[client, proxy] cancel context before stopping engine on embedded client

This fixes an issue where L4 service update is not possible when proxy clusters don't support custom ports

.github/workflows/golang-test-linux.yml

@@ -158,7 +158,7 @@ jobs:
         run: git --no-pager diff --exit-code
 
       - name: Test
-        run: CGO_ENABLED=1 GOARCH=${{ matrix.arch }} CI=true go test -race -coverprofile=coverage.txt -tags devcert -exec 'sudo' -timeout 10m -p 1 $(go list ./... | grep -v -e /management -e /signal -e /relay -e /proxy -e /combined)
+        run: CGO_ENABLED=1 GOARCH=${{ matrix.arch }} CI=true go test -coverprofile=coverage.txt -tags devcert -exec 'sudo' -timeout 10m -p 1 $(go list ./... | grep -v -e /management -e /signal -e /relay -e /proxy -e /combined)
 
       - name: Upload coverage reports to Codecov
         if: matrix.arch == 'amd64'
@@ -478,7 +478,7 @@ jobs:
           CGO_ENABLED=1 GOARCH=${{ matrix.arch }} \
           NETBIRD_STORE_ENGINE=${{ matrix.store }} \
             CI=true \
-          go test -race -tags=devcert -coverprofile=coverage.txt \
+          go test -tags=devcert -coverprofile=coverage.txt \
             -exec "sudo --preserve-env=CI,NETBIRD_STORE_ENGINE" \
             -timeout 20m ./management/... ./shared/management/...
 

client/cmd/kubernetes.go

@@ -0,0 +1,301 @@
+package cmd
+
+import (
+	"context"
+	"crypto/tls"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"io"
+	"net"
+	"net/http"
+	"net/url"
+	"os"
+	"path/filepath"
+	"slices"
+	"strings"
+
+	"github.com/goccy/go-yaml"
+	log "github.com/sirupsen/logrus"
+	"github.com/spf13/cobra"
+
+	"github.com/netbirdio/netbird/client/proto"
+)
+
+const (
+	KubernetesDNSSuffix = "netbird-kubeapi-proxy"
+)
+
+var kubernetesCmd = &cobra.Command{
+	Use:   "kubernetes",
+	Short: "Kubernetes cluster commands.",
+	Long:  "Kubernetes cluster commands.",
+}
+
+var kubernetesListCmd = &cobra.Command{
+	Use:   "list",
+	RunE:  kubernetesList,
+	Short: "List Kubernetes clusters.",
+	Long:  "List Kubernetes clusters by discovering NetBird peers running netbird-kubeapi-proxy.",
+}
+
+var kubernetesWriteKubeconfigCmd = &cobra.Command{
+	Use:   "write-kubeconfig",
+	RunE:  kubernetesWriteKubeconfig,
+	Args:  cobra.ExactArgs(1),
+	Short: "Write kubeconfig for a Kubernetes cluster.",
+	Long:  "Updates kubeconfig in place to allow token-less access to the Kubernetes cluster through NetBird.",
+}
+
+func init() {
+	kubernetesWriteKubeconfigCmd.Flags().String("kubeconfig", "", "path to kubeconfig file")
+}
+
+func kubernetesList(cmd *cobra.Command, _ []string) error {
+	conn, err := getClient(cmd)
+	if err != nil {
+		return err
+	}
+	defer conn.Close()
+	client := proto.NewDaemonServiceClient(conn)
+	statusResp, err := client.Status(cmd.Context(), &proto.StatusRequest{GetFullPeerStatus: true})
+	if err != nil {
+		return err
+	}
+
+	kcs, err := getKubernetesClusters(cmd.Context(), statusResp.FullStatus.Peers, "")
+	if err != nil {
+		return err
+	}
+	if len(kcs) == 0 {
+		cmd.Println("No Kubernetes clusters available.")
+		return nil
+	}
+	cmd.Println("Available Kubernetes clusters:")
+	for _, k := range kcs {
+		cmd.Printf("\n  - Name: %s\n    FQDN: %s\n    Version: %s\n", k.name, k.url.Host, k.version)
+	}
+	return nil
+}
+
+func kubernetesWriteKubeconfig(cmd *cobra.Command, args []string) error {
+	kubeconfigPath, err := resolveKubeconfigPath(cmd)
+	if err != nil {
+		return err
+	}
+
+	conn, err := getClient(cmd)
+	if err != nil {
+		return err
+	}
+	defer conn.Close()
+	client := proto.NewDaemonServiceClient(conn)
+	statusResp, err := client.Status(cmd.Context(), &proto.StatusRequest{GetFullPeerStatus: true})
+	if err != nil {
+		return err
+	}
+
+	clusterName := args[0]
+	kcs, err := getKubernetesClusters(cmd.Context(), statusResp.FullStatus.Peers, clusterName)
+	if err != nil {
+		return err
+	}
+	if len(kcs) == 0 {
+		return fmt.Errorf("kubernetes cluster named %s not found", clusterName)
+	}
+	if len(kcs) > 1 {
+		return fmt.Errorf("too many Kubernetes clusters returned")
+	}
+	err = writeKubeconfig(kubeconfigPath, kcs[0])
+	if err != nil {
+		return err
+	}
+	return nil
+}
+
+type kubernetesCluster struct {
+	name    string
+	url     *url.URL
+	version string
+}
+
+func getKubernetesClusters(ctx context.Context, peers []*proto.PeerState, nameFilter string) ([]kubernetesCluster, error) {
+	transport := http.DefaultTransport.(*http.Transport).Clone()
+	transport.TLSClientConfig = &tls.Config{
+		InsecureSkipVerify: true,
+	}
+	httpClient := &http.Client{
+		Transport: transport,
+	}
+	resolver := net.Resolver{
+		// Required so both DNS records are returned.
+		// https://github.com/golang/go/issues/17093
+		PreferGo: true,
+	}
+
+	kcs := []kubernetesCluster{}
+	attempted := map[string]struct{}{}
+	for _, peer := range peers {
+		fqdns, err := resolver.LookupAddr(ctx, peer.IP)
+		if err != nil {
+			return nil, err
+		}
+		for _, fqdn := range fqdns {
+			if _, ok := attempted[fqdn]; ok {
+				continue
+			}
+			attempted[fqdn] = struct{}{}
+			comps := strings.Split(fqdn, ".")
+			if len(comps) < 2 {
+				continue
+			}
+			if comps[1] != KubernetesDNSSuffix {
+				continue
+			}
+			if nameFilter != "" && nameFilter != comps[0] {
+				continue
+			}
+			clusterURL, clusterVersion, err := fingerprintClusters(ctx, httpClient, fqdn)
+			if err != nil {
+				log.Debugf("could not fingerprint Kubernetes cluster %s %q", fqdn, err)
+				continue
+			}
+			kc := kubernetesCluster{
+				name:    comps[0],
+				url:     clusterURL,
+				version: clusterVersion,
+			}
+			if nameFilter != "" {
+				return []kubernetesCluster{kc}, nil
+			}
+			kcs = append(kcs, kc)
+		}
+	}
+	return kcs, nil
+}
+
+func fingerprintClusters(ctx context.Context, httpClient *http.Client, fqdn string) (*url.URL, string, error) {
+	clusterURL, err := url.Parse("https://" + fqdn)
+	if err != nil {
+		return nil, "", err
+	}
+	versionURL, err := clusterURL.Parse("/version")
+	if err != nil {
+		return nil, "", err
+	}
+	req, err := http.NewRequestWithContext(ctx, http.MethodGet, versionURL.String(), nil)
+	if err != nil {
+		return nil, "", err
+	}
+	resp, err := httpClient.Do(req)
+	if err != nil {
+		return nil, "", err
+	}
+	defer resp.Body.Close()
+	if resp.StatusCode != http.StatusOK {
+		return nil, "", fmt.Errorf("expected %d response but got %s", http.StatusOK, resp.Status)
+	}
+	b, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return nil, "", err
+	}
+	versionData := map[string]string{}
+	err = json.Unmarshal(b, &versionData)
+	if err != nil {
+		return nil, "", err
+	}
+	version, ok := versionData["gitVersion"]
+	if !ok {
+		return nil, "", errors.New("no version found in response")
+	}
+	return clusterURL, version, nil
+}
+
+func resolveKubeconfigPath(cmd *cobra.Command) (string, error) {
+	if cmd.Flags().Changed("kubeconfig") {
+		path, err := cmd.Flags().GetString("kubeconfig")
+		if err != nil {
+			return "", err
+		}
+		return path, nil
+	}
+	if env := os.Getenv("KUBECONFIG"); env != "" {
+		return env, nil
+	}
+	home, err := os.UserHomeDir()
+	if err != nil {
+		return "", fmt.Errorf("could not determine home directory: %w", err)
+	}
+	return filepath.Join(home, ".kube", "config"), nil
+}
+
+func writeKubeconfig(kubeconfigPath string, kc kubernetesCluster) error {
+	b, err := os.ReadFile(kubeconfigPath)
+	if err != nil && !errors.Is(err, os.ErrNotExist) {
+		return err
+	}
+	var cfg map[string]any
+	if err := yaml.Unmarshal(b, &cfg); err != nil {
+		return err
+	}
+	if cfg == nil {
+		cfg = map[string]any{
+			"apiVersion": "v1",
+			"kind":       "Config",
+		}
+	}
+
+	cfg["clusters"] = appendWithName(cfg["clusters"], map[string]any{
+		"name": kc.name,
+		"cluster": map[string]any{
+			"server":                   kc.url.String(),
+			"insecure-skip-tls-verify": true,
+		},
+	})
+	cfg["users"] = appendWithName(cfg["users"], map[string]any{
+		"name": "netbird",
+		"user": map[string]any{
+			"token": "none",
+		},
+	})
+	cfg["contexts"] = appendWithName(cfg["contexts"], map[string]any{
+		"name": kc.name,
+		"context": map[string]any{
+			"cluster":   kc.name,
+			"user":      "netbird",
+			"namespace": "default",
+		},
+	})
+	cfg["current-context"] = kc.name
+
+	out, err := yaml.Marshal(cfg)
+	if err != nil {
+		return err
+	}
+	if err := os.WriteFile(kubeconfigPath, out, 0o600); err != nil {
+		return err
+	}
+	return nil
+}
+
+func appendWithName(data any, add map[string]any) any {
+	if data == nil {
+		return []any{add}
+	}
+	v, ok := data.([]any)
+	if !ok {
+		return []any{add}
+	}
+	i := slices.IndexFunc(v, func(item any) bool {
+		m, ok := item.(map[string]any)
+		if !ok {
+			return false
+		}
+		return m["name"] == add["name"]
+	})
+	if i == -1 {
+		return append(v, add)
+	}
+	v[i] = add
+	return v
+}

