Merge pull request #3656 from chethanv28/topic/chethanv28/c-pick-from-c1e90f3-to-f3467b5

Cherry-pick commits from c1e90f3 to f3467b5

Author: chethanv28

manifests/supervisorcluster/1.29/cns-csi.yaml

@@ -401,7 +401,7 @@ spec:
         - name: vsphere-csi-controller
           image: localhost:5000/vmware/vsphere-csi:<vsphere_csi_ver>
           args:
-            - "--enable-profile-server=false"
+            - "--enable-profile-server=true"
           ports:
             - containerPort: 2112
               name: prometheus
@@ -473,7 +473,7 @@ spec:
             - "--leader-election-retry-period=30s"
             - "--storagequota-sync-interval=10m"
             - "--webhook-client-cert-verification"
-            - "--enable-profile-server=false"
+            - "--enable-profile-server=true"
           env:
             - name: CLUSTER_FLAVOR
               value: "WORKLOAD"

manifests/supervisorcluster/1.30/cns-csi.yaml

@@ -404,7 +404,7 @@ spec:
         - name: vsphere-csi-controller
           image: localhost:5000/vmware/vsphere-csi:<vsphere_csi_ver>
           args:
-            - "--enable-profile-server=false"
+            - "--enable-profile-server=true"
           ports:
             - containerPort: 2112
               name: prometheus
@@ -476,7 +476,7 @@ spec:
             - "--leader-election-retry-period=30s"
             - "--storagequota-sync-interval=10m"
             - "--webhook-client-cert-verification"
-            - "--enable-profile-server=false"
+            - "--enable-profile-server=true"
           env:
             - name: CLUSTER_FLAVOR
               value: "WORKLOAD"

manifests/supervisorcluster/1.31/cns-csi.yaml

@@ -404,7 +404,7 @@ spec:
         - name: vsphere-csi-controller
           image: localhost:5000/vmware/vsphere-csi:<vsphere_csi_ver>
           args:
-            - "--enable-profile-server=false"
+            - "--enable-profile-server=true"
           ports:
             - containerPort: 2112
               name: prometheus
@@ -476,7 +476,7 @@ spec:
             - "--leader-election-retry-period=30s"
             - "--storagequota-sync-interval=10m"
             - "--webhook-client-cert-verification"
-            - "--enable-profile-server=false"
+            - "--enable-profile-server=true"
           env:
             - name: CLUSTER_FLAVOR
               value: "WORKLOAD"

manifests/supervisorcluster/1.32/cns-csi.yaml

@@ -404,7 +404,7 @@ spec:
         - name: vsphere-csi-controller
           image: localhost:5000/vmware/vsphere-csi:<vsphere_csi_ver>
           args:
-            - "--enable-profile-server=false"
+            - "--enable-profile-server=true"
           ports:
             - containerPort: 2112
               name: prometheus
@@ -476,7 +476,7 @@ spec:
             - "--leader-election-retry-period=30s"
             - "--storagequota-sync-interval=10m"
             - "--webhook-client-cert-verification"
-            - "--enable-profile-server=false"
+            - "--enable-profile-server=true"
           env:
             - name: CLUSTER_FLAVOR
               value: "WORKLOAD"

pkg/csi/service/logger/logger.go

@@ -4,6 +4,7 @@ import (
 	"context"
 	"errors"
 	"fmt"
+	"sync"
 
 	"github.com/google/uuid"
 	"go.uber.org/zap"
@@ -26,7 +27,11 @@ const (
 	LogCtxIDKey = "TraceId"
 )
 
-var defaultLogLevel LogLevel
+var (
+	defaultLogLevel LogLevel
+	logger          *zap.Logger
+	loggerMutex     sync.Mutex
+)
 
 // loggerKey holds the context key used for loggers.
 type loggerKey struct{}
@@ -76,9 +81,22 @@ func withFields(ctx context.Context, fields ...zapcore.Field) context.Context {
 	return context.WithValue(ctx, loggerKey{}, getLogger(ctx).With(fields...))
 }
 
