Reduce node events caused by zone cluster controller

On every node event, ZCC will call kube patch. Reduce it to a single
time. Before patch:

trozet@fedora:~/go/src/github.com/ovn-org/ovn-kubernetes/go-controller$ go test -mod=vendor -v ./pkg/clustermanager -ginkgo.v -ginkgo.focus=".*Node subnet allocations.*Linux nodes$" | grep -i "setting annotations"
I0218 11:18:14.168191  310203 kube.go:130] Setting annotations map[k8s.ovn.org/node-id:3] on node node1
I0218 11:18:14.168187  310203 kube.go:130] Setting annotations map[k8s.ovn.org/node-id:2] on node node2
I0218 11:18:14.168200  310203 kube.go:130] Setting annotations map[k8s.ovn.org/node-id:4] on node node3
I0218 11:18:14.168964  310203 kube.go:130] Setting annotations map[k8s.ovn.org/node-id:3] on node node1
I0218 11:18:14.169120  310203 kube.go:130] Setting annotations map[k8s.ovn.org/node-id:4] on node node3
I0218 11:18:14.169152  310203 kube.go:130] Setting annotations map[k8s.ovn.org/node-id:2] on node node2
I0218 11:18:14.169395  310203 kube.go:130] Setting annotations map[k8s.ovn.org/node-id:3] on node node1
I0218 11:18:14.169430  310203 kube.go:130] Setting annotations map[k8s.ovn.org/node-id:2] on node node2
I0218 11:18:14.169492  310203 kube.go:130] Setting annotations map[k8s.ovn.org/node-id:4] on node node3

After patch:
trozet@fedora:~/go/src/github.com/ovn-org/ovn-kubernetes/go-controller$ go test -mod=vendor -v ./pkg/clustermanager -ginkgo.v -ginkgo.focus=".*Node subnet allocations.*Linux nodes$" | grep -i "setting annotations"
I0218 11:28:16.991114  338949 kube.go:130] Setting annotations map[k8s.ovn.org/node-id:2] on node node2
I0218 11:28:16.991133  338949 kube.go:130] Setting annotations map[k8s.ovn.org/node-id:4] on node node3
I0218 11:28:16.991130  338949 kube.go:130] Setting annotations map[k8s.ovn.org/node-id:3] on node node1

Signed-off-by: Tim Rozet <[email protected]>

Author: trozet

go-controller/pkg/clustermanager/zone_cluster_controller.go

@@ -131,6 +131,23 @@ func (zcc *zoneClusterController) Stop() {
 	}
 }
 
+func needsZoneAllocation(node *corev1.Node) bool {
+	if config.HybridOverlay.Enabled && util.NoHostSubnet(node) {
+		// skip hybrid overlay nodes
+		return false
+	}
+
+	if _, ok := node.Annotations[util.OvnNodeID]; !ok {
+		return true
+	}
+	if config.OVNKubernetesFeature.EnableInterconnect {
+		if _, ok := node.Annotations[util.OvnTransitSwitchPortAddr]; !ok {
+			return true
+		}
+	}
+	return false
+}
+
 // handleAddUpdateNodeEvent handles the add or update node event
 func (zcc *zoneClusterController) handleAddUpdateNodeEvent(node *corev1.Node) error {
 	if config.HybridOverlay.Enabled && util.NoHostSubnet(node) {
@@ -226,6 +243,8 @@ type zoneClusterControllerEventHandler struct {
 	objType  reflect.Type
 	zcc      *zoneClusterController
 	syncFunc func([]interface{}) error
+
+	nodeSyncFailed sync.Map
 }
 
 func (h *zoneClusterControllerEventHandler) FilterOutResource(_ interface{}) bool {
@@ -246,9 +265,11 @@ func (h *zoneClusterControllerEventHandler) AddResource(obj interface{}, _ bool)
 			return fmt.Errorf("could not cast %T object to *corev1.Node", obj)
 		}
 		if err = h.zcc.handleAddUpdateNodeEvent(node); err != nil {
+			h.nodeSyncFailed.Store(node.Name, true)
 			return fmt.Errorf("node add failed for %s, will try again later: %w",
 				node.Name, err)
 		}
+		h.nodeSyncFailed.Delete(node.Name)
 	default:
 		return fmt.Errorf("no add function for object type %s", h.objType)
 	}
@@ -267,10 +288,16 @@ func (h *zoneClusterControllerEventHandler) UpdateResource(_, newObj interface{}
 		if !ok {
 			return fmt.Errorf("could not cast %T object to *corev1.Node", newObj)
 		}
+		_, nodeFailed := h.nodeSyncFailed.Load(node.GetName())
+		if !nodeFailed && !needsZoneAllocation(node) {
+			// node ID and transit switch IP are assigned by us and cannot change
+			return nil
+		}
 		if err = h.zcc.handleAddUpdateNodeEvent(node); err != nil {
 			return fmt.Errorf("node update failed for %s, will try again later: %w",
 				node.Name, err)
 		}
+		h.nodeSyncFailed.Delete(node.GetName())
 	default:
 		return fmt.Errorf("no update function for object type %s", h.objType)
 	}
@@ -286,7 +313,11 @@ func (h *zoneClusterControllerEventHandler) DeleteResource(obj, _ interface{}) e
 		if !ok {
 			return fmt.Errorf("could not cast obj of type %T to *knet.Node", obj)
 		}
-		return h.zcc.handleDeleteNode(node)
+		err := h.zcc.handleDeleteNode(node)
+		if err != nil {
+			return err
+		}
+		h.nodeSyncFailed.Delete(node.GetName())
 	}
 	return nil
 }

go-controller/pkg/util/node_annotations.go

@@ -129,12 +129,12 @@ const (
 	OvnNodeMigratedZoneName = "k8s.ovn.org/remote-zone-migrated"
 	/** HACK END **/
 