client/cmd/kubernetes_test.go

@@ -0,0 +1,120 @@
+package cmd
+
+import (
+	"net/http"
+	"net/http/httptest"
+	"net/url"
+	"os"
+	"path/filepath"
+	"testing"
+
+	"github.com/spf13/cobra"
+	"github.com/stretchr/testify/require"
+)
+
+func TestFingerprintClusters(t *testing.T) {
+	t.Parallel()
+
+	srv := httptest.NewTLSServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		//nolint: errcheck
+		w.Write([]byte(`{"gitVersion": "foobar"}`))
+	}))
+	defer srv.Close()
+
+	clusterURL, clusterVersion, err := fingerprintClusters(t.Context(), srv.Client(), srv.Listener.Addr().String())
+	require.NoError(t, err)
+	require.Equal(t, srv.URL, clusterURL.String())
+	require.Equal(t, "foobar", clusterVersion)
+}
+
+func TestResolveKubeconfigPath(t *testing.T) {
+	home, err := os.UserHomeDir()
+	if err != nil {
+		t.Fatalf("could not determine home directory: %v", err)
+	}
+	defaultPath := filepath.Join(home, ".kube", "config")
+	path, err := resolveKubeconfigPath(&cobra.Command{})
+	require.NoError(t, err)
+	require.Equal(t, defaultPath, path)
+
+	flagPath := "flag-path"
+	cmd := &cobra.Command{}
+	cmd.Flags().String("kubeconfig", "", "")
+	err = cmd.Flags().Set("kubeconfig", flagPath)
+	require.NoError(t, err)
+	path, err = resolveKubeconfigPath(cmd)
+	require.NoError(t, err)
+	require.Equal(t, flagPath, path)
+
+	envPath := "env-path"
+	t.Setenv("KUBECONFIG", envPath)
+	path, err = resolveKubeconfigPath(&cobra.Command{})
+	require.NoError(t, err)
+	require.Equal(t, envPath, path)
+}
+
+func TestWriteKubeconfig(t *testing.T) {
+	t.Parallel()
+
+	tests := []struct {
+		name     string
+		existing string
+	}{
+		{
+			name: "empty file",
+		},
+		{
+			name: "existing content",
+			existing: `apiVersion: v1
+clusters:
+- cluster:
+    insecure-skip-tls-verify: true
+    server: https://foobar.com
+  name: foo
+current-context: test
+kind: Config
+users: []
+`,
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			t.Parallel()
+
+			kubeconfigPath := filepath.Join(t.TempDir(), "config")
+			err := os.WriteFile(kubeconfigPath, []byte(tt.existing), 0o644)
+			require.NoError(t, err)
+
+			kc := kubernetesCluster{
+				name: "foo",
+				url:  &url.URL{Scheme: "https", Host: "example.com"},
+			}
+			err = writeKubeconfig(kubeconfigPath, kc)
+			require.NoError(t, err)
+
+			b, err := os.ReadFile(kubeconfigPath)
+			require.NoError(t, err)
+			expected := `apiVersion: v1
+clusters:
+- cluster:
+    insecure-skip-tls-verify: true
+    server: https://example.com
+  name: foo
+contexts:
+- context:
+    cluster: foo
+    namespace: default
+    user: netbird
+  name: foo
+current-context: foo
+kind: Config
+users:
+- name: netbird
+  user:
+    token: none
+`
+			require.Equal(t, expected, string(b))
+		})
+	}
+
+}

client/cmd/root.go

@@ -169,6 +169,11 @@ func init() {
 	debugCmd.AddCommand(forCmd)
 	debugCmd.AddCommand(persistenceCmd)
 
