Fix bugs (#319)

* fix arm arch

* fix bugs

Dockerfile.multiarch

@@ -1,12 +1,12 @@
-FROM --platform=$BUILDPLATFORM alpine:3.18.2
+FROM alpine:3.18.2
 
-ARG BUILDPLATFORM
+ARG TARGETPLATFORM
 
 RUN addgroup -S -g 65532 ng-user && \
     adduser -S -D -H -u 65532 \
     -s /sbin/nologin -G ng-user -g ng-user ng-user
 
-ADD bin/${BUILDPLATFORM}/controller-manager /usr/local/bin/controller-manager
-ADD bin/${BUILDPLATFORM}/autoscaler /usr/local/bin/autoscaler
-ADD bin/${BUILDPLATFORM}/scheduler /usr/local/bin/scheduler
+ADD bin/${TARGETPLATFORM}/controller-manager /usr/local/bin/controller-manager
+ADD bin/${TARGETPLATFORM}/autoscaler /usr/local/bin/autoscaler
+ADD bin/${TARGETPLATFORM}/scheduler /usr/local/bin/scheduler
 USER 65532:65532

alpine.multiarch

@@ -1,4 +1,6 @@
-FROM --platform=$BUILDPLATFORM alpine:3.18.2
+FROM alpine:3.18.2
+
+ARG TARGETPLATFORM
 
 RUN apk update \
   && apk upgrade \

apis/apps/v1alpha1/nebulacluster_common.go

@@ -182,6 +182,13 @@ func getClientCertsVolume(sslCerts *SSLCertsSpec) []corev1.Volume {
 	}
 }
 
+func rollingUpdateDone(workloadStatus WorkloadStatus) bool {
+	return workloadStatus.UpdatedReplicas == workloadStatus.Replicas &&
+		workloadStatus.ReadyReplicas == workloadStatus.Replicas &&
+		workloadStatus.CurrentReplicas == workloadStatus.UpdatedReplicas &&
+		workloadStatus.CurrentRevision == workloadStatus.UpdateRevision
+}
+
 func upgradeStatefulSet(sts *appsv1.StatefulSet) (*kruisev1alpha1.StatefulSet, error) {
 	data, err := json.Marshal(sts)
 	if err != nil {

apis/apps/v1alpha1/nebulacluster_graphd.go

@@ -129,7 +129,8 @@ func (c *graphdComponent) GetEndpoints(portName string) []string {
 }
 
 func (c *graphdComponent) IsReady() bool {
-	return *c.nc.Spec.Graphd.Replicas == c.nc.Status.Graphd.Workload.ReadyReplicas
+	return *c.nc.Spec.Graphd.Replicas == c.nc.Status.Graphd.Workload.ReadyReplicas &&
+		rollingUpdateDone(c.nc.Status.Graphd.Workload)
 }
 
 func (c *graphdComponent) GenerateLabels() map[string]string {

apis/apps/v1alpha1/nebulacluster_metad.go

@@ -146,7 +146,8 @@ func (c *metadComponent) GetEndpoints(portName string) []string {
 }
 
 func (c *metadComponent) IsReady() bool {
-	return *c.nc.Spec.Metad.Replicas == c.nc.Status.Metad.Workload.ReadyReplicas
+	return *c.nc.Spec.Metad.Replicas == c.nc.Status.Metad.Workload.ReadyReplicas &&
+		rollingUpdateDone(c.nc.Status.Metad.Workload)
 }
 
 func (c *metadComponent) GenerateLabels() map[string]string {

apis/apps/v1alpha1/nebulacluster_storaged.go

@@ -147,7 +147,8 @@ func (c *storagedComponent) GetEndpoints(portName string) []string {
 }
 
 func (c *storagedComponent) IsReady() bool {
-	return *c.nc.Spec.Storaged.Replicas == c.nc.Status.Storaged.Workload.ReadyReplicas
+	return *c.nc.Spec.Storaged.Replicas == c.nc.Status.Storaged.Workload.ReadyReplicas &&
+		rollingUpdateDone(c.nc.Status.Storaged.Workload)
 }
 
 func (c *storagedComponent) GenerateLabels() map[string]string {

apis/apps/v1alpha1/nebulacluster_types.go

@@ -89,7 +89,7 @@ type NebulaClusterSpec struct {
 	Affinity *corev1.Affinity `json:"affinity,omitempty"`
 
 	// +optional
-	Tolerations []corev1.Toleration `json:"toleration,omitempty"`
+	Tolerations []corev1.Toleration `json:"tolerations,omitempty"`
 
 	// Flag to enable/disable sidecar container nebula-agent injection, default false.
 	// +optional
@@ -153,11 +153,11 @@ type WorkloadStatus struct {
 	// +optional
 	ObservedGeneration int64 `json:"observedGeneration,omitempty"`
 
-	// The number of ready replicas.
+	// ReadyReplicas is the number of pods with a Ready Condition.
 	// +optional
 	ReadyReplicas int32 `json:"readyReplicas,omitempty"`
 
-	// Replicas is the most recently observed number of replicas.
+	// Replicas is the number of pods created by the Workload controller.
 	Replicas int32 `json:"replicas"`
 
 	// The number of pods in updatedRevision.
@@ -554,6 +554,7 @@ type ServiceSpec struct {
 // +kubebuilder:object:root=true
 // +kubebuilder:subresource:status
 // +kubebuilder:resource:shortName=nc
+// +kubebuilder:printcolumn:name="READY",type=string,JSONPath=`.status.conditions[?(@.type=="Ready")].status`
 // +kubebuilder:printcolumn:name="GRAPHD-DESIRED",type="string",JSONPath=".spec.graphd.replicas",description="The desired number of graphd pods."
 // +kubebuilder:printcolumn:name="GRAPHD-READY",type="string",JSONPath=".status.graphd.workload.readyReplicas",description="The number of graphd pods ready."
 // +kubebuilder:printcolumn:name="METAD-DESIRED",type="string",JSONPath=".spec.metad.replicas",description="The desired number of metad pods."

apis/apps/v1alpha1/template.go

@@ -36,16 +36,9 @@ var DynamicFlags = map[string]string{
 	"wal_ttl":                                     "14400",
 	"query_concurrently":                          "true",
 	"auto_remove_invalid_space":                   "true",
-	"num_io_threads":                              "16",
-	"num_worker_threads":                          "0",
-	"max_concurrent_subtasks":                     "10",
 	"snapshot_part_rate_limit":                    "10485760",
 	"snapshot_batch_size":                         "1048576",
 	"rebuild_index_part_rate_limit":               "4194304",
-	"rocksdb_db_options":                          "{}",
-	"rocksdb_column_family_options":               `{"write_buffer_size":"67108864","max_write_buffer_number":"4","max_bytes_for_level_base":"268435456"}`,
-	"rocksdb_block_based_table_options":           `{"block_size":"8192"}`,
-	"prioritize_intra_zone_reading":               "false",
 	"stick_to_intra_zone_on_failure":              "false",
 	"sync_meta_when_use_space":                    "false",
 	"validate_session_timestamp":                  "true",

charts/nebula-operator/crds/nebulaclusters.yaml

@@ -18,6 +18,9 @@ spec:
   scope: Namespaced
   versions:
   - additionalPrinterColumns:
+    - jsonPath: .status.conditions[?(@.type=="Ready")].status
+      name: READY
+      type: string
     - description: The desired number of graphd pods.
       jsonPath: .spec.graphd.replicas
       name: GRAPHD-DESIRED
@@ -11599,7 +11602,7 @@ spec:
                 type: object
               suspend:
                 type: boolean
-              toleration:
+              tolerations:
                 items:
                   properties:
                     effect:

config/crd/bases/apps.nebula-graph.io_nebulaclusters.yaml

@@ -18,6 +18,9 @@ spec:
   scope: Namespaced
   versions:
   - additionalPrinterColumns:
+    - jsonPath: .status.conditions[?(@.type=="Ready")].status
+      name: READY
+      type: string
     - description: The desired number of graphd pods.
       jsonPath: .spec.graphd.replicas
       name: GRAPHD-DESIRED
@@ -11599,7 +11602,7 @@ spec:
                 type: object
               suspend:
                 type: boolean
-              toleration:
+              tolerations:
                 items:
                   properties:
                     effect:

