🌱 ClusterCacheTracker: fix accessor deletion on health check failure

[sbueringer]

Signed-off-by: Stefan Büringer [email protected]

What this PR does / why we need it:

Which issue(s) this PR fixes (optional, in fixes #<issue number>(, fixes #<issue_number>, ...) format, will close the issue(s) when PR gets merged):
Part of #8948 (should fix the issue for main and release-1.5)

[sbueringer]

/assign @fabriziopandini @chrischdi @killianmuldoon

I think we should have a follow-up to ensure we improve the test coverage of this critical component for cases like that workload clusters become unreachable

[sbueringer]

/cherry-pick release-1.5

This issue was introduced in the 1.5 cycle during the CR bump. So we shouldn't need it in v1.4.
I'll verify if this behavior works in v1.4

EDIT: PR to fix the other half of the issue #9028

[k8s-infra-cherrypick-robot]

@sbueringer: once the present PR merges, I will cherry-pick it on top of release-1.5 in a new PR and assign it to you.

In response to this:

/cherry-pick release-1.5

This issue was introduced in the 1.5 cycle during the CR bump. So we shouldn't need it in v1.4.
I'll verify if this behavior works in v1.4

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.

[chrischdi]

/lgtm

[k8s-ci-robot]

LGTM label has been added.

Git tree hash: de607ad529c83b17301f7729c470877bedc92a45

[sbueringer]

/assign @vincepri

[fabriziopandini]

/lgtm
nice find!

[fabriziopandini]

/approve

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: fabriziopandini

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 [fabriziopandini]

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

[killianmuldoon]

/cherry-pick release-1.5

[k8s-infra-cherrypick-robot]

@killianmuldoon: once the present PR merges, I will cherry-pick it on top of release-1.5 in a new PR and assign it to you.

In response to this:

/cherry-pick release-1.5

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.

[k8s-infra-cherrypick-robot]

@sbueringer: new pull request created: #9031

In response to this:

/cherry-pick release-1.5

This issue was introduced in the 1.5 cycle during the CR bump. So we shouldn't need it in v1.4.
I'll verify if this behavior works in v1.4

EDIT: PR to fix the other half of the issue #9028

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.

[sbueringer]

Just more context for future us. We don't need this on release-1.4 or older because there we still could precisely check for the error returned by the wait func. This changed and with >= release-1.5 we can't diferentiate between wait func timeout and client-go timeout.

1.4 code:

	err := wait.PollImmediateUntil(in.interval, runHealthCheckWithThreshold, ctx.Done())
	// An error returned implies the health check has failed a sufficient number of
	// times for the cluster to be considered unhealthy
	// NB. we are ignoring ErrWaitTimeout because this error happens when the channel is close, that in this case
	// happens when the cache is explicitly stopped.
	if err != nil && err != wait.ErrWaitTimeout {
		t.log.Error(err, "Error health checking cluster", "Cluster", klog.KRef(in.cluster.Namespace, in.cluster.Name))
		t.deleteAccessor(ctx, in.cluster)
	}

(but the old code was a bit brittle anyway so I'm happy with the new one in any case)

[killianmuldoon]

/area clustercachetracker