+	// kubernetes commands
+	rootCmd.AddCommand(kubernetesCmd)
+	kubernetesCmd.AddCommand(kubernetesListCmd)
+	kubernetesCmd.AddCommand(kubernetesWriteKubeconfigCmd)
+
 	// profile commands
 	profileCmd.AddCommand(profileListCmd)
 	profileCmd.AddCommand(profileAddCmd)

client/embed/embed.go

@@ -279,6 +279,10 @@ func (c *Client) Start(startCtx context.Context) error {
 
 	select {
 	case <-startCtx.Done():
+		// Cancel the client context before stopping: Engine.Start blocks on the
+		// signal stream while holding the engine mutex and only unblocks on
+		// cancellation. Stopping first would deadlock on that mutex.
+		cancel()
 		if stopErr := client.Stop(); stopErr != nil {
 			return fmt.Errorf("stop error after context done. Stop error: %w. Context done: %w", stopErr, startCtx.Err())
 		}

client/embed/embed_test.go

@@ -0,0 +1,168 @@
+package embed
+
+import (
+	"context"
+	"net"
+	"testing"
+	"time"
+
+	"github.com/golang/mock/gomock"
+	"github.com/stretchr/testify/require"
+	"google.golang.org/grpc"
+
+	"github.com/netbirdio/netbird/management/internals/controllers/network_map/controller"
+	"github.com/netbirdio/netbird/management/internals/controllers/network_map/update_channel"
+	"github.com/netbirdio/netbird/management/internals/modules/peers"
+	"github.com/netbirdio/netbird/management/internals/modules/peers/ephemeral/manager"
+	"github.com/netbirdio/netbird/management/internals/server/config"
+	nbgrpc "github.com/netbirdio/netbird/management/internals/shared/grpc"
+	mgmt "github.com/netbirdio/netbird/management/server"
+	"github.com/netbirdio/netbird/management/server/activity"
+	nbcache "github.com/netbirdio/netbird/management/server/cache"
+	"github.com/netbirdio/netbird/management/server/groups"
+	"github.com/netbirdio/netbird/management/server/integrations/integrated_validator/validator"
+	"github.com/netbirdio/netbird/management/server/integrations/port_forwarding"
+	"github.com/netbirdio/netbird/management/server/job"
+	"github.com/netbirdio/netbird/management/server/permissions"
+	"github.com/netbirdio/netbird/management/server/settings"
+	"github.com/netbirdio/netbird/management/server/store"
+	"github.com/netbirdio/netbird/management/server/telemetry"
+	"github.com/netbirdio/netbird/management/server/types"
+	mgmtProto "github.com/netbirdio/netbird/shared/management/proto"
+	"github.com/netbirdio/netbird/util"
+)
+
+const testSetupKey = "A2C8E62B-38F5-4553-B31E-DD66C696CEBB"
+
+// TestClientStartTimeoutRollback reproduces a deadlock between Engine.Start and
+// Engine.Stop. The signal endpoint accepts gRPC connections but never serves the
+// SignalExchange service, so Engine.Start parks in WaitStreamConnected while
+// holding the engine mutex. When the Start context expires, the rollback path
+// calls ConnectClient.Stop, which must not block forever acquiring that mutex.
+func TestClientStartTimeoutRollback(t *testing.T) {
+	signalAddr := startBlackholeSignal(t)
+	mgmAddr := startManagement(t, signalAddr)
+
+	wgPort := 0
+	client, err := New(Options{
+		DeviceName:    "embed-rollback-test",
+		SetupKey:      testSetupKey,
+		ManagementURL: "http://" + mgmAddr,
+		WireguardPort: &wgPort,
+	})
+	require.NoError(t, err, "embed client creation must succeed")
+
+	startCtx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
+	defer cancel()
+
+	startErr := make(chan error, 1)
+	go func() {
+		startErr <- client.Start(startCtx)
+	}()
+
+	select {
+	case err := <-startErr:
+		require.ErrorIs(t, err, context.DeadlineExceeded)
+	case <-time.After(60 * time.Second):
+		t.Fatal("client.Start did not return after its context expired: Engine.Stop deadlocked against Engine.Start waiting for the signal stream")
+	}
+}
+
+// startBlackholeSignal starts a gRPC server without the SignalExchange service
+// registered. Connections succeed, but the signal stream can never be
+// established, which keeps Engine.Start parked in WaitStreamConnected.
+func startBlackholeSignal(t *testing.T) string {
+	t.Helper()
+
+	lis, err := net.Listen("tcp", "localhost:0")
+	require.NoError(t, err)
+
+	s := grpc.NewServer()
+	go func() {
+		if err := s.Serve(lis); err != nil {
+			t.Error(err)
+		}
+	}()
+	t.Cleanup(s.Stop)
+
+	return lis.Addr().String()
+}
+
+func startManagement(t *testing.T, signalAddr string) string {
+	t.Helper()
+
+	cfg := &config.Config{
+		Stuns:      []*config.Host{},
+		TURNConfig: &config.TURNConfig{},
+		Relay: &config.Relay{
+			Addresses:      []string{"127.0.0.1:1234"},
+			CredentialsTTL: util.Duration{Duration: time.Hour},
+			Secret:         "222222222222222222",
+		},
+		Signal: &config.Host{
+			Proto: "http",
+			URI:   signalAddr,
+		},
+		Datadir:    t.TempDir(),
+		HttpConfig: nil,
+	}
+
+	lis, err := net.Listen("tcp", "localhost:0")
+	require.NoError(t, err)
+
+	s := grpc.NewServer()
+
+	testStore, cleanUp, err := store.NewTestStoreFromSQL(context.Background(), "../testdata/store.sql", cfg.Datadir)
+	require.NoError(t, err)
+	t.Cleanup(cleanUp)
+
+	eventStore := &activity.InMemoryEventStore{}
+
+	permissionsManager := permissions.NewManager(testStore)
+	peersManager := peers.NewManager(testStore, permissionsManager)
+	jobManager := job.NewJobManager(nil, testStore, peersManager)
+
+	cacheStore, err := nbcache.NewStore(context.Background(), 100*time.Millisecond, 300*time.Millisecond, 100)
+	require.NoError(t, err)
+
+	iv, _ := validator.NewIntegratedValidator(context.Background(), peersManager, nil, eventStore, cacheStore)
+
+	metrics, err := telemetry.NewDefaultAppMetrics(context.Background())
+	require.NoError(t, err)
+
+	ctrl := gomock.NewController(t)
+	t.Cleanup(ctrl.Finish)
+	settingsMockManager := settings.NewMockManager(ctrl)
+	settingsMockManager.EXPECT().
+		GetSettings(gomock.Any(), gomock.Any(), gomock.Any()).
+		Return(&types.Settings{}, nil).
+		AnyTimes()
+	settingsMockManager.EXPECT().
+		GetExtraSettings(gomock.Any(), gomock.Any()).
+		Return(&types.ExtraSettings{}, nil).
+		AnyTimes()
+
+	groupsManager := groups.NewManagerMock()
+
+	updateManager := update_channel.NewPeersUpdateManager(metrics)
+	requestBuffer := mgmt.NewAccountRequestBuffer(context.Background(), testStore)
+	networkMapController := controller.NewController(context.Background(), testStore, metrics, updateManager, requestBuffer, mgmt.MockIntegratedValidator{}, settingsMockManager, "netbird.selfhosted", port_forwarding.NewControllerMock(), manager.NewEphemeralManager(testStore, peersManager), cfg)
+	accountManager, err := mgmt.BuildManager(context.Background(), cfg, testStore, networkMapController, jobManager, nil, "", eventStore, nil, false, iv, metrics, port_forwarding.NewControllerMock(), settingsMockManager, permissionsManager, false, cacheStore)
+	require.NoError(t, err)
+
+	secretsManager, err := nbgrpc.NewTimeBasedAuthSecretsManager(updateManager, cfg.TURNConfig, cfg.Relay, settingsMockManager, groupsManager)
+	require.NoError(t, err)
+
+	mgmtServer, err := nbgrpc.NewServer(cfg, accountManager, settingsMockManager, jobManager, secretsManager, nil, nil, &mgmt.MockIntegratedValidator{}, networkMapController, nil, nil)
+	require.NoError(t, err)
+	mgmtProto.RegisterManagementServiceServer(s, mgmtServer)
+
+	go func() {
+		if err := s.Serve(lis); err != nil {
+			t.Error(err)
+		}
+	}()
+	t.Cleanup(s.Stop)
+
+	return lis.Addr().String()
+}

