Merge pull request loft-sh#2168 from johannesfrey/vcluster-numeric-name

chart/templates/_helper.tpl

@@ -15,3 +15,11 @@
 {{ .repository }}:{{ .tag }}
 {{- end -}}
 {{- end -}}
+
+{{- define "vcluster.name" -}}
+{{- if regexMatch "^[0-9]+$" .Release.Name -}}
+{{- printf "vc-%s" .Release.Name -}}
+{{- else -}}
+{{- printf "%s" .Release.Name }}
+{{- end -}}
+{{- end -}}

chart/templates/headless-service.yaml

@@ -3,7 +3,7 @@
 apiVersion: v1
 kind: Service
 metadata:
-  name: {{ .Release.Name }}-headless
+  name: {{ template "vcluster.name" . }}-headless
   namespace: {{ .Release.Namespace }}
   labels:
     app: vcluster

chart/templates/service.yaml

@@ -2,7 +2,7 @@
 apiVersion: v1
 kind: Service
 metadata:
-  name: {{ .Release.Name }}
+  name: {{ template "vcluster.name" . }}
   namespace: {{ .Release.Namespace }}
   labels:
     app: vcluster
@@ -41,6 +41,6 @@ spec:
   {{- if and (not .Values.controlPlane.service.spec.selector) (not .Values.experimental.isolatedControlPlane.headless) }}
   selector:
     app: vcluster
-    release: {{ .Release.Name }}
+    release: {{ .Release.Name | quote }}
   {{- end }}
 {{- end }}

chart/templates/statefulset.yaml

@@ -2,7 +2,7 @@
 apiVersion: apps/v1
 kind: {{ include "vcluster.kind" . }}
 metadata:
-  name: {{ .Release.Name }}
+  name: {{ .Release.Name | quote }}
   namespace: {{ .Release.Namespace }}
   labels:
     app: vcluster
@@ -21,13 +21,13 @@ spec:
   selector:
     matchLabels:
       app: vcluster
-      release: {{ .Release.Name }}
+      release: {{ .Release.Name | quote }}
   {{- if eq (include "vcluster.kind" .) "StatefulSet" }}
   {{- if ge (int .Capabilities.KubeVersion.Minor) 27 }}
   persistentVolumeClaimRetentionPolicy:
     whenDeleted: {{ .Values.controlPlane.statefulSet.persistence.volumeClaim.retentionPolicy }}
   {{- end }}
-  serviceName: {{ .Release.Name }}-headless
+  serviceName: {{ template "vcluster.name" . }}-headless
   podManagementPolicy: {{ .Values.controlPlane.statefulSet.scheduling.podManagementPolicy }}
 {{ include "vcluster.persistence" . | indent 2 }}
   {{- else }}
@@ -51,7 +51,7 @@ spec:
       {{- end }}
       labels:
         app: vcluster
-        release: {{ .Release.Name }}
+        release: {{ .Release.Name | quote }}
         {{- if .Values.controlPlane.statefulSet.pods.labels }}
 {{ toYaml .Values.controlPlane.statefulSet.pods.labels | indent 8 }}
         {{- end }}
@@ -87,7 +87,7 @@ spec:
       dnsPolicy: {{ .Values.controlPlane.statefulSet.dnsPolicy }}
       {{- end }}
       {{- if .Values.controlPlane.statefulSet.dnsConfig }}
-      dnsConfig: 
+      dnsConfig:
 {{ toYaml .Values.controlPlane.statefulSet.dnsConfig | indent 8 }}
       {{- end }}
       {{- if .Values.controlPlane.advanced.serviceAccount.name }}

chart/tests/headless-service_test.yaml

@@ -61,6 +61,16 @@ tests:
           path: metadata.namespace
           value: my-namespace
 
+  - it: should prepend vc to service name when only digits are given
+    release:
+      name: 1234
+    asserts:
+      - hasDocuments:
+          count: 1
+      - equal:
+          path: metadata.name
+          value: vc-1234-headless
+
   - it: embedded-etcd
     set:
       controlPlane:

chart/tests/service_test.yaml

@@ -93,6 +93,16 @@ tests:
           path: spec.ports
           count: 2
 
+  - it: should prepend vc to service name when only digits are given
+    release:
+      name: 1234
+    asserts:
+      - hasDocuments:
+          count: 1
+      - equal:
+          path: metadata.name
+          value: vc-1234
+
   - it: isolated control plane
     release:
       name: my-release

devspace.yaml

@@ -9,6 +9,7 @@ vars:
   COMMON_VALUES: ./test/commonValues.yaml
   PRO_VALUES: ./test/proValues.yaml
   VALUES_FILE: ./test/e2e/values.yaml
+  RELEASE_NAME: "vcluster"
 