-// newLogger creates and return a new logger depending logLevel set.
+// newLogger creates and returns a logger according to the logLevel set.
+// Uses singleton pattern to create only one instance of logger.
 func newLogger() *zap.Logger {
-	var logger *zap.Logger
+	if logger != nil {
+		// logger is already initialized
+		return logger
+	}
+
+	loggerMutex.Lock()
+	defer loggerMutex.Unlock()
+	// this check ensures that multiple threads don't override the logger
+	// created by the first thread acquiring the lock.
+	if logger != nil {
+		return logger
+	}
+
 	if defaultLogLevel == DevelopmentLogLevel {
 		logger, _ = zap.NewDevelopment()
 	} else {
@@ -87,6 +105,7 @@ func newLogger() *zap.Logger {
 		loggerConfig.EncoderConfig.EncodeTime = zapcore.RFC3339NanoTimeEncoder
 		logger, _ = loggerConfig.Build()
 	}
+
 	return logger
 }
 

pkg/csi/service/vanilla/controller.go

@@ -551,7 +551,6 @@ func (c *controller) createBlockVolumeWithPlacementEngineForMultiVC(ctx context.
 	// Check if requested volume size and source snapshot size matches.
 	volumeSource := req.GetVolumeContentSource()
 	var contentSourceSnapshotID, snapshotDatastoreURL string
-	var vCenterWithTransactionSupport bool
 	if volumeSource != nil {
 		sourceSnapshot := volumeSource.GetSnapshot()
 		if sourceSnapshot == nil {
@@ -582,13 +581,6 @@ func (c *controller) createBlockVolumeWithPlacementEngineForMultiVC(ctx context.
 				"VC %q does not support snapshot operations", vCenterHost)
 		}
 
-		vCenterWithTransactionSupport, err = c.managers.VcenterManager.IsCnsTransactionSupported(ctx, vCenterHost)
-		if err != nil {
-			return nil, csifault.CSIUnimplementedFault, logger.LogNewErrorCodef(log, codes.Internal,
-				"failed to check if cns transaction APIs are supported on VC %q due to error: %v",
-				vCenterHost, err)
-		}
-
 		// Query capacity in MB and datastore url for block volume snapshot.
 		volumeIds := []cnstypes.CnsVolumeId{{Id: cnsVolumeID}}
 		cnsVolumeDetailsMap, err := utils.QueryVolumeDetailsUtil(ctx, volumeManager, volumeIds)
@@ -650,8 +642,6 @@ func (c *controller) createBlockVolumeWithPlacementEngineForMultiVC(ctx context.
 		}
 		break
 	}
-
-	isTransactionAPIsSupported := vCenterWithTransactionSupport && isCSITransactionSupportEnabled
 	volumeOperationDetails, err := operationStore.GetRequestDetails(ctx, req.Name)
 	if err != nil {
 		if apierrors.IsNotFound(err) {
@@ -661,7 +651,7 @@ func (c *controller) createBlockVolumeWithPlacementEngineForMultiVC(ctx context.
 				"error occurred while getting CreateVolume task details for block volume %q. Error: %+v",
 				req.Name, err)
 		}
-	} else if !isTransactionAPIsSupported && volumeOperationDetails.OperationDetails != nil {
+	} else if !isCSITransactionSupportEnabled && volumeOperationDetails.OperationDetails != nil {
 		if volumeOperationDetails.OperationDetails.TaskStatus ==
 			cnsvolumeoperationrequest.TaskInvocationStatusSuccess &&
 			volumeOperationDetails.VolumeID != "" {
@@ -689,7 +679,7 @@ func (c *controller) createBlockVolumeWithPlacementEngineForMultiVC(ctx context.
 				},
 			}
 			volTaskAlreadyRegistered = true
-		} else if !isTransactionAPIsSupported && cnsvolume.IsTaskPending(volumeOperationDetails) {
+		} else if !isCSITransactionSupportEnabled && cnsvolume.IsTaskPending(volumeOperationDetails) {
 			// If task is already created in CNS for this volume but task is in progress,
 			// we need to monitor the task to check if volume creation is complete or not.
 			log.Infof("Volume with name %s has CreateVolume task %s pending on VC %q.",
@@ -883,7 +873,7 @@ func (c *controller) createBlockVolumeWithPlacementEngineForMultiVC(ctx context.
 						ClusterFlavor:        cnstypes.CnsClusterFlavorVanilla,
 					},
 					common.CreateBlockVolumeOptions{
-						IsCSITransactionSupportEnabled: isTransactionAPIsSupported,
+						IsCSITransactionSupportEnabled: isCSITransactionSupportEnabled,
 					})
 				if err != nil {
 					if cnsvolume.IsNotSupportedFaultType(ctx, faultType) {
@@ -968,7 +958,7 @@ func (c *controller) createBlockVolumeWithPlacementEngineForMultiVC(ctx context.
 					ClusterFlavor:        cnstypes.CnsClusterFlavorVanilla,
 				},
 				common.CreateBlockVolumeOptions{
-					IsCSITransactionSupportEnabled: isTransactionAPIsSupported,
+					IsCSITransactionSupportEnabled: isCSITransactionSupportEnabled,
 				})
 			if err != nil {
 				if cnsvolume.IsNotSupportedFaultType(ctx, faultType) {
@@ -2570,15 +2560,9 @@ func (c *controller) CreateSnapshot(ctx context.Context, req *csi.CreateSnapshot
 		// VolumeID and SnapshotID as the input, while corresponding snapshot APIs in upstream CSI require SnapshotID.
 		// So, we need to bridge the gap in vSphere CSI driver and return a combined SnapshotID to CSI Snapshotter.
 
-		vCenterWithTransactionSupport, err := vCenterManager.IsCnsTransactionSupported(ctx, vCenterHost)
-		if err != nil {
-			return nil, logger.LogNewErrorCodef(log, codes.Internal,
-				"failed to check if cns transaction APIs are supported on VC %q due to error: %v",
-				vCenterHost, err)
-		}
 		snapshotID, cnsSnapshotInfo, err := common.CreateSnapshotUtil(ctx, volumeManager,
 			volumeID, req.Name, &cnsvolume.CreateSnapshotExtraParams{
-				IsCSITransactionSupportEnabled: isCSITransactionSupportEnabled && vCenterWithTransactionSupport,
+				IsCSITransactionSupportEnabled: isCSITransactionSupportEnabled,
 			})
 		if err != nil {
 			return nil, logger.LogNewErrorCodef(log, codes.Internal,

pkg/csi/service/wcp/controller_helper_test.go

@@ -13,11 +13,9 @@ import (
 	k8stesting "k8s.io/client-go/testing"
 	"sigs.k8s.io/vsphere-csi-driver/v3/pkg/common/unittestcommon"
 	"sigs.k8s.io/vsphere-csi-driver/v3/pkg/csi/service/common/commonco"
-	"sigs.k8s.io/vsphere-csi-driver/v3/pkg/csi/service/logger"
 )
 
 func TestGetPodVMUUID(t *testing.T) {
-	logger.SetLoggerLevel(logger.DevelopmentLogLevel)
 	containerOrchOriginal := commonco.ContainerOrchestratorUtility
 	commonco.ContainerOrchestratorUtility = &unittestcommon.FakeK8SOrchestrator{}
 	newK8sClientOriginal := newK8sClient

pkg/syncer/cnsoperator/controller/cnsfileaccessconfig/cnsfileaccessconfig_controller.go

@@ -27,6 +27,7 @@ import (
 
 	vmoperatortypes "github.com/vmware-tanzu/vm-operator/api/v1alpha5"
 	cnstypes "github.com/vmware/govmomi/cns/types"
+	vsanfstypes "github.com/vmware/govmomi/vsan/vsanfs/types"
 	v1 "k8s.io/api/core/v1"
 	apierrors "k8s.io/apimachinery/pkg/api/errors"
 	"k8s.io/apimachinery/pkg/runtime"
@@ -39,13 +40,10 @@ import (
 	"sigs.k8s.io/controller-runtime/pkg/client/config"
 	"sigs.k8s.io/controller-runtime/pkg/controller"
 	"sigs.k8s.io/controller-runtime/pkg/controller/controllerutil"
-
 	"sigs.k8s.io/controller-runtime/pkg/handler"
 	"sigs.k8s.io/controller-runtime/pkg/manager"
 	"sigs.k8s.io/controller-runtime/pkg/reconcile"
 	"sigs.k8s.io/controller-runtime/pkg/source"
-
-	vsanfstypes "github.com/vmware/govmomi/vsan/vsanfs/types"
 	cnsoperatorapis "sigs.k8s.io/vsphere-csi-driver/v3/pkg/apis/cnsoperator"
 	cnsfileaccessconfigv1alpha1 "sigs.k8s.io/vsphere-csi-driver/v3/pkg/apis/cnsoperator/cnsfileaccessconfig/v1alpha1"
 	volumes "sigs.k8s.io/vsphere-csi-driver/v3/pkg/common/cns-lib/volume"
@@ -57,13 +55,14 @@ import (
 	k8s "sigs.k8s.io/vsphere-csi-driver/v3/pkg/kubernetes"
 	"sigs.k8s.io/vsphere-csi-driver/v3/pkg/syncer"
 	cnsoperatortypes "sigs.k8s.io/vsphere-csi-driver/v3/pkg/syncer/cnsoperator/types"
-	cnsoperatorutil "sigs.k8s.io/vsphere-csi-driver/v3/pkg/syncer/cnsoperator/util"
+	"sigs.k8s.io/vsphere-csi-driver/v3/pkg/syncer/cnsoperator/util"
 )
 