pkg/controller/autoscaler/autoscaler.go

@@ -592,6 +592,7 @@ type NormalizationArg struct {
 // 4. Apply the stabilization (i.e. add no more than 4 pods per minute, and pick the smallest recommendation during last 5 minutes)
 func (a *HorizontalController) normalizeDesiredReplicasWithBehaviors(hpa *v1alpha1.NebulaAutoscaler, key string, currentReplicas, prenormalizedDesiredReplicas, minReplicas int32) int32 {
 	a.maybeInitScaleDownStabilizationWindow(hpa)
+	a.maybeInitSelectPolicy(hpa)
 	normalizationArg := NormalizationArg{
 		Key:               key,
 		ScaleUpBehavior:   hpa.Spec.GraphdPolicy.Behavior.ScaleUp,
@@ -626,6 +627,20 @@ func (a *HorizontalController) maybeInitScaleDownStabilizationWindow(hpa *v1alph
 	}
 }
 
+func (a *HorizontalController) maybeInitSelectPolicy(hpa *v1alpha1.NebulaAutoscaler) {
+	behavior := hpa.Spec.GraphdPolicy.Behavior
+	if behavior != nil {
+		if behavior.ScaleUp != nil && behavior.ScaleUp.SelectPolicy == nil {
+			selectPolicy := autoscalingv2.MaxChangePolicySelect
+			hpa.Spec.GraphdPolicy.Behavior.ScaleUp.SelectPolicy = &selectPolicy
+		}
+		if behavior.ScaleDown != nil && behavior.ScaleDown.SelectPolicy == nil {
+			selectPolicy := autoscalingv2.MaxChangePolicySelect
+			hpa.Spec.GraphdPolicy.Behavior.ScaleDown.SelectPolicy = &selectPolicy
+		}
+	}
+}
+
 // getReplicasChangePerPeriod function find all the replica changes per period
 func getReplicasChangePerPeriod(periodSeconds int32, scaleEvents []timestampedScaleEvent) int32 {
 	period := time.Second * time.Duration(periodSeconds)

pkg/controller/component/helper.go

@@ -492,7 +492,6 @@ func isPending(pod *corev1.Pod) bool {
 	return pod.Status.Phase == corev1.PodPending
 }
 
-// 1. new pvc is not nil, old pvc is not nil, update storage
 func syncPVC(
 	component v1alpha1.NebulaClusterComponent,
 	pvcClient kube.PersistentVolumeClaim) error {
@@ -513,7 +512,8 @@ func syncPVC(
 			if oldPVC == nil {
 				continue
 			}
-			if volumeClaim.Spec.Resources.Requests.Storage().Cmp(*oldPVC.Spec.Resources.Requests.Storage()) != 0 {
+			if volumeClaim.Spec.Resources.Requests.Storage().Cmp(*oldPVC.Spec.Resources.Requests.Storage()) == 1 {
+				klog.Infof("expand PVC %s size to %s", pvcName, volumeClaim.Spec.Resources.Requests.Storage().String())
 				// only update storage
 				oldPVC.Spec.Resources.Requests = volumeClaim.Spec.Resources.Requests
 				if err = pvcClient.UpdatePVC(oldPVC); err != nil {

pkg/controller/component/storaged_cluster.go

@@ -169,7 +169,6 @@ func (c *storagedCluster) syncStoragedWorkload(nc *v1alpha1.NebulaCluster) error
 		nc.Status.Storaged.Phase == v1alpha1.UpdatePhase {
 		oldVolumeClaims := extender.GetDataVolumeClaims(oldWorkload)
 		newVolumeClaims := extender.GetDataVolumeClaims(newWorkload)
-		// TODO: webhook validating
 		if len(oldVolumeClaims) != len(newVolumeClaims) {
 			return fmt.Errorf("update storage data volume claims is not supported")
 		}