PR Branch Commit: 07c92dd) [test] |
|
I'm seeing a consistent failure in the ansible install at the "Reconcile Cluster Roles and Cluster Role Bindings and Security Context Constraints." step:
|
|
all-in-one master is dying with |
|
looks like all the controller roles moved out of the block that auto-reconciles them on server start, which means ansible reconciliation races controllers which kill the process if they hit permission errors long enough |
|
Evaluated for origin test up to 04fdc79 |
|
continuous-integration/openshift-jenkins/test SUCCESS (https://ci.openshift.redhat.com/jenkins/job/test_pull_request_origin/3338/) (Base Commit: 4c2392b) (PR Branch Commit: 04fdc79) |
|
@deads2k PTAL at the second commit for 3.6 |
|
lgtm |
|
[merge] |
|
green tests on HEAD^, contains fix for flake other merge jobs are hitting, merging |
Pick of #15353 and more contained version of #15364
The SDN controller was registering shared informer event handlers in a goroutine, so registration raced with informer start. If the registration lost, then SDN event handlers would never get namespace events.