 # Images DevSpace will build for vcluster
 images:
@@ -23,7 +24,7 @@ images:
 deployments:
   vcluster-k8s:
     helm:
-      releaseName: vcluster
+      releaseName: ${RELEASE_NAME}
       chart:
         name: ./chart
       values:
@@ -51,7 +52,7 @@ deployments:
 
   vcluster-k3s:
     helm:
-      releaseName: vcluster
+      releaseName: ${RELEASE_NAME}
       chart:
         name: ./chart
       values:
@@ -79,7 +80,7 @@ deployments:
 
   vcluster-k0s:
     helm:
-      releaseName: vcluster
+      releaseName: ${RELEASE_NAME}
       chart:
         name: ./chart
       values:

pkg/certs/ensure.go

@@ -107,9 +107,7 @@ func EnsureCerts(
 
 	ownerRef := []metav1.OwnerReference{}
 	if options.Experimental.SyncSettings.SetOwner {
-		// options.ServiceName gets rewritten to the workload service name so we use options.Name as the helm chart
-		// directly uses the release name for the service name
-		controlPlaneService, err := currentNamespaceClient.CoreV1().Services(currentNamespace).Get(ctx, options.Name, metav1.GetOptions{})
+		controlPlaneService, err := currentNamespaceClient.CoreV1().Services(currentNamespace).Get(ctx, options.ControlPlaneService, metav1.GetOptions{})
 		if err != nil {
 			return fmt.Errorf("get vcluster service: %w", err)
 		}

pkg/cli/connect_helm.go

@@ -97,7 +97,7 @@ func (cmd *connectHelm) connect(ctx context.Context, vCluster *find.VCluster, co
 	}
 
 	// retrieve vcluster kube config
-	kubeConfig, err := cmd.getVClusterKubeConfig(ctx, vCluster.Name, command)
+	kubeConfig, err := cmd.getVClusterKubeConfig(ctx, vCluster, command)
 	if err != nil {
 		return err
 	}
@@ -272,15 +272,15 @@ func (cmd *connectHelm) prepare(ctx context.Context, vCluster *find.VCluster) er
 	return nil
 }
 
-func (cmd *connectHelm) getVClusterKubeConfig(ctx context.Context, vclusterName string, command []string) (*clientcmdapi.Config, error) {
+func (cmd *connectHelm) getVClusterKubeConfig(ctx context.Context, vcluster *find.VCluster, command []string) (*clientcmdapi.Config, error) {
 	var err error
 	podName := cmd.PodName
 	if podName == "" {
 		waitErr := wait.PollUntilContextTimeout(ctx, time.Second, time.Second*30, true, func(ctx context.Context) (bool, error) {
 			// get vcluster pod name
 			var pods *corev1.PodList
 			pods, err = cmd.kubeClient.CoreV1().Pods(cmd.Namespace).List(ctx, metav1.ListOptions{
-				LabelSelector: "app=vcluster,release=" + vclusterName,
+				LabelSelector: "app=vcluster,release=" + vcluster.Name,
 			})
 			if err != nil {
 				return false, err
@@ -310,7 +310,7 @@ func (cmd *connectHelm) getVClusterKubeConfig(ctx context.Context, vclusterName
 	cmd.Log.Debugf("Successfully found vCluster pod for connecting %s", podName)
 
 	// get the kube config from the Secret
-	kubeConfig, err := clihelper.GetKubeConfig(ctx, cmd.kubeClient, vclusterName, cmd.Namespace, cmd.Log)
+	kubeConfig, err := clihelper.GetKubeConfig(ctx, cmd.kubeClient, vcluster.Name, cmd.Namespace, cmd.Log)
 	if err != nil {
 		return nil, fmt.Errorf("failed to parse kube config: %w", err)
 	}
@@ -322,14 +322,14 @@ func (cmd *connectHelm) getVClusterKubeConfig(ctx context.Context, vclusterName
 	}
 
 	// exchange context name in virtual kube config
-	err = cmd.exchangeContextName(kubeConfig, vclusterName)
+	err = cmd.exchangeContextName(kubeConfig, vcluster.Name)
 	if err != nil {
 		return nil, err
 	}
 
