Merge pull request #4485 from fabriziopandini/stop-updating-ClusterStatus-for-KubernetesVersionGTE-v1.22.0

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

Author: k8s-ci-robot

controlplane/kubeadm/controllers/controller.go

@@ -536,7 +536,12 @@ func (r *KubeadmControlPlaneReconciler) reconcileEtcdMembers(ctx context.Context
 		return ctrl.Result{}, errors.Wrap(err, "cannot get remote client to workload cluster")
 	}
 
-	removedMembers, err := workloadCluster.ReconcileEtcdMembers(ctx, nodeNames)
+	kubernetesVersion := controlPlane.KCP.Spec.Version
+	parsedVersion, err := semver.ParseTolerant(kubernetesVersion)
+	if err != nil {
+		return ctrl.Result{}, errors.Wrapf(err, "failed to parse kubernetes version %q", kubernetesVersion)
+	}
+	removedMembers, err := workloadCluster.ReconcileEtcdMembers(ctx, nodeNames, parsedVersion)
 	if err != nil {
 		return ctrl.Result{}, errors.Wrap(err, "failed attempt to reconcile etcd members")
 	}

controlplane/kubeadm/controllers/fakes_test.go

@@ -64,7 +64,7 @@ func (f fakeWorkloadCluster) ForwardEtcdLeadership(_ context.Context, _ *cluster
 	return nil
 }
 
-func (f fakeWorkloadCluster) ReconcileEtcdMembers(ctx context.Context, nodeNames []string) ([]string, error) {
+func (f fakeWorkloadCluster) ReconcileEtcdMembers(ctx context.Context, nodeNames []string, version semver.Version) ([]string, error) {
 	return nil, nil
 }
 
@@ -100,7 +100,7 @@ func (f fakeWorkloadCluster) RemoveEtcdMemberForMachine(ctx context.Context, mac
 	return nil
 }
 
-func (f fakeWorkloadCluster) RemoveMachineFromKubeadmConfigMap(ctx context.Context, machine *clusterv1.Machine) error {
+func (f fakeWorkloadCluster) RemoveMachineFromKubeadmConfigMap(ctx context.Context, machine *clusterv1.Machine, version semver.Version) error {
 	return nil
 }
 

controlplane/kubeadm/controllers/remediation.go

@@ -20,6 +20,7 @@ import (
 	"context"
 	"fmt"
 
+	"github.com/blang/semver"
 	"github.com/pkg/errors"
 	clusterv1 "sigs.k8s.io/cluster-api/api/v1alpha3"
 	controlplanev1 "sigs.k8s.io/cluster-api/controlplane/kubeadm/api/v1alpha3"
@@ -135,7 +136,12 @@ func (r *KubeadmControlPlaneReconciler) reconcileUnhealthyMachines(ctx context.C
 		}
 	}
 
-	if err := workloadCluster.RemoveMachineFromKubeadmConfigMap(ctx, machineToBeRemediated); err != nil {
+	kubernetesVersion := controlPlane.KCP.Spec.Version
+	parsedVersion, err := semver.ParseTolerant(kubernetesVersion)
+	if err != nil {
+		return ctrl.Result{}, errors.Wrapf(err, "failed to parse kubernetes version %q", kubernetesVersion)
+	}
+	if err := workloadCluster.RemoveMachineFromKubeadmConfigMap(ctx, machineToBeRemediated, parsedVersion); err != nil {
 		logger.Error(err, "Failed to remove machine from kubeadm ConfigMap")
 		return ctrl.Result{}, err
 	}
@@ -164,7 +170,7 @@ func (r *KubeadmControlPlaneReconciler) reconcileUnhealthyMachines(ctx context.C
 // - etc.
 //
 // NOTE: this func assumes the list of members in sync with the list of machines/nodes, it is required to call reconcileEtcdMembers
-// ans well as reconcileControlPlaneConditions before this.
+// and well as reconcileControlPlaneConditions before this.
 func (r *KubeadmControlPlaneReconciler) canSafelyRemoveEtcdMember(ctx context.Context, controlPlane *internal.ControlPlane, machineToBeRemediated *clusterv1.Machine) (bool, error) {
 	logger := r.Log.WithValues("namespace", controlPlane.KCP.Namespace, "kubeadmControlPlane", controlPlane.KCP.Name, "cluster", controlPlane.Cluster.Name)
 