 const (
-	defaultMaxWorkerThreadsForFileAccessConfig = 10
-	capvVmLabelKey                             = "capv.vmware.com"
-	devopsUserLabelKey                         = "cns.vmware.com/user-created"
+	workerThreadsEnvVar     = "WORKER_THREADS_CNS_FILE_ACCESS_CONFIG"
+	defaultMaxWorkerThreads = 10
+	capvVmLabelKey          = "capv.vmware.com"
+	devopsUserLabelKey      = "cns.vmware.com/user-created"
 )
 
 // backOffDuration is a map of cnsfileaccessconfig name's to the time after
@@ -87,18 +86,16 @@ func Add(mgr manager.Manager, clusterFlavor cnstypes.CnsClusterFlavor,
 		return nil
 	}
 
-	if clusterFlavor == cnstypes.CnsClusterFlavorWorkload {
-		if commonco.ContainerOrchestratorUtility.IsFSSEnabled(ctx, common.TKGsHA) {
-			clusterComputeResourceMoIds, _, err := common.GetClusterComputeResourceMoIds(ctx)
-			if err != nil {
-				log.Errorf("failed to get clusterComputeResourceMoIds. err: %v", err)
-				return err
-			}
-			if len(clusterComputeResourceMoIds) > 1 &&
-				!commonco.ContainerOrchestratorUtility.IsFSSEnabled(ctx, common.WorkloadDomainIsolation) {
-				log.Infof("Not initializing the CnsFileAccessConfig Controller as stretched supervisor is detected.")
-				return nil
-			}
+	if commonco.ContainerOrchestratorUtility.IsFSSEnabled(ctx, common.TKGsHA) {
+		clusterComputeResourceMoIds, _, err := common.GetClusterComputeResourceMoIds(ctx)
+		if err != nil {
+			log.Errorf("failed to get clusterComputeResourceMoIds. err: %v", err)
+			return err
+		}
+		if len(clusterComputeResourceMoIds) > 1 &&
+			!commonco.ContainerOrchestratorUtility.IsFSSEnabled(ctx, common.WorkloadDomainIsolation) {
+			log.Infof("Not initializing the CnsFileAccessConfig Controller as stretched supervisor is detected.")
+			return nil
 		}
 	}
 	volumePermissionLockMap = &sync.Map{}
@@ -164,17 +161,23 @@ func Add(mgr manager.Manager, clusterFlavor cnstypes.CnsClusterFlavor,
 func newReconciler(mgr manager.Manager, configInfo *commonconfig.ConfigurationInfo,
 	volumeManager volumes.Manager, vmOperatorClient client.Client, dynamicClient dynamic.Interface,
 	recorder record.EventRecorder) reconcile.Reconciler {
-	return &ReconcileCnsFileAccessConfig{client: mgr.GetClient(), scheme: mgr.GetScheme(),
-		configInfo: configInfo, volumeManager: volumeManager, vmOperatorClient: vmOperatorClient,
-		dynamicClient: dynamicClient, recorder: recorder}
+	return &ReconcileCnsFileAccessConfig{
+		client:           mgr.GetClient(),
+		scheme:           mgr.GetScheme(),
+		configInfo:       configInfo,
+		volumeManager:    volumeManager,
+		vmOperatorClient: vmOperatorClient,
+		dynamicClient:    dynamicClient,
+		recorder:         recorder,
+	}
 }
 
 // add adds a new Controller to mgr with r as the reconcile.Reconciler.
 func add(mgr manager.Manager, r reconcile.Reconciler) error {
 	ctx, log := logger.GetNewContextWithLogger()
 