client/internal/engine.go

@@ -880,62 +880,25 @@ func (e *Engine) handleSync(update *mgmProto.SyncResponse) error {
 		e.handleAutoUpdateVersion(update.NetworkMap.PeerConfig.AutoUpdate)
 	}
 
-	if update.GetNetbirdConfig() != nil {
-		wCfg := update.GetNetbirdConfig()
-		err := e.updateTURNs(wCfg.GetTurns())
-		if err != nil {
-			return fmt.Errorf("update TURNs: %w", err)
-		}
+	if err := e.updateNetbirdConfig(update.GetNetbirdConfig()); err != nil {
+		return err
+	}
 
-		err = e.updateSTUNs(wCfg.GetStuns())
-		if err != nil {
-			return fmt.Errorf("update STUNs: %w", err)
-		}
-
-		var stunTurn []*stun.URI
-		stunTurn = append(stunTurn, e.STUNs...)
-		stunTurn = append(stunTurn, e.TURNs...)
-		e.stunTurn.Store(stunTurn)
-
-		err = e.handleRelayUpdate(wCfg.GetRelay())
-		if err != nil {
-			return err
-		}
-
-		err = e.handleFlowUpdate(wCfg.GetFlow())
-		if err != nil {
-			return fmt.Errorf("handle the flow configuration: %w", err)
-		}
-
-		if err := e.PopulateNetbirdConfig(wCfg, nil); err != nil {
-			log.Warnf("Failed to update DNS server config: %v", err)
-		}
-
-		// todo update signal
+	// Posture checks are bound to the network map presence:
+	//   NetworkMap != nil, checks present -> apply the received checks
+	//   NetworkMap != nil, checks nil     -> posture checks were removed, clear them
+	//   NetworkMap == nil                 -> config-only update (e.g. relay token rotation),
+	//                                        leave the previously applied checks untouched
+	nm := update.GetNetworkMap()
+	if nm == nil {
+		return nil
 	}
 
 	if err := e.updateChecksIfNew(update.Checks); err != nil {
 		return err
 	}
 
-	nm := update.GetNetworkMap()
-	if nm == nil {
-		return nil
-	}
-
-	// Persist sync response under the dedicated lock (syncRespMux), not under syncMsgMux.
-	// A non-nil syncStore is what marks persistence as enabled. Hold the lock for
-	// the whole Set so the store cannot be cleared (disabled / engine close)
-	// mid-call and have this write resurrect a file that was just removed.
-	e.syncRespMux.RLock()
-	if e.syncStore != nil {
-		if err := e.syncStore.Set(update); err != nil {
-			log.Errorf("failed to persist sync response: %v", err)
-		} else {
-			log.Debugf("sync response persisted with serial %d", nm.GetSerial())
-		}
-	}
-	e.syncRespMux.RUnlock()
+	e.persistSyncResponse(update)
 
 	// only apply new changes and ignore old ones
 	if err := e.updateNetworkMap(nm); err != nil {
@@ -947,6 +910,64 @@ func (e *Engine) handleSync(update *mgmProto.SyncResponse) error {
 	return nil
 }
 
+// updateNetbirdConfig applies the management-provided NetBird configuration:
+// STUN/TURN and relay servers, flow logging and DNS settings. A nil config is a no-op,
+// which is the case for sync updates carrying only a network map.
+func (e *Engine) updateNetbirdConfig(wCfg *mgmProto.NetbirdConfig) error {
+	if wCfg == nil {
+		return nil
+	}
+
+	if err := e.updateTURNs(wCfg.GetTurns()); err != nil {
+		return fmt.Errorf("update TURNs: %w", err)
+	}
+
+	if err := e.updateSTUNs(wCfg.GetStuns()); err != nil {
+		return fmt.Errorf("update STUNs: %w", err)
+	}
+
+	var stunTurn []*stun.URI
+	stunTurn = append(stunTurn, e.STUNs...)
+	stunTurn = append(stunTurn, e.TURNs...)
+	e.stunTurn.Store(stunTurn)
+
+	if err := e.handleRelayUpdate(wCfg.GetRelay()); err != nil {
+		return err
+	}
+
+	if err := e.handleFlowUpdate(wCfg.GetFlow()); err != nil {
+		return fmt.Errorf("handle the flow configuration: %w", err)
+	}
+
+	if err := e.PopulateNetbirdConfig(wCfg, nil); err != nil {
+		log.Warnf("Failed to update DNS server config: %v", err)
+	}
+
+	// todo update signal
+
+	return nil
+}
+
+// persistSyncResponse stores the full sync response so it can be restored on the next
+// startup. Persistence is enabled only when syncStore is set. The dedicated syncRespMux
+// (not syncMsgMux) is held for the whole Set so the store cannot be cleared (disabled /
+// engine close) mid-call and have this write resurrect a file that was just removed.
+func (e *Engine) persistSyncResponse(update *mgmProto.SyncResponse) {
+	e.syncRespMux.RLock()
+	defer e.syncRespMux.RUnlock()
+
+	if e.syncStore == nil {
+		return
+	}
+
+	if err := e.syncStore.Set(update); err != nil {
+		log.Errorf("failed to persist sync response: %v", err)
+		return
+	}
+
+	log.Debugf("sync response persisted with serial %d", update.GetNetworkMap().GetSerial())
+}
+
 func (e *Engine) handleRelayUpdate(update *mgmProto.RelayConfig) error {
 	if update != nil {
 		// when we receive token we expect valid address list too

go.mod

@@ -2,6 +2,8 @@ module github.com/netbirdio/netbird
 
 go 1.25.5
 
+toolchain go1.25.11
+
 require (
 	cunicu.li/go-rosenpass v0.5.42
 	github.com/cenkalti/backoff/v4 v4.3.0
@@ -54,6 +56,7 @@ require (
 	github.com/fsnotify/fsnotify v1.9.0
 	github.com/gliderlabs/ssh v0.3.8
 	github.com/go-jose/go-jose/v4 v4.1.4
+	github.com/goccy/go-yaml v1.18.0
 	github.com/godbus/dbus/v5 v5.1.0
 	github.com/golang-jwt/jwt/v5 v5.3.1
 	github.com/golang/mock v1.6.0
@@ -211,10 +214,9 @@ require (
 	github.com/go-viper/mapstructure/v2 v2.5.0 // indirect
 	github.com/go-webauthn/webauthn v0.16.4 // indirect
 	github.com/go-webauthn/x v0.2.3 // indirect
-	github.com/goccy/go-yaml v1.18.0 // indirect
 	github.com/gogo/protobuf v1.3.2 // indirect
 	github.com/golang-jwt/jwt/v4 v4.5.2 // indirect
-	github.com/google/btree v1.1.2 // indirect
+	github.com/google/btree v1.1.3 // indirect
 	github.com/google/go-querystring v1.1.0 // indirect
 	github.com/google/go-tpm v0.9.8 // indirect
 	github.com/google/s2a-go v0.1.9 // indirect

go.sum

@@ -275,8 +275,8 @@ github.com/golang/protobuf v1.5.2/go.mod h1:XVQd3VNwM+JqD3oG2Ue2ip4fOMUkwXdXDdiu
 github.com/golang/protobuf v1.5.3/go.mod h1:XVQd3VNwM+JqD3oG2Ue2ip4fOMUkwXdXDdiuN0vRsmY=
 github.com/golang/protobuf v1.5.4 h1:i7eJL8qZTpSEXOPTxNKhASYpMn+8e5Q6AdndVa1dWek=
 github.com/golang/protobuf v1.5.4/go.mod h1:lnTiLA8Wa4RWRcIUkrtSVa5nRhsEGBg48fD6rSs7xps=
-github.com/google/btree v1.1.2 h1:xf4v41cLI2Z6FxbKm+8Bu+m8ifhj15JuZ9sa0jZCMUU=
-github.com/google/btree v1.1.2/go.mod h1:qOPhT0dTNdNzV6Z/lhRX0YXUafgPLFUh+gZMl761Gm4=
+github.com/google/btree v1.1.3 h1:CVpQJjYgC4VbzxeGVHfvZrv1ctoYCAI8vbl07Fcxlyg=
+github.com/google/btree v1.1.3/go.mod h1:qOPhT0dTNdNzV6Z/lhRX0YXUafgPLFUh+gZMl761Gm4=
 github.com/google/go-cmp v0.3.0/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
 github.com/google/go-cmp v0.3.1/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
 github.com/google/go-cmp v0.4.0/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=

management/internals/modules/reverseproxy/service/manager/l4_port_test.go

@@ -488,6 +488,195 @@ func TestUpdate_AllowsPortChange(t *testing.T) {
 	assert.Equal(t, uint16(54321), updated.ListenPort, "explicit port change should be applied")
 }
 
+func TestUpdate_PreservesPortWhenCustomPortsNotSupported(t *testing.T) {
+	mgr, testStore, _ := setupL4Test(t, boolPtr(false))
+	ctx := context.Background()
+
+	existing := seedService(t, testStore, "tcp-svc", "tcp", testCluster, testCluster, 12345)
+
+	updated := &rpservice.Service{
+		ID:           existing.ID,
+		AccountID:    testAccountID,
+		Name:         "tcp-svc-renamed",
+		Mode:         "tcp",
+		Domain:       testCluster,
+		ProxyCluster: testCluster,
+		ListenPort:   0,
+		Enabled:      true,
+		Targets: []*rpservice.Target{
+			{AccountID: testAccountID, TargetId: testPeerID, TargetType: rpservice.TargetTypePeer, Protocol: "tcp", Port: 9090, Enabled: true},
+		},
+	}
+
+	_, err := mgr.persistServiceUpdate(ctx, testAccountID, updated)
+	require.NoError(t, err, "update must not be rejected by the custom-port capability check")
+	assert.Equal(t, uint16(12345), updated.ListenPort, "existing listen port should be preserved on unsupported cluster")
+}
+
+func TestUpdate_PreservesPortWhenCustomPortsUnknown(t *testing.T) {
+	mgr, testStore, _ := setupL4Test(t, nil)
+	ctx := context.Background()
+
+	existing := seedService(t, testStore, "tcp-svc", "tcp", testCluster, testCluster, 12345)
+
+	updated := &rpservice.Service{
+		ID:           existing.ID,
+		AccountID:    testAccountID,
+		Name:         "tcp-svc-renamed",
+		Mode:         "tcp",
+		Domain:       testCluster,
+		ProxyCluster: testCluster,
+		ListenPort:   0,
+		Enabled:      true,
+		Targets: []*rpservice.Target{
+			{AccountID: testAccountID, TargetId: testPeerID, TargetType: rpservice.TargetTypePeer, Protocol: "tcp", Port: 9090, Enabled: true},
+		},
+	}
+
+	_, err := mgr.persistServiceUpdate(ctx, testAccountID, updated)
+	require.NoError(t, err, "update must not be rejected when cluster capability is unknown")
+	assert.Equal(t, uint16(12345), updated.ListenPort, "existing listen port should be preserved when capability is unknown")
+}
+
+func TestUpdate_RejectsPortChangeWhenCustomPortsNotSupported(t *testing.T) {
+	mgr, testStore, _ := setupL4Test(t, boolPtr(false))
+	ctx := context.Background()
+
+	existing := seedService(t, testStore, "tcp-svc", "tcp", testCluster, testCluster, 12345)
+
+	updated := &rpservice.Service{
+		ID:           existing.ID,
+		AccountID:    testAccountID,
+		Name:         "tcp-svc",
+		Mode:         "tcp",
+		Domain:       testCluster,
+		ProxyCluster: testCluster,
+		ListenPort:   54321,
+		Enabled:      true,
+		Targets: []*rpservice.Target{
+			{AccountID: testAccountID, TargetId: testPeerID, TargetType: rpservice.TargetTypePeer, Protocol: "tcp", Port: 9090, Enabled: true},
+		},
+	}
+
+	_, err := mgr.persistServiceUpdate(ctx, testAccountID, updated)
+	require.Error(t, err, "explicit port change on update must be rejected on unsupported clusters")
+	assert.Contains(t, err.Error(), "custom ports not supported on target cluster")
+}
+
+func TestUpdate_TLSPortChangeAllowedWhenNotSupported(t *testing.T) {
+	mgr, testStore, _ := setupL4Test(t, boolPtr(false))
+	ctx := context.Background()
+
+	existing := seedService(t, testStore, "tls-svc", "tls", "app.example.com", testCluster, 443)
+
+	updated := &rpservice.Service{
+		ID:           existing.ID,
+		AccountID:    testAccountID,
+		Name:         "tls-svc",
+		Mode:         "tls",
+		Domain:       "app.example.com",
+		ProxyCluster: testCluster,
+		ListenPort:   9999,
+		Enabled:      true,
+		Targets: []*rpservice.Target{
+			{AccountID: testAccountID, TargetId: testPeerID, TargetType: rpservice.TargetTypePeer, Protocol: "tcp", Port: 8443, Enabled: true},
+		},
+	}
+
+	_, err := mgr.persistServiceUpdate(ctx, testAccountID, updated)
+	require.NoError(t, err, "TLS port change uses SNI routing and is exempt from the custom-port check")
+	assert.Equal(t, uint16(9999), updated.ListenPort, "TLS port change should be applied")
+}
+
+func TestValidateL4PortDiffOnClusterDiff(t *testing.T) {
+	tests := []struct {
+		name        string
+		mode        string
+		customPorts *bool
+		newPort     uint16
+		oldPort     uint16
+		wantErr     bool
+	}{
+		{"tcp port change unsupported", "tcp", boolPtr(false), 54321, 12345, true},
+		{"tcp port change unknown capability", "tcp", nil, 54321, 12345, true},
+		{"udp port change unsupported", "udp", boolPtr(false), 54321, 12345, true},
+		{"tcp first port assignment unsupported", "tcp", boolPtr(false), 54321, 0, true},
+		{"tcp port change supported", "tcp", boolPtr(true), 54321, 12345, false},
+		{"tcp port unchanged unsupported", "tcp", boolPtr(false), 12345, 12345, false},
+		{"tcp zero port unsupported", "tcp", boolPtr(false), 0, 12345, false},
+		{"tls port change unsupported", "tls", boolPtr(false), 9999, 443, false},
+		{"http mode ignored", "http", boolPtr(false), 54321, 12345, false},
+		{"empty mode ignored", "", boolPtr(false), 54321, 12345, false},
+	}
+
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			newSvc := &rpservice.Service{Mode: tc.mode, ListenPort: tc.newPort, ProxyCluster: testCluster}
+			oldSvc := &rpservice.Service{Mode: tc.mode, ListenPort: tc.oldPort, ProxyCluster: testCluster}
+
+			err := validateL4PortDiffOnClusterDiff(tc.customPorts, newSvc, oldSvc)
+			if tc.wantErr {
+				assert.Error(t, err, "port diff should be rejected for %s", tc.name)
+			} else {
+				assert.NoError(t, err, "port diff should be allowed for %s", tc.name)
+			}
+		})
+	}
+}
+
+func TestUpdate_PortConflictRejected(t *testing.T) {
+	mgr, testStore, _ := setupL4Test(t, boolPtr(true))
+	ctx := context.Background()
+
+	seedService(t, testStore, "tcp-a", "tcp", "tcp-a."+testCluster, testCluster, 5432)
+	svcB := seedService(t, testStore, "tcp-b", "tcp", "tcp-b."+testCluster, testCluster, 6543)
+
+	updated := &rpservice.Service{
+		ID:           svcB.ID,
+		AccountID:    testAccountID,
+		Name:         "tcp-b",
+		Mode:         "tcp",
+		Domain:       "tcp-b." + testCluster,
+		ProxyCluster: testCluster,
+		ListenPort:   5432,
+		Enabled:      true,
+		Targets: []*rpservice.Target{
+			{AccountID: testAccountID, TargetId: testPeerID, TargetType: rpservice.TargetTypePeer, Protocol: "tcp", Port: 9090, Enabled: true},
+		},
+	}
+
+	_, err := mgr.persistServiceUpdate(ctx, testAccountID, updated)
+	require.Error(t, err, "updating to a port held by another service should be rejected")
+	assert.Contains(t, err.Error(), "already in use")
+}
+
+func TestUpdate_AutoAssignsWhenNoPort(t *testing.T) {
+	mgr, testStore, _ := setupL4Test(t, boolPtr(false))
+	ctx := context.Background()
+
+	existing := seedService(t, testStore, "tcp-svc", "tcp", testCluster, testCluster, 0)
+
+	updated := &rpservice.Service{
+		ID:           existing.ID,
+		AccountID:    testAccountID,
+		Name:         "tcp-svc",
+		Mode:         "tcp",
+		Domain:       testCluster,
+		ProxyCluster: testCluster,
+		ListenPort:   0,
+		Enabled:      true,
+		Targets: []*rpservice.Target{
+			{AccountID: testAccountID, TargetId: testPeerID, TargetType: rpservice.TargetTypePeer, Protocol: "tcp", Port: 9090, Enabled: true},
+		},
+	}
+
+	_, err := mgr.persistServiceUpdate(ctx, testAccountID, updated)
+	require.NoError(t, err)
+	assert.True(t, updated.ListenPort >= autoAssignPortMin && updated.ListenPort <= autoAssignPortMax,
+		"auto-assigned port %d should be in range [%d, %d]", updated.ListenPort, autoAssignPortMin, autoAssignPortMax)
+	assert.True(t, updated.PortAutoAssigned, "PortAutoAssigned should be set when update triggers auto-assignment")
+}
+
 func TestCreateServiceFromPeer_TCP(t *testing.T) {
 	mgr, _, _ := setupL4Test(t, boolPtr(false))
 	ctx := context.Background()

management/internals/modules/reverseproxy/service/manager/manager.go

@@ -338,7 +338,7 @@ func (m *Manager) persistNewService(ctx context.Context, accountID string, svc *
 			}
 		}
 
-		if err := m.ensureL4Port(ctx, transaction, svc, customPorts); err != nil {
+		if err := m.ensureL4Port(ctx, transaction, svc, customPorts, false); err != nil {
 			return err
 		}
 
@@ -367,11 +367,11 @@ func (m *Manager) clusterCustomPorts(ctx context.Context, svc *service.Service)
 
 // ensureL4Port auto-assigns a listen port when needed and validates cluster support.
 // customPorts must be pre-computed via clusterCustomPorts before entering a transaction.
-func (m *Manager) ensureL4Port(ctx context.Context, tx store.Store, svc *service.Service, customPorts *bool) error {
+func (m *Manager) ensureL4Port(ctx context.Context, tx store.Store, svc *service.Service, customPorts *bool, serviceUpdate bool) error {
 	if !service.IsL4Protocol(svc.Mode) {
 		return nil
 	}
-	if service.IsPortBasedProtocol(svc.Mode) && svc.ListenPort > 0 && (customPorts == nil || !*customPorts) {
+	if service.IsPortBasedProtocol(svc.Mode) && svc.ListenPort > 0 && !serviceUpdate && (customPorts == nil || !*customPorts) {
 		if svc.Source != service.SourceEphemeral {
 			return status.Errorf(status.InvalidArgument, "custom ports not supported on cluster %s", svc.ProxyCluster)
 		}
@@ -465,7 +465,7 @@ func (m *Manager) persistNewEphemeralService(ctx context.Context, accountID, pee
 			return err
 		}
 
-		if err := m.ensureL4Port(ctx, transaction, svc, customPorts); err != nil {
+		if err := m.ensureL4Port(ctx, transaction, svc, customPorts, false); err != nil {
 			return err
 		}
 
@@ -651,12 +651,22 @@ func (m *Manager) executeServiceUpdate(ctx context.Context, transaction store.St
 	m.preserveListenPort(service, existingService)
 	updateInfo.serviceEnabledChanged = existingService.Enabled != service.Enabled
 
-	if err := m.ensureL4Port(ctx, transaction, service, customPorts); err != nil {
+	// if the service is being updated, and we decide in the future to allow mode update,
+	// we should reconsider the currently assigned port if not 0 for clusters that don't support custom ports
+	if err := validateL4PortDiffOnClusterDiff(customPorts, service, existingService); err != nil {
 		return err
 	}
+
+	if err := m.ensureL4Port(ctx, transaction, service, customPorts, true); err != nil {
+		return err
+	}
+
+	// we can try carrying the previous service port into a new cluster, if this becomes a problem for multiple users,
+	// we should reconsider adding another check
 	if err := m.checkPortConflict(ctx, transaction, service); err != nil {
 		return err
 	}
+
 	if err := transaction.UpdateService(ctx, service); err != nil {
 		return fmt.Errorf("update service: %w", err)
 	}
@@ -664,6 +674,21 @@ func (m *Manager) executeServiceUpdate(ctx context.Context, transaction store.St
 	return nil
 }
 
+// validateL4PortDiffOnClusterDiff checks if custom L4 ports are configured and validates port changes across clusters.
+// It ensures no port changes if custom ports are unsupported for a given cluster and protocol mode.
+// Returns an error if validation fails, otherwise returns nil.
+func validateL4PortDiffOnClusterDiff(customPorts *bool, newSVC, oldSVC *service.Service) error {
+	if !service.IsPortBasedProtocol(newSVC.Mode) || (customPorts != nil && *customPorts) {
+		return nil
+	}
+
+	if newSVC.ListenPort != 0 && newSVC.ListenPort != oldSVC.ListenPort {
+		return status.Errorf(status.InvalidArgument, "custom ports not supported on target cluster %s", newSVC.ProxyCluster)
+	}
+
+	return nil
+}
+
 // handleDomainChange validates the new domain is free inside the transaction
 // and applies the pre-resolved cluster (computed outside the tx by
 // resolveEffectiveCluster). It must NOT call clusterDeriver here: that talks

management/internals/shared/grpc/conversion.go

@@ -8,6 +8,8 @@ import (
 	"strings"
 	"time"
 
+	"github.com/hashicorp/go-version"
+	nbversion "github.com/netbirdio/netbird/version"
 	log "github.com/sirupsen/logrus"
 	goproto "google.golang.org/protobuf/proto"
 	"google.golang.org/protobuf/types/known/timestamppb"
@@ -28,6 +30,23 @@ import (
 	"github.com/netbirdio/netbird/shared/sshauth"
 )
 
+const (
+	// deprecatedRemotePeersVersion is the version of Netbird that introduced the NetworkMap.RemotePeers field, deprecated in favor of RemotePeers.
+	deprecatedRemotePeersVersion = "0.29.3"
+)
+
+// precomputedDeprecatedRemotePeersConstraint is the parsed ">= 0.29.3" constraint,
+// built once at init since the bound is a compile-time constant.
+var precomputedDeprecatedRemotePeersConstraint version.Constraints
+
+func init() {
+	constraint, err := version.NewConstraint(">= " + deprecatedRemotePeersVersion)
+	if err != nil {
+		panic("parse deprecated remote peers version constraint: " + err.Error())
+	}
+	precomputedDeprecatedRemotePeersConstraint = constraint
+}
+
 func toNetbirdConfig(config *nbconfig.Config, turnCredentials *Token, relayToken *Token, extraSettings *types.ExtraSettings) *proto.NetbirdConfig {
 	if config == nil {
 		return nil
@@ -155,7 +174,11 @@ func ToSyncResponse(ctx context.Context, config *nbconfig.Config, httpConfig *nb
 
 	remotePeers := make([]*proto.RemotePeerConfig, 0, len(networkMap.Peers)+len(networkMap.OfflinePeers))
 	remotePeers = appendRemotePeerConfig(remotePeers, networkMap.Peers, dnsName, includeIPv6)
-	response.RemotePeers = remotePeers
+
+	if !shouldSkipSendingDeprecatedRemotePeers(peer.Meta.WtVersion) {
+		response.RemotePeers = remotePeers
+	}
+
 	response.NetworkMap.RemotePeers = remotePeers
 	response.RemotePeersIsEmpty = len(remotePeers) == 0
 	response.NetworkMap.RemotePeersIsEmpty = response.RemotePeersIsEmpty
@@ -246,6 +269,19 @@ func buildAuthorizedUsersProto(ctx context.Context, authorizedUsers map[string]m
 	return hashedUsers, machineUsers
 }
 
+func shouldSkipSendingDeprecatedRemotePeers(peerVersion string) bool {
+	if nbversion.IsDevelopmentVersion(peerVersion) {
+		return true
+	}
+
+	peerNBVersion, err := version.NewVersion(peerVersion)
+	if err != nil {
+		return false
+	}
+
+	return precomputedDeprecatedRemotePeersConstraint.Check(peerNBVersion)
+}
+
 func appendRemotePeerConfig(dst []*proto.RemotePeerConfig, peers []*nbpeer.Peer, dnsName string, includeIPv6 bool) []*proto.RemotePeerConfig {
 	for _, rPeer := range peers {
 		allowedIPs := []string{rPeer.IP.String() + "/32"}
@@ -363,7 +399,6 @@ func toProtocolFirewallRules(rules []*types.FirewallRule, includeIPv6, useSource
 	return result
 }
 
-
 // populateSourcePrefixes sets SourcePrefixes on fwRule and returns any
 // additional rules needed (e.g. a v6 wildcard clone when the peer IP is unspecified).
 func populateSourcePrefixes(fwRule *proto.FirewallRule, rule *types.FirewallRule, includeIPv6 bool) []*proto.FirewallRule {

management/internals/shared/grpc/conversion_test.go

@@ -202,6 +202,42 @@ func TestBuildJWTConfig_Audiences(t *testing.T) {
 	}
 }
 
+// TestShouldSkipSendingDeprecatedRemotePeers covers the version gate that
+// stops populating the deprecated top-level SyncResponse.RemotePeers field for
+// peers new enough to read RemotePeers off the NetworkMap. Development builds
+// are treated as latest and skip the field. The gate otherwise fails safe: a
+// release version older than the boundary, or one that can't be parsed (empty,
+// garbage, prereleases of the boundary) still receives the deprecated field so
+// older/unknown clients keep working.
+func TestShouldSkipSendingDeprecatedRemotePeers(t *testing.T) {
+	tests := []struct {
+		name        string
+		peerVersion string
+		wantSkip    bool
+	}{
+		{"exact boundary skips", "0.29.3", true},
+		{"newer patch skips", "0.29.4", true},
+		{"newer minor skips", "0.30.0", true},
+		{"newer major skips", "1.0.0", true},
+		{"v-prefixed newer skips", "v0.30.0", true},
+		{"development build skips", "development", true},
+		{"development build with commit skips", "development-abc123def456-dirty", true},
+		{"older patch keeps field", "0.29.2", false},
+		{"older minor keeps field", "0.28.0", false},
+		{"prerelease of boundary keeps field", "0.29.3-SNAPSHOT", false},
+		{"tagged dev prerelease keeps field", "v0.31.1-dev", false},
+		{"empty version keeps field", "", false},
+		{"garbage version keeps field", "not-a-version", false},
+	}
+
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			got := shouldSkipSendingDeprecatedRemotePeers(tc.peerVersion)
+			assert.Equal(t, tc.wantSkip, got, "skip decision for peer version %q", tc.peerVersion)
+		})
+	}
+}
+
 // TestEncodeSessionExpiresAt pins the wire encoding the client's
 // applySessionDeadline depends on:
 //

proxy/server.go

@@ -24,6 +24,7 @@ import (
 	"time"
 
 	"github.com/cenkalti/backoff/v4"
+	"github.com/google/uuid"
 	"github.com/pires/go-proxyproto"
 	prometheus2 "github.com/prometheus/client_golang/prometheus"
 	"github.com/prometheus/client_golang/prometheus/promhttp"
@@ -75,29 +76,30 @@ type portRouter struct {
 }
 
 type Server struct {
-	ctx           context.Context
-	mgmtClient    proto.ProxyServiceClient
-	proxy         *proxy.ReverseProxy
-	netbird       *roundtrip.NetBird
-	acme          *acme.Manager
-	auth          *auth.Middleware
-	http          *http.Server
-	https         *http.Server
-	debug         *http.Server
-	healthServer  *health.Server
-	healthChecker *health.Checker
-	meter         *proxymetrics.Metrics
-	accessLog     *accesslog.Logger
-	mainRouter    *nbtcp.Router
-	mainPort      uint16
-	udpMu         sync.Mutex
-	udpRelays     map[types.ServiceID]*udprelay.Relay
-	udpRelayWg    sync.WaitGroup
-	portMu        sync.RWMutex
-	portRouters   map[uint16]*portRouter
-	svcPorts      map[types.ServiceID][]uint16
-	lastMappings  map[types.ServiceID]*proto.ProxyMapping
-	portRouterWg  sync.WaitGroup
+	ctx               context.Context
+	mgmtClient        proto.ProxyServiceClient
+	proxy             *proxy.ReverseProxy
+	netbird           *roundtrip.NetBird
+	acme              *acme.Manager
+	staticCertWatcher *certwatch.Watcher
+	auth              *auth.Middleware
+	http              *http.Server
+	https             *http.Server
+	debug             *http.Server
+	healthServer      *health.Server
+	healthChecker     *health.Checker
+	meter             *proxymetrics.Metrics
+	accessLog         *accesslog.Logger
+	mainRouter        *nbtcp.Router
+	mainPort          uint16
+	udpMu             sync.Mutex
+	udpRelays         map[types.ServiceID]*udprelay.Relay
+	udpRelayWg        sync.WaitGroup
+	portMu            sync.RWMutex
+	portRouters       map[uint16]*portRouter
+	svcPorts          map[types.ServiceID][]uint16
+	lastMappings      map[types.ServiceID]*proto.ProxyMapping
+	portRouterWg      sync.WaitGroup
 
 	// hijackTracker tracks hijacked connections (e.g. WebSocket upgrades)
 	// so they can be closed during graceful shutdown, since http.Server.Shutdown
@@ -614,7 +616,7 @@ func (s *Server) initDefaults() {
 
 	// If no ID is set then one can be generated.
 	if s.ID == "" {
-		s.ID = "netbird-proxy-" + s.startTime.Format("20060102150405")
+		s.ID = fmt.Sprintf("netbird-proxy-%s", uuid.NewString())
 	}
 	// Fallback version option in case it is not set.
 	if s.Version == "" {
@@ -792,6 +794,7 @@ func (s *Server) configureTLS(ctx context.Context) (*tls.Config, error) {
 			return nil, fmt.Errorf("initialize certificate watcher: %w", err)
 		}
 		go certWatcher.Watch(ctx)
+		s.staticCertWatcher = certWatcher
 		tlsConfig.GetCertificate = certWatcher.GetCertificate
 		return tlsConfig, nil
 	}
@@ -1623,6 +1626,8 @@ func (s *Server) setupHTTPMapping(ctx context.Context, mapping *proto.ProxyMappi
 	var wildcardHit bool
 	if s.acme != nil {
 		wildcardHit = s.acme.AddDomain(d, accountID, svcID)
+	} else {
+		wildcardHit = s.staticCertCovers(d)
 	}
 	httpRoute := nbtcp.Route{
 		Type:      nbtcp.RouteHTTP,
@@ -1647,6 +1652,26 @@ func (s *Server) setupHTTPMapping(ctx context.Context, mapping *proto.ProxyMappi
 	return nil
 }
 
+// staticCertCovers reports whether the static certificate loaded when ACME is
+// disabled covers the given domain, making it certificate-ready immediately —
+// the equivalent of a wildcard hit in the ACME path. Domains the certificate
+// does not cover are logged: clients connecting to them will get TLS errors.
+func (s *Server) staticCertCovers(d domain.Domain) bool {
+	if s.staticCertWatcher == nil {
+		return false
+	}
+	leaf := s.staticCertWatcher.Leaf()
+	if leaf == nil {
+		return false
+	}
+	name := d.PunycodeString()
+	if err := leaf.VerifyHostname(name); err != nil {
+		s.Logger.Warnf("static certificate (SANs %v) does not cover domain %q: %v", leaf.DNSNames, name, err)
+		return false
+	}
+	return true
+}
+
 // setupTCPMapping sets up a TCP port-forwarding fallback route on the listen port.
 func (s *Server) setupTCPMapping(ctx context.Context, mapping *proto.ProxyMapping) error {
 	svcID := types.ServiceID(mapping.GetId())

proxy/static_cert_test.go

@@ -0,0 +1,89 @@
+package proxy
+
+import (
+	"crypto/ecdsa"
+	"crypto/elliptic"
+	"crypto/rand"
+	"crypto/x509"
+	"crypto/x509/pkix"
+	"encoding/pem"
+	"math/big"
+	"os"
+	"path/filepath"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/proxy/internal/certwatch"
+	"github.com/netbirdio/netbird/shared/management/domain"
+)
+
+func generateCertWithSANs(t *testing.T, dnsNames []string) (certPEM, keyPEM []byte) {
+	t.Helper()
+
+	key, err := ecdsa.GenerateKey(elliptic.P256(), rand.Reader)
+	require.NoError(t, err)
+
+	template := &x509.Certificate{
+		SerialNumber: big.NewInt(1),
+		Subject:      pkix.Name{CommonName: dnsNames[0]},
+		DNSNames:     dnsNames,
+		NotBefore:    time.Now().Add(-time.Hour),
+		NotAfter:     time.Now().Add(24 * time.Hour),
+	}
+
+	certDER, err := x509.CreateCertificate(rand.Reader, template, template, &key.PublicKey, key)
+	require.NoError(t, err)
+	certPEM = pem.EncodeToMemory(&pem.Block{Type: "CERTIFICATE", Bytes: certDER})
+
+	keyDER, err := x509.MarshalECPrivateKey(key)
+	require.NoError(t, err)
+	keyPEM = pem.EncodeToMemory(&pem.Block{Type: "EC PRIVATE KEY", Bytes: keyDER})
+
+	return certPEM, keyPEM
+}
+
+func newStaticWatcher(t *testing.T, dnsNames []string) *certwatch.Watcher {
+	t.Helper()
+
+	dir := t.TempDir()
+	certPEM, keyPEM := generateCertWithSANs(t, dnsNames)
+	certPath := filepath.Join(dir, "tls.crt")
+	keyPath := filepath.Join(dir, "tls.key")
+	require.NoError(t, os.WriteFile(certPath, certPEM, 0o600))
+	require.NoError(t, os.WriteFile(keyPath, keyPEM, 0o600))
+
+	w, err := certwatch.NewWatcher(certPath, keyPath, quietLifecycleLogger())
+	require.NoError(t, err)
+	return w
+}
+
+func TestStaticCertCovers(t *testing.T) {
+	s := &Server{
+		Logger:            quietLifecycleLogger(),
+		staticCertWatcher: newStaticWatