-
Notifications
You must be signed in to change notification settings - Fork 40.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Revert "Optimization on running prePreEnqueuePlugins before adding pods into activeQ" #117194
Conversation
Please note that we're already in Test Freeze for the Fast forwards are scheduled to happen every 6 hours, whereas the most recent run was: Tue Apr 11 02:03:33 UTC 2023. |
This issue is currently awaiting triage. If a SIG or subproject determines this is a relevant issue, they will accept it by applying the The 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. |
/cc @ahg-g @Huang-Wei |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
/approve
We will see if we can make it to 1.27, or we can wait for the first 1.27 patch release.
/retest |
That seems like a good candidate for inclusion in a 1.27.1 patch release, and noting the issue as a known issue for someone using custom scheduler plug-ins in 1.27.0. |
as a follow-up, I'd recommend scheduling folks open an issue to track improving the testing of external plugins in a way that would have prevented this in the first place |
ok, I'll include the change to add the description as the known issue to the release note once v1.27 is released. |
@sanposhiho can you add an integration test? |
Creating #117215 to track follow-ups. |
We can have a separate test. And have this PR easy to back-port. |
…ds into activeQ" This reverts commit c01fa82.
1246652
to
02d8fc2
Compare
added the known issue section on the release note v1.26. (or should I create a separate PR to update the release note? |
Sure. I'll follow up this in another PR. |
@alculquicondor @Huang-Wei @ahg-g Could anyone take a look? 🙏 |
You already have Wei's approval, you are missing one for CHANGELOG |
/approve |
/unhold |
cherry-pick for 1.27 #117308 |
/approve |
LGTM label has been added. Git tree hash: 320579c4fa054169c4ca9538b78cf28495a68c2c
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: dims, Huang-Wei, sanposhiho, Verolop The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
|
||
### The PreEnqueue extension point doesn't work for Pods going to activeQ through backoffQ | ||
|
||
In v1.26.0, we've found the bug that the PreEnqueue extension point doesn't work for Pods going to activeQ through backoffQ. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
1.27.0
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
oops, I'll fix in another PR 🙏
In v1.26.0, we've found the bug that the PreEnqueue extension point doesn't work for Pods going to activeQ through backoffQ. | ||
It doesn't affect any of the vanilla Kubernetes behavior, but, may break custom PreEnqueue plugins. | ||
|
||
The cause PR is [reverted](https://github.com/kubernetes/kubernetes/pull/117194) by v1.26.1. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
1.27.1
Seems unrelated. |
…of-#117194-upstream-release-1.27 Automated cherry pick of #117194: Revert "Optimization on running prePreEnqueuePlugins
What type of PR is this?
/kind bug
What this PR does / why we need it:
This reverts commit c01fa82.
The discussion: #115583 (comment)
This PR fixes that the PreEnqueue plugins aren't executed if Pods proceed to activeQ through backoffQ.
It doesn't affect the vanilla kubernetes because if the scheduling gate plugin once allow Pods to go to activeQ, it always allows them afterward. But, it may break custom PreEnqueue plugins.
This PR also adds the description of this issue as the known issue on the release note of v1.27.
Which issue(s) this PR fixes:
Fixes #
Special notes for your reviewer:
Does this PR introduce a user-facing change?
Additional documentation e.g., KEPs (Kubernetes Enhancement Proposals), usage docs, etc.: