Let LB bind unschedulable node

[imroc]

What would you like to be added:
Let LB bind unschedulable node or not bind only if there is no pod of service with local externalTrafficPolicy on that node

Why is this needed:
service controller will not bind the node if it is unschedulable, see code:
https://github.com/kubernetes/kubernetes/blob/master/pkg/controller/service/service_controller.go#L605-L611

func getNodeConditionPredicate() corelisters.NodeConditionPredicate {
	return func(node *v1.Node) bool {
		// We add the master to the node list, but its unschedulable.  So we use this to filter
		// the master.
		if node.Spec.Unschedulable {
			return false
		}

It is avoid to bind the master node according to the comment, but imagine a scenario, there is a service with externalTrafficPolicy=Local, its all instances are in one node, but this node is been cordon (marked unschedulable), just to prevent the new pod from being scheduled onto it for some reason, the exist pod on that node continues to run, and service controller do not let LB bind the unschedulable node, so the service described above will be inaccessible.

[imroc]

/sig usability

[imroc]

@kubernetes/sig-usability-feature-requests

[imroc]

cc @brendandburns

[neolit123]

/sig network

[vllry]

/triage unresolved

[vllry]

/assign @andrewsykim

[andrewsykim]

Related KEP: kubernetes/enhancements#1083

Please coordinate design there

[imroc]

Related KEP: kubernetes/enhancements#1083

Please coordinate design there

Yes, this proposal can solve this problem.

[imroc]

Reopen this, reminding to change the logic after the proposal is implemented. Simply remove this in getNodeConditionPredicate():

		// We add the master to the node list, but its unschedulable.  So we use this to filter
		// the master.
		if node.Spec.Unschedulable {
			return false
		}

[andrewsykim]

/remove-triage unresolved

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[andrewsykim]

/remove-lifecycle stale

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[fejta-bot]

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten

[fejta-bot]

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close

[k8s-ci-robot]

@fejta-bot: Closing this issue.

In response to this:

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close

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.