-	maxWorkerThreads := getMaxWorkerThreadsToReconcileCnsFileAccessConfig(ctx)
-
+	maxWorkerThreads := util.GetMaxWorkerThreads(ctx,
+		workerThreadsEnvVar, defaultMaxWorkerThreads)
 	// Create a new controller.
 	c, err := controller.New("cnsfileaccessconfig-controller", mgr,
 		controller.Options{Reconciler: r, MaxConcurrentReconciles: maxWorkerThreads})
@@ -261,7 +264,7 @@ func (r *ReconcileCnsFileAccessConfig) Reconcile(ctx context.Context,
 				instance.Name, instance.Spec.VMName)
 			// Fetch the PVC and PV instance and get volume ID
 			skipConfigureVolumeACL := false
-			volumeID, err := cnsoperatorutil.GetVolumeID(ctx, r.client, instance.Spec.PvcName, instance.Namespace)
+			volumeID, err := util.GetVolumeID(ctx, r.client, instance.Spec.PvcName, instance.Namespace)
 			if err != nil {
 				if apierrors.IsNotFound(err) {
 					// If PVC instance is NotFound (deleted), then there is no need to configure ACL on file volume.
@@ -330,7 +333,7 @@ func (r *ReconcileCnsFileAccessConfig) Reconcile(ctx context.Context,
 	if instance.DeletionTimestamp != nil {
 		log.Infof("CnsFileAccessConfig instance %q has deletion timestamp set", instance.Name)
 		volumeExists := true
-		volumeID, err := cnsoperatorutil.GetVolumeID(ctx, r.client, instance.Spec.PvcName, instance.Namespace)
+		volumeID, err := util.GetVolumeID(ctx, r.client, instance.Spec.PvcName, instance.Namespace)
 		if err != nil {
 			if ifFileVolumesWithVmserviceVmsSupported &&
 				apierrors.IsNotFound(err) {
@@ -447,7 +450,7 @@ func (r *ReconcileCnsFileAccessConfig) Reconcile(ctx context.Context,
 	log.Infof("Reconciling CnsFileAccessConfig with instance: %q from namespace: %q. timeout %q seconds",
 		instance.Name, instance.Namespace, timeout)
 	if !instance.Status.Done {
-		volumeID, err := cnsoperatorutil.GetVolumeID(ctx, r.client, instance.Spec.PvcName, instance.Namespace)
+		volumeID, err := util.GetVolumeID(ctx, r.client, instance.Spec.PvcName, instance.Namespace)
 		if err != nil {
 			msg := fmt.Sprintf("Failed to get volumeID from pvcName: %q. Error: %+v", instance.Spec.PvcName, err)
 			log.Error(msg)
@@ -780,7 +783,7 @@ func (r *ReconcileCnsFileAccessConfig) configureVolumeACLs(ctx context.Context,
 func (r *ReconcileCnsFileAccessConfig) getVMExternalIP(ctx context.Context,
 	vm *vmoperatortypes.VirtualMachine) (string, error) {
 	log := logger.GetLogger(ctx)
-	networkProvider, err := cnsoperatorutil.GetNetworkProvider(ctx)
+	networkProvider, err := util.GetNetworkProvider(ctx)
 	if err != nil {
 		return "", logger.LogNewErrorf(log, "Failed to identify the network provider. Error: %+v", err)
 	}
@@ -789,11 +792,11 @@ func (r *ReconcileCnsFileAccessConfig) getVMExternalIP(ctx context.Context,
 		return "", logger.LogNewError(log, "unable to find network provider information")
 	}
 
-	networkTypes := []string{cnsoperatorutil.NSXTNetworkProvider, cnsoperatorutil.
+	networkTypes := []string{util.NSXTNetworkProvider, util.
 		VDSNetworkProvider}
 
 	if commonco.ContainerOrchestratorUtility.IsFSSEnabled(ctx, common.VPCCapabilitySupervisor) {
-		networkTypes = append(networkTypes, cnsoperatorutil.VPCNetworkProvider)
+		networkTypes = append(networkTypes, util.VPCNetworkProvider)
 	}
 
 	supported_found := false
@@ -808,7 +811,7 @@ func (r *ReconcileCnsFileAccessConfig) getVMExternalIP(ctx context.Context,
 		return "", logger.LogNewErrorf(log, "Unknown network provider. Error: %+v", err)
 	}
 
-	tkgVMIP, err := cnsoperatorutil.GetTKGVMIP(ctx, r.vmOperatorClient,
+	tkgVMIP, err := util.GetTKGVMIP(ctx, r.vmOperatorClient,
 		r.dynamicClient, vm.Namespace, vm.Name, networkProvider)
 	if err != nil {
 		return "", logger.LogNewErrorf(log, "Failed to get external facing IP address for VM %q/%q. Err: %+v",