controlplane/kubeadm/controllers/remediation_test.go

@@ -221,6 +221,7 @@ func TestReconcileUnhealthyMachines(t *testing.T) {
 		controlPlane := &internal.ControlPlane{
 			KCP: &controlplanev1.KubeadmControlPlane{Spec: controlplanev1.KubeadmControlPlaneSpec{
 				Replicas: utilpointer.Int32Ptr(2),
+				Version:  "v1.19.1",
 			}},
 			Cluster:  &clusterv1.Cluster{},
 			Machines: internal.NewFilterableMachineCollection(m1, m2),
@@ -270,6 +271,7 @@ func TestReconcileUnhealthyMachines(t *testing.T) {
 		controlPlane := &internal.ControlPlane{
 			KCP: &controlplanev1.KubeadmControlPlane{Spec: controlplanev1.KubeadmControlPlaneSpec{
 				Replicas: utilpointer.Int32Ptr(3),
+				Version:  "v1.19.1",
 			}},
 			Cluster:  &clusterv1.Cluster{},
 			Machines: internal.NewFilterableMachineCollection(m1, m2, m3),
@@ -320,6 +322,7 @@ func TestReconcileUnhealthyMachines(t *testing.T) {
 		controlPlane := &internal.ControlPlane{
 			KCP: &controlplanev1.KubeadmControlPlane{Spec: controlplanev1.KubeadmControlPlaneSpec{
 				Replicas: utilpointer.Int32Ptr(4),
+				Version:  "v1.19.1",
 			}},
 			Cluster:  &clusterv1.Cluster{},
 			Machines: internal.NewFilterableMachineCollection(m1, m2, m3, m4),

controlplane/kubeadm/controllers/scale.go

@@ -20,6 +20,7 @@ import (
 	"context"
 	"strings"
 
+	"github.com/blang/semver"
 	"github.com/pkg/errors"
 	corev1 "k8s.io/api/core/v1"
 	apierrors "k8s.io/apimachinery/pkg/api/errors"
@@ -128,7 +129,12 @@ func (r *KubeadmControlPlaneReconciler) scaleDownControlPlane(
 		}
 	}
 
-	if err := workloadCluster.RemoveMachineFromKubeadmConfigMap(ctx, machineToDelete); err != nil {
+	kubernetesVersion := controlPlane.KCP.Spec.Version
+	parsedVersion, err := semver.ParseTolerant(kubernetesVersion)
+	if err != nil {
+		return ctrl.Result{}, errors.Wrapf(err, "failed to parse kubernetes version %q", kubernetesVersion)
+	}
+	if err := workloadCluster.RemoveMachineFromKubeadmConfigMap(ctx, machineToDelete, parsedVersion); err != nil {
 		logger.Error(err, "Failed to remove machine from kubeadm ConfigMap")
 		return ctrl.Result{}, err
 	}

controlplane/kubeadm/controllers/scale_test.go

@@ -192,7 +192,11 @@ func TestKubeadmControlPlaneReconciler_scaleDownControlPlane_NoError(t *testing.
 		}
 
 		cluster := &clusterv1.Cluster{}
-		kcp := &controlplanev1.KubeadmControlPlane{}
+		kcp := &controlplanev1.KubeadmControlPlane{
+			Spec: controlplanev1.KubeadmControlPlaneSpec{
+				Version: "v1.19.1",
+			},
+		}
 		setKCPHealthy(kcp)
 		controlPlane := &internal.ControlPlane{
 			KCP:      kcp,
@@ -230,7 +234,11 @@ func TestKubeadmControlPlaneReconciler_scaleDownControlPlane_NoError(t *testing.
 		}
 