-	// ovnTransitSwitchPortAddr is the annotation to store the node Transit switch port ips.
+	// OvnTransitSwitchPortAddr is the annotation to store the node Transit switch port ips.
 	// It is set by cluster manager.
-	ovnTransitSwitchPortAddr = "k8s.ovn.org/node-transit-switch-port-ifaddr"
+	OvnTransitSwitchPortAddr = "k8s.ovn.org/node-transit-switch-port-ifaddr"
 
-	// ovnNodeID is the id (of type integer) of a node. It is set by cluster-manager.
-	ovnNodeID = "k8s.ovn.org/node-id"
+	// OvnNodeID is the id (of type integer) of a node. It is set by cluster-manager.
+	OvnNodeID = "k8s.ovn.org/node-id"
 
 	// InvalidNodeID indicates an invalid node id
 	InvalidNodeID = -1
@@ -646,11 +646,11 @@ func parseJoinSubnetAnnotation(nodeAnnotations map[string]string, annotationName
 // CreateNodeTransitSwitchPortAddrAnnotation creates the node annotation for the node's Transit switch port addresses.
 func CreateNodeTransitSwitchPortAddrAnnotation(nodeAnnotation map[string]interface{}, nodeIPNetv4,
 	nodeIPNetv6 *net.IPNet) (map[string]interface{}, error) {
-	return createPrimaryIfAddrAnnotation(ovnTransitSwitchPortAddr, nodeAnnotation, nodeIPNetv4, nodeIPNetv6)
+	return createPrimaryIfAddrAnnotation(OvnTransitSwitchPortAddr, nodeAnnotation, nodeIPNetv4, nodeIPNetv6)
 }
 
 func NodeTransitSwitchPortAddrAnnotationChanged(oldNode, newNode *corev1.Node) bool {
-	return oldNode.Annotations[ovnTransitSwitchPortAddr] != newNode.Annotations[ovnTransitSwitchPortAddr]
+	return oldNode.Annotations[OvnTransitSwitchPortAddr] != newNode.Annotations[OvnTransitSwitchPortAddr]
 }
 
 // CreateNodeMasqueradeSubnetAnnotation sets the IPv4 / IPv6 values of the node's Masquerade subnet.
@@ -870,7 +870,7 @@ func ParseNodeGatewayRouterJoinAddrs(node *corev1.Node, netName string) ([]*net.
 // ParseNodeTransitSwitchPortAddrs returns the IPv4 and/or IPv6 addresses for the node's transit switch port
 // stored in the 'ovnTransitSwitchPortAddr' annotation
 func ParseNodeTransitSwitchPortAddrs(node *corev1.Node) ([]*net.IPNet, error) {
-	return parsePrimaryIfAddrAnnotation(node, ovnTransitSwitchPortAddr)
+	return parsePrimaryIfAddrAnnotation(node, OvnTransitSwitchPortAddr)
 }
 
 // ParseNodeMasqueradeSubnet returns the IPv4 and/or IPv6 networks for the node's gateway router port
@@ -1228,22 +1228,22 @@ func GetSecondaryHostNetworkContainingIP(node *corev1.Node, ip net.IP) (string,
 	return match.String(), nil
 }
 
-// UpdateNodeIDAnnotation updates the ovnNodeID annotation with the node id in the annotations map
+// UpdateNodeIDAnnotation updates the OvnNodeID annotation with the node id in the annotations map
 // and returns it.
 func UpdateNodeIDAnnotation(annotations map[string]interface{}, nodeID int) map[string]interface{} {
 	if annotations == nil {
 		annotations = make(map[string]interface{})
 	}
 
-	annotations[ovnNodeID] = strconv.Itoa(nodeID)
+	annotations[OvnNodeID] = strconv.Itoa(nodeID)
 	return annotations
 }
 
-// GetNodeID returns the id of the node set in the 'ovnNodeID' node annotation.
-// Returns InvalidNodeID (-1) if the 'ovnNodeID' node annotation is not set or if the value is
+// GetNodeID returns the id of the node set in the 'OvnNodeID' node annotation.
+// Returns InvalidNodeID (-1) if the 'OvnNodeID' node annotation is not set or if the value is
 // not an integer value.
 func GetNodeID(node *corev1.Node) int {
-	nodeID, ok := node.Annotations[ovnNodeID]
+	nodeID, ok := node.Annotations[OvnNodeID]
 	if !ok {
 		return InvalidNodeID
 	}
@@ -1255,9 +1255,9 @@ func GetNodeID(node *corev1.Node) int {
 	return id
 }
 
-// NodeIDAnnotationChanged returns true if the ovnNodeID in the corev1.Nodes doesn't match
+// NodeIDAnnotationChanged returns true if the OvnNodeID in the corev1.Nodes doesn't match
 func NodeIDAnnotationChanged(oldNode, newNode *corev1.Node) bool {
-	return oldNode.Annotations[ovnNodeID] != newNode.Annotations[ovnNodeID]
+	return oldNode.Annotations[OvnNodeID] != newNode.Annotations[OvnNodeID]
 }
 
 // SetNodeZone sets the node's zone in the 'ovnNodeZoneName' node annotation.