 	// check if the vcluster is exposed and set server
-	if vclusterName != "" && cmd.Server == "" && len(command) == 0 {
-		err = cmd.setServerIfExposed(ctx, vclusterName, kubeConfig)
+	if vcluster.Name != "" && cmd.Server == "" && len(command) == 0 {
+		err = cmd.setServerIfExposed(ctx, vcluster, kubeConfig)
 		if err != nil {
 			return nil, err
 		}
@@ -338,7 +338,7 @@ func (cmd *connectHelm) getVClusterKubeConfig(ctx context.Context, vclusterName
 		if cmd.Server == "" && cmd.BackgroundProxy {
 			if localkubernetes.IsDockerInstalledAndUpAndRunning() {
 				// start background container
-				server, err := localkubernetes.CreateBackgroundProxyContainer(ctx, vclusterName, cmd.Namespace, cmd.kubeClientConfig, kubeConfig, cmd.LocalPort, cmd.Log)
+				server, err := localkubernetes.CreateBackgroundProxyContainer(ctx, vcluster.Name, cmd.Namespace, cmd.kubeClientConfig, kubeConfig, cmd.LocalPort, cmd.Log)
 				if err != nil {
 					cmd.Log.Warnf("Error exposing local vcluster, will fallback to port-forwarding: %v", err)
 					cmd.BackgroundProxy = false
@@ -419,12 +419,12 @@ func (cmd *connectHelm) getVClusterKubeConfig(ctx context.Context, vclusterName
 	return kubeConfig, nil
 }
 
-func (cmd *connectHelm) setServerIfExposed(ctx context.Context, vClusterName string, vClusterConfig *clientcmdapi.Config) error {
+func (cmd *connectHelm) setServerIfExposed(ctx context.Context, vcluster *find.VCluster, vClusterConfig *clientcmdapi.Config) error {
 	printedWaiting := false
 	err := wait.PollUntilContextTimeout(ctx, time.Second*2, time.Minute*5, true, func(ctx context.Context) (done bool, err error) {
 		// first check for load balancer service, look for the other service if it's not there
 		loadBalancerMissing := false
-		service, err := cmd.kubeClient.CoreV1().Services(cmd.Namespace).Get(ctx, vClusterName, metav1.GetOptions{})
+		service, err := cmd.kubeClient.CoreV1().Services(cmd.Namespace).Get(ctx, vcluster.ServiceName, metav1.GetOptions{})
 		if err != nil {
 			if kerrors.IsNotFound(err) {
 				loadBalancerMissing = true
@@ -433,7 +433,7 @@ func (cmd *connectHelm) setServerIfExposed(ctx context.Context, vClusterName str
 			}
 		}
 		if loadBalancerMissing {
-			service, err = cmd.kubeClient.CoreV1().Services(cmd.Namespace).Get(ctx, vClusterName, metav1.GetOptions{})
+			service, err = cmd.kubeClient.CoreV1().Services(cmd.Namespace).Get(ctx, vcluster.ServiceName, metav1.GetOptions{})
 			if kerrors.IsNotFound(err) {
 				return true, nil
 			} else if err != nil {
@@ -446,7 +446,7 @@ func (cmd *connectHelm) setServerIfExposed(ctx context.Context, vClusterName str
 
 		// not a load balancer? Then don't wait
 		if service.Spec.Type == corev1.ServiceTypeNodePort {
-			server, err := localkubernetes.ExposeLocal(ctx, vClusterName, cmd.Namespace, &cmd.rawConfig, vClusterConfig, service, cmd.LocalPort, cmd.Log)
+			server, err := localkubernetes.ExposeLocal(ctx, vcluster.Name, cmd.Namespace, &cmd.rawConfig, vClusterConfig, service, cmd.LocalPort, cmd.Log)
 			if err != nil {
 				cmd.Log.Warnf("Error exposing local vcluster, will fallback to port-forwarding: %v", err)
 			}
@@ -476,7 +476,7 @@ func (cmd *connectHelm) setServerIfExposed(ctx context.Context, vClusterName str
 			return false, nil
 		}
 
-		cmd.Log.Infof("Using vcluster %s load balancer endpoint: %s", vClusterName, cmd.Server)
+		cmd.Log.Infof("Using vcluster %s load balancer endpoint: %s", vcluster.Name, cmd.Server)
 		return true, nil
 	})
 	if err != nil {

pkg/cli/find/find.go

@@ -27,11 +27,14 @@ import (
 
 const VirtualClusterSelector = "app=vcluster"
 
+var digitsOnlyRegex = regexp.MustCompile("^[0-9]+$")
+
 type VCluster struct {
 	ClientFactory clientcmd.ClientConfig `json:"-"`
 	Created       metav1.Time
 	Name          string
 	Namespace     string
+	ServiceName   string
 	Annotations   map[string]string
 	Labels        map[string]string
 	Status        Status
@@ -399,6 +402,11 @@ func getVCluster(ctx context.Context, object client.Object, context, release str
 		status = string(StatusUnknown)
 	}
 
+	service, err := getService(ctx, client, kubeClientConfig, namespace, release)
+	if err != nil {
+		return VCluster{}, err
+	}
+
 	switch vclusterObject := object.(type) {
 	case *appsv1.StatefulSet:
 		for _, container := range vclusterObject.Spec.Template.Spec.Containers {
@@ -425,6 +433,7 @@ func getVCluster(ctx context.Context, object client.Object, context, release str
 	return VCluster{
 		Name:          release,
 		Namespace:     namespace,
+		ServiceName:   service.Name,
 		Annotations:   object.GetAnnotations(),
 		Labels:        object.GetLabels(),
 		Status:        Status(status),
@@ -435,6 +444,32 @@ func getVCluster(ctx context.Context, object client.Object, context, release str
 	}, nil
 }
 
+func getService(ctx context.Context, client *kubernetes.Clientset, kubeClientConfig clientcmd.ClientConfig, namespace, name string) (*corev1.Service, error) {
+	ctx, cancel := context.WithTimeout(ctx, time.Second*30)
+	defer cancel()
+
+	var svcName string
+	if digitsOnlyRegex.MatchString(name) {
+		svcName = fmt.Sprintf("vc-%s", name)
+	} else {
+		svcName = name
+	}
+
+	service, err := client.CoreV1().Services(namespace).Get(ctx, svcName, metav1.GetOptions{})
+	if err != nil {
+		if kerrors.IsForbidden(err) {
+			// try the current namespace instead
+			if namespace, err = getAccessibleNS(kubeClientConfig); err != nil {
+				return nil, err
+			}
+			return client.CoreV1().Services(namespace).Get(ctx, svcName, metav1.GetOptions{})
+		}
+		return nil, err
+	}
+
+	return service, nil
+}
+
 func getPods(ctx context.Context, client *kubernetes.Clientset, kubeClientConfig clientcmd.ClientConfig, namespace, podSelector string) (*corev1.PodList, error) {
 	ctx, cancel := context.WithTimeout(ctx, time.Second*30)
 	defer cancel()

pkg/config/parse.go

@@ -3,13 +3,16 @@ package config
 import (
 	"fmt"
 	"os"
+	"regexp"
 
 	"github.com/loft-sh/vcluster/config"
 	"github.com/loft-sh/vcluster/pkg/strvals"
 	"github.com/pkg/errors"
 	"sigs.k8s.io/yaml"
 )
 
+var digitsOnlyRegex = regexp.MustCompile("^[0-9]+$")
+
 func ParseConfig(path, name string, setValues []string) (*VirtualClusterConfig, error) {
 	// check if name is empty
 	if name == "" {
@@ -37,10 +40,18 @@ func ParseConfig(path, name string, setValues []string) (*VirtualClusterConfig,
 	}
 
 	// build config
+
+	var svcName string
+	if digitsOnlyRegex.MatchString(name) {
+		svcName = fmt.Sprintf("vc-%s", name)
+	} else {
+		svcName = name
+	}
+
 	retConfig := &VirtualClusterConfig{
 		Config:              *rawConfig,
 		Name:                name,
-		ControlPlaneService: name,
+		ControlPlaneService: svcName,
 	}
 
 	// validate config

pkg/controllers/resources/nodes/nodeservice/node_service.go

@@ -123,7 +123,7 @@ func (n *nodeServiceProvider) Unlock() {
 }
 
 func (n *nodeServiceProvider) GetNodeIP(ctx context.Context, name string) (string, error) {
-	serviceName := translate.SafeConcatName(translate.VClusterName, "node", strings.ReplaceAll(name, ".", "-"))
+	serviceName := translate.SafeConcatName(translate.VClusterServiceName, "node", strings.ReplaceAll(name, ".", "-"))
 
 	service := &corev1.Service{}
 	err := n.currentNamespaceClient.Get(ctx, types.NamespacedName{Name: serviceName, Namespace: n.currentNamespace}, service)

pkg/setup/config.go

@@ -26,6 +26,7 @@ func InitAndValidateConfig(ctx context.Context, vConfig *config.VirtualClusterCo
 
 	// set global vCluster name
 	translate.VClusterName = vConfig.Name
+	translate.VClusterServiceName = vConfig.WorkloadService
 
 	// set workload namespace
 	err = os.Setenv("NAMESPACE", vConfig.WorkloadNamespace)

pkg/util/translate/types.go

@@ -25,7 +25,8 @@ var (
 	NamespaceLabelPrefix = "vcluster.loft.sh/ns-label"
 
 	// VClusterName is the vcluster name, usually set at start time
-	VClusterName = "suffix"
+	VClusterName        = "suffix"
+	VClusterServiceName = "vcluster"
 
 	ManagedAnnotationsAnnotation = "vcluster.loft.sh/managed-annotations"
 	ManagedLabelsAnnotation      = "vcluster.loft.sh/managed-labels"