 		cluster := &clusterv1.Cluster{}
-		kcp := &controlplanev1.KubeadmControlPlane{}
+		kcp := &controlplanev1.KubeadmControlPlane{
+			Spec: controlplanev1.KubeadmControlPlaneSpec{
+				Version: "v1.19.1",
+			},
+		}
 		controlPlane := &internal.ControlPlane{
 			KCP:      kcp,
 			Cluster:  cluster,

controlplane/kubeadm/internal/workload_cluster.go

@@ -82,13 +82,13 @@ type WorkloadCluster interface {
 	UpdateKubeProxyImageInfo(ctx context.Context, kcp *controlplanev1.KubeadmControlPlane) error
 	UpdateCoreDNS(ctx context.Context, kcp *controlplanev1.KubeadmControlPlane) error
 	RemoveEtcdMemberForMachine(ctx context.Context, machine *clusterv1.Machine) error
-	RemoveMachineFromKubeadmConfigMap(ctx context.Context, machine *clusterv1.Machine) error
-	RemoveNodeFromKubeadmConfigMap(ctx context.Context, nodeName string) error
+	RemoveMachineFromKubeadmConfigMap(ctx context.Context, machine *clusterv1.Machine, version semver.Version) error
+	RemoveNodeFromKubeadmConfigMap(ctx context.Context, nodeName string, version semver.Version) error
 	ForwardEtcdLeadership(ctx context.Context, machine *clusterv1.Machine, leaderCandidate *clusterv1.Machine) error
 	AllowBootstrapTokensToGetNodes(ctx context.Context) error
 
 	// State recovery tasks.
-	ReconcileEtcdMembers(ctx context.Context, nodeNames []string) ([]string, error)
+	ReconcileEtcdMembers(ctx context.Context, nodeNames []string, version semver.Version) ([]string, error)
 }
 
 // Workload defines operations on workload clusters.
@@ -296,17 +296,28 @@ func (w *Workload) UpdateSchedulerInKubeadmConfigMap(ctx context.Context, schedu
 }
 
 // RemoveMachineFromKubeadmConfigMap removes the entry for the machine from the kubeadm configmap.
-func (w *Workload) RemoveMachineFromKubeadmConfigMap(ctx context.Context, machine *clusterv1.Machine) error {
+func (w *Workload) RemoveMachineFromKubeadmConfigMap(ctx context.Context, machine *clusterv1.Machine, version semver.Version) error {
 	if machine == nil || machine.Status.NodeRef == nil {
 		// Nothing to do, no node for Machine
 		return nil
 	}
 
-	return w.RemoveNodeFromKubeadmConfigMap(ctx, machine.Status.NodeRef.Name)
+	return w.RemoveNodeFromKubeadmConfigMap(ctx, machine.Status.NodeRef.Name, version)
 }
 
+var (
+	// Starting from v1.22.0 kubeadm dropped usage of the ClusterStatus entry from the kubeadm-config ConfigMap
+	// so it isn't necessary anymore to remove API endpoints for control plane nodes after deletion.
+	// NOTE: This assume kubeadm version equals to Kubernetes version.
+	minKubernetesVersionWithoutClusterStatus = semver.MustParse("1.22.0")
+)
+
 // RemoveNodeFromKubeadmConfigMap removes the entry for the node from the kubeadm configmap.
-func (w *Workload) RemoveNodeFromKubeadmConfigMap(ctx context.Context, name string) error {
+func (w *Workload) RemoveNodeFromKubeadmConfigMap(ctx context.Context, name string, version semver.Version) error {
+	if version.GTE(minKubernetesVersionWithoutClusterStatus) {
+		return nil
+	}
+
 	return util.Retry(func() (bool, error) {
 		configMapKey := ctrlclient.ObjectKey{Name: kubeadmConfigKey, Namespace: metav1.NamespaceSystem}
 		kubeadmConfigMap, err := w.getConfigMap(ctx, configMapKey)

controlplane/kubeadm/internal/workload_cluster_etcd.go

@@ -19,6 +19,7 @@ package internal
 import (
 	"context"
 
