Bugfix: don't GC instance groups while multi-cluster ingresses exist

[skmatti]

Fixes #836

Existing GC workflow deletes instance groups as soon all GCE ingresses are deleted. But, instance groups should be retained as they are shared by multi-cluster ingresses as well.

[k8s-ci-robot]

Hi @skmatti. Thanks for your PR.

I'm waiting for a kubernetes member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

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.

[skmatti]

/assign mrhohn

[skmatti]

/assign bowei

[MrHohn]

/ok-to-test

[MrHohn]

LGTM overall, some minor comments

[MrHohn]

BTW should we fix the GC logic for L7ILB here as well? It may be as simple as tweaking these:

ingress-gce/pkg/utils/utils.go

Lines 288 to 305 in 78801b3

	// IsGCEMultiClusterIngress returns true if the given Ingress has
	// ingress.class annotation set to "gce-multi-cluster".
	func IsGCEMultiClusterIngress(ing *v1beta1.Ingress) bool {
	class := annotations.FromIngress(ing).IngressClass()
	return class == annotations.GceMultiIngressClass
	}
	// IsGCEL7ILBIngress returns true if the given Ingress has
	// ingress.class annotation set to "gce-l7-ilb"
	func IsGCEL7ILBIngress(ing *v1beta1.Ingress) bool {
	class := annotations.FromIngress(ing).IngressClass()
	return class == annotations.GceL7ILBIngressClass
	}
	// IsGLBCIngress returns true if the given Ingress should be processed by GLBC
	func IsGLBCIngress(ing *v1beta1.Ingress) bool {
	return IsGCEIngress(ing) || IsGCEMultiClusterIngress(ing)
	}

cc @spencerhance

[MrHohn]

BTW should we fix the GC logic for L7ILB here as well?

Discussed offline, seems like that works properly as is and doesn't need any fix :)

[MrHohn]

Will let @bowei also take a look.
/approve

[bowei]

Do we want to do the filter inside or outside this function?

[skmatti]

We either needed to have toKeepGceIngresses and toKeepIngresses both or handle this filter in the downstream. I think second one is straightforward. I improved the doc of both helper methods to specify that it returns a list of svc ports owned by this controller.

[bowei]

see comment above

[bowei]

you might want to sync up with rohit about why this was interface

[skmatti]

Discussed offline with Rohit and interface is not required anymore.

[skmatti]

/assign @rramkumar1

[bowei]

/lgtm
/approve

It seems strange that toCleanup is not related at all with the GC action (e.g. it is only after GC deletes the resources for a given Ingress that the finalizer is removed). It seems worthwhile to have a follow on change that makes it more explicit connection, as this will lead to bugs.

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: bowei, MrHohn, skmatti

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:

• OWNERS [MrHohn,bowei]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment