Bug 2071139: delete default ingress pod if it is scheduled where another router pod already is#720
Conversation
|
/hold |
4e3a09a to
2779ca2
Compare
|
/test all |
|
/retest |
|
/hold cancel The controller and unit tests merged to library-go, so this bump is now real. |
|
/retest |
/retest |
|
/retest |
|
Fix here openshift/library-go#1344, will pick when it lands. |
|
bumped /hold cancel |
|
Thanks! |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: deads2k, Miciah The full list of commands accepted by this bot can be found here. The pull request process is described here DetailsNeeds approval from an approver in each of these files:
Approvers can indicate their approval by writing |
|
/retest-required Please review the full test history for this PR and help us cut down flakes. |
|
/retest-required Please review the full test history for this PR and help us cut down flakes. |
|
@deads2k: all tests passed! Full PR test history. Your PR dashboard. DetailsInstructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here. |
|
/retitle Bug 2071139: delete default ingress pod if it is scheduled where another router pod already is |
|
@deads2k: This pull request references Bugzilla bug 2071139, which is valid. The bug has been moved to the POST state. The bug has been updated to refer to the pull request using the external bug tracker. 3 validation(s) were run on this bug
Requesting review from QA contact: DetailsIn response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
|
@deads2k: All pull requests linked via external trackers have merged: Bugzilla bug 2071139 has been moved to the MODIFIED state. DetailsIn response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
Anti-affinity in the scheduler has a bug and isn't properly spreading pods. This fits a new library-go controller into the ingress-operator to delete a mal-scheduled pod. If it finds two non-deleted pods using the selector on the same node, then the oldest pod is deleted. Oldest is chosen so that a deployment controller could still make progress. This different that is commonly done to preserve the oldest to disrupt the fewest connections, but it's better to avoid blocking the deployment controller.
I chose to start with ingress because it has the biggest multiplier and this doesn't fully resolve the bug for this component, it only impacts the default ingress.
/hold
This PR is proof for openshift/library-go#1330