+	"github.com/blang/semver"
 	"github.com/pkg/errors"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	kerrors "k8s.io/apimachinery/pkg/util/errors"
@@ -35,7 +36,7 @@ type etcdClientFor interface {
 
 // ReconcileEtcdMembers iterates over all etcd members and finds members that do not have corresponding nodes.
 // If there are any such members, it deletes them from etcd and removes their nodes from the kubeadm configmap so that kubeadm does not run etcd health checks on them.
-func (w *Workload) ReconcileEtcdMembers(ctx context.Context, nodeNames []string) ([]string, error) {
+func (w *Workload) ReconcileEtcdMembers(ctx context.Context, nodeNames []string, version semver.Version) ([]string, error) {
 	removedMembers := []string{}
 	errs := []error{}
 	for _, nodeName := range nodeNames {
@@ -73,7 +74,7 @@ func (w *Workload) ReconcileEtcdMembers(ctx context.Context, nodeNames []string)
 				errs = append(errs, err)
 			}
 
-			if err := w.RemoveNodeFromKubeadmConfigMap(ctx, member.Name); err != nil {
+			if err := w.RemoveNodeFromKubeadmConfigMap(ctx, member.Name, version); err != nil {
 				errs = append(errs, err)
 			}
 		}

controlplane/kubeadm/internal/workload_cluster_etcd_test.go

@@ -21,6 +21,7 @@ import (
 	"errors"
 	"testing"
 
+	"github.com/blang/semver"
 	. "github.com/onsi/gomega"
 
 	"go.etcd.io/etcd/clientv3"
@@ -466,21 +467,24 @@ func TestReconcileEtcdMembers(t *testing.T) {
 			Namespace: metav1.NamespaceSystem,
 		},
 		Data: map[string]string{
-			clusterStatusKey: `apiEndpoints:
-  ip-10-0-0-1.ec2.internal:
-    advertiseAddress: 10.0.0.1
-    bindPort: 6443
-  ip-10-0-0-2.ec2.internal:
-    advertiseAddress: 10.0.0.2
-    bindPort: 6443
-    someFieldThatIsAddedInTheFuture: bar
-  ip-10-0-0-3.ec2.internal:
-    advertiseAddress: 10.0.0.3
-    bindPort: 6443
-apiVersion: kubeadm.k8s.io/vNbetaM
-kind: ClusterStatus`,
+			clusterStatusKey: "apiEndpoints:\n" +
+				"  ip-10-0-0-1.ec2.internal:\n" +
+				"    advertiseAddress: 10.0.0.1\n" +
+				"    bindPort: 6443\n" +
+				"  ip-10-0-0-2.ec2.internal:\n" +
+				"    advertiseAddress: 10.0.0.2\n" +
+				"    bindPort: 6443\n" +
+				"    someFieldThatIsAddedInTheFuture: bar\n" +
+				"  ip-10-0-0-3.ec2.internal:\n" +
+				"    advertiseAddress: 10.0.0.3\n" +
+				"    bindPort: 6443\n" +
+				"apiVersion: kubeadm.k8s.io/vNbetaM\n" +
+				"kind: ClusterStatus\n",
 		},
 	}
+	kubeadmConfigWithoutClusterStatus := kubeadmConfig.DeepCopy()
+	delete(kubeadmConfigWithoutClusterStatus.Data, clusterStatusKey)
+
 	node1 := &corev1.Node{
 		ObjectMeta: metav1.ObjectMeta{
 			Name:      "ip-10-0-0-1.ec2.internal",
@@ -508,25 +512,62 @@ kind: ClusterStatus`,
 
