-
Notifications
You must be signed in to change notification settings - Fork 38.7k
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
KCCM: fix GCP ILB by reintroducing readiness predicate for eTP:Local #121116
KCCM: fix GCP ILB by reintroducing readiness predicate for eTP:Local #121116
Conversation
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. |
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.
We really don't want readiness to exclude nodes in case of etp=Local, but the current non-determinism is worse.
/approve
/lgtm
@@ -1002,6 +1002,7 @@ var ( | |||
etpLocalNodePredicates []NodeConditionPredicate = []NodeConditionPredicate{ | |||
nodeIncludedPredicate, | |||
nodeUnTaintedPredicate, | |||
nodeReadyPredicate, |
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.
Is the plan to backport this?
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.
Yes!
@@ -1002,6 +1002,7 @@ var ( | |||
etpLocalNodePredicates []NodeConditionPredicate = []NodeConditionPredicate{ |
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.
allNodePredicates
and etpLocalNodePredicates
are identical now. If I understand, we could just merge them, but since this is all replaced by stableNodeSetPredicates
(which is beta) it's moot, and a smaller delta is preferable, right?
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.
Yes, this was the idea.
LGTM label has been added. Git tree hash: 63ae40e8ca669c8d03f80441fd930d1f19e1974c
|
/retest |
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: alexanderConstantinescu, thockin 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 |
/cherry-pick release-1.28 |
…rry-pick-of-#121116-upstream-release-1.28 Automated cherry pick of #121116: KCCM: fix GCP ILB by reintroducing readiness predicate for
…rry-pick-of-#121116-upstream-release-1.27 Automated cherry pick of #121116: KCCM: fix GCP ILB by reintroducing readiness predicate for
…rry-pick-of-#121116-upstream-release-1.26 Automated cherry pick of #121116: KCCM: fix GCP ILB by reintroducing readiness predicate for
What type of PR is this?
/kind bug
What this PR does / why we need it:
As mentioned in the linked issue: service exposed through GCP ILBs might have their SLOs impacted as a consequence of applying a different set of predicates to different services (eTP: Cluster/Local), since all load balancers point to the same InstanceGroup. The fix here is therefore that we re-introduce the readiness predicate for eTP:Local services so that the predicates align across all classes of services. We can't do the inverse and remove the readiness predicate for eTP:Cluster, because that's the KEP-3458.
/sig network
/sig cloud-provider
/assign @thockin
/cc @aojea
This PR should have no effect on >= 1.27 since the predicates are already aligned under the feature gate
StableLoadBalancerNodeSet
, but this should get in so that we guard against the broken behaviour should someone turn that feature gate off.Which issue(s) this PR fixes:
Fixes #121094
Special notes for your reviewer:
Does this PR introduce a user-facing change?
Additional documentation e.g., KEPs (Kubernetes Enhancement Proposals), usage docs, etc.: