⚠️ KCP: Stop updating and using Kubeadm's ClusterStatus with Kubernetes v1.22

[fabriziopandini]

What this PR does / why we need it:
This PR makes it possible for CAPI 0.4.x to work with Kubernetes/kubeadm v1.22.0, where the ClusterStatus field inside the Kubeadm-config has been dropped.

Which issue(s) this PR fixes:
Fixes #4423

[fabriziopandini]

/hold
I need to check CI Ferros, they are not happening locally 🤔

[sbueringer]

@fabriziopandini I'm not sure if you're aware so I thought I just mention it. Locally you're testing exactly this branch. Prow tests the commit which is created after this branch is merged into the base branch. Maybe that explains the error. It only really leads to differing results in rare cases. (but we had a few of them on-prem)

[fabriziopandini]

Agree, rebase is usually the first things to do....

[vincepri]

Could it be worth to switch to multiline strings here?

[fabriziopandini]

IMO multiline strings makes the code less readable when using yaml, because the indentation gets abruptly interrupted (and this is even worst if we are in a middle in of table tests / a series of test cases).
Also, we already did this for recent changes in KCP (e.g. all the refactor on the kubeadm config update)...

[vincepri]

Could you provide an example? I'm not sure how this is better than a straight multiline string

[fabriziopandini]

I'm happy to change if you feel this is better, but here it is an example of indentation hard to read due to multiline strings

cluster-api/controlplane/kubeadm/internal/workload_cluster_test.go

Lines 809 to 821 in 7478817

	{
	name: "updates the config map",
	scheduler: kubeadmv1beta1.ControlPlaneComponent{ExtraArgs: map[string]string{"foo": "bar"}},
	objs: []client.Object{kubeadmConfig},
	expectErr: false,
	expectedChanged: true,
	expectedScheduler: `apiVersion: kubeadm.k8s.io/v1beta2
	kind: ClusterConfiguration
	scheduler:
	extraArgs:
	foo: bar
	`,
	},

This is ( much more complex) an example without multiline strings; indentation is consisted and easier to read

cluster-api/controlplane/kubeadm/internal/workload_cluster_test.go

Lines 986 to 1018 in 5c19ea1

	{
	name: "it should set the scheduler config",
	clusterConfigurationData: "apiVersion: kubeadm.k8s.io/v1beta2\n" +
	"kind: ClusterConfiguration\n",
	newScheduler: bootstrapv1.ControlPlaneComponent{
	ExtraArgs: map[string]string{
	"bar": "baz",
	"someKey": "someVal",
	},
	ExtraVolumes: []bootstrapv1.HostPathMount{
	{
	Name: "mount2",
	HostPath: "/bar/baz",
	MountPath: "/foo/bar",
	},
	},
	},
	wantClusterConfiguration: "apiServer: {}\n" +
	"apiVersion: kubeadm.k8s.io/v1beta2\n" +
	"controllerManager: {}\n" +
	"dns: {}\n" +
	"etcd: {}\n" +
	"kind: ClusterConfiguration\n" +
	"networking: {}\n" +
	"scheduler:\n" +
	" extraArgs:\n" +
	" bar: baz\n" +
	" someKey: someVal\n" +
	" extraVolumes:\n" +
	" - hostPath: /bar/baz\n" +
	" mountPath: /foo/bar\n" +
	" name: mount2\n",
	},

[vincepri]

You could start the multi-line string with a new line and use TrimSpace https://play.golang.org/p/YShA372kY82 ?

[vincepri]

Let's tackle this in a new PR?

[vincepri]

Same as above, re:multiline strings

[vincepri]

Same as above, re:multiline strings

[vincepri]

Why is this PR marked as a breaking change?

[fabriziopandini]

Why is this PR marked as a breaking change?

I made this breaking change to raise the attention on the fact the there ClusterStatus is going away, in case some tools on top of CAPI is not paying attention to kubeadm changes.

[vincepri]

/retitle ⚠️ KCP: Stop updating and using Kubeadm's ClusterStatus with Kuberntes v1.22

[vincepri]

/milestone v0.4

[vincepri]

/approve
/lgtm

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: vincepri

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 [vincepri]

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

[vincepri]

@CecileRobertMichon over to you

[CecileRobertMichon]

I need to check CI Ferros, they are not happening locally

is the hold still in place intentionally @fabriziopandini ?

[CecileRobertMichon]

/lgtm

[fabriziopandini]

/hold cancel
Now everything is fixed

[vincepri]

/retest

[fabriziopandini]

rebased

[fabriziopandini]

/retest

[CecileRobertMichon]

/retitle ⚠️ KCP: Stop updating and using Kubeadm's ClusterStatus with Kubernetes v1.22

[CecileRobertMichon]

/lgtm