 	tests := []struct {
 		name                string
+		kubernetesVersion   semver.Version
 		objs                []runtime.Object
 		nodes               []string
 		etcdClientGenerator etcdClientFor
 		expectErr           bool
-		assert              func(*WithT)
+		assert              func(*WithT, client.Client)
 	}{
 		{
 			// the node to be removed is ip-10-0-0-3.ec2.internal since the
 			// other two have nodes
-			name:  "successfully removes the etcd member without a node and removes the node from kubeadm config",
-			objs:  []runtime.Object{node1.DeepCopy(), node2.DeepCopy(), kubeadmConfig.DeepCopy()},
-			nodes: []string{node1.Name, node2.Name},
+			name:              "successfully removes the etcd member without a node and removes the node from kubeadm config for Kubernetes version < 1.22.0",
+			kubernetesVersion: kubernetesVersionWithClusterStatus, // Kubernetes version < 1.22.0 has ClusterStatus
+			objs:              []runtime.Object{node1.DeepCopy(), node2.DeepCopy(), kubeadmConfig.DeepCopy()},
+			nodes:             []string{node1.Name, node2.Name},
+			etcdClientGenerator: &fakeEtcdClientGenerator{
+				forNodesClient: &etcd.Client{
+					EtcdClient: fakeEtcdClient,
+				},
+			},
+			expectErr: false,
+			assert: func(g *WithT, c client.Client) {
+				g.Expect(fakeEtcdClient.RemovedMember).To(Equal(uint64(3)))
+
+				var actualConfig corev1.ConfigMap
+				g.Expect(c.Get(
+					ctx,
+					ctrlclient.ObjectKey{Name: kubeadmConfigKey, Namespace: metav1.NamespaceSystem},
+					&actualConfig,
+				)).To(Succeed())
+				g.Expect(actualConfig.Data[clusterStatusKey]).To(Equal("apiEndpoints:\n" +
+					"  ip-10-0-0-1.ec2.internal:\n" +
+					"    advertiseAddress: 10.0.0.1\n" +
+					"    bindPort: 6443\n" +
+					"  ip-10-0-0-2.ec2.internal:\n" +
+					"    advertiseAddress: 10.0.0.2\n" +
+					"    bindPort: 6443\n" +
+					"    someFieldThatIsAddedInTheFuture: bar\n" +
+					"apiVersion: kubeadm.k8s.io/vNbetaM\n" +
+					"kind: ClusterStatus\n"))
+
+			},
+		},
+		{
+			// the node to be removed is ip-10-0-0-3.ec2.internal since the
+			// other two have nodes
+			name:              "successfully removes the etcd member without a node for Kubernetes version >= 1.22.0",
+			kubernetesVersion: minKubernetesVersionWithoutClusterStatus, // Kubernetes version >= 1.22.0 should not manage ClusterStatus
+			objs:              []runtime.Object{node1.DeepCopy(), node2.DeepCopy(), kubeadmConfigWithoutClusterStatus.DeepCopy()},
+			nodes:             []string{node1.Name, node2.Name},
 			etcdClientGenerator: &fakeEtcdClientGenerator{
 				forNodesClient: &etcd.Client{
 					EtcdClient: fakeEtcdClient,
 				},
 			},
 			expectErr: false,
-			assert: func(g *WithT) {
+			assert: func(g *WithT, c client.Client) {
 				g.Expect(fakeEtcdClient.RemovedMember).To(Equal(uint64(3)))
 			},
 		},
@@ -559,15 +600,15 @@ kind: ClusterStatus`,
 				etcdClientGenerator: tt.etcdClientGenerator,
 			}
 			ctx := context.TODO()
-			_, err := w.ReconcileEtcdMembers(ctx, tt.nodes)
+			_, err := w.ReconcileEtcdMembers(ctx, tt.nodes, tt.kubernetesVersion)
 			if tt.expectErr {
 				g.Expect(err).To(HaveOccurred())
 				return
 			}
 			g.Expect(err).ToNot(HaveOccurred())
 
 			if tt.assert != nil {
-				tt.assert(g)
+				tt.assert(g, testEnv.Client)
 			}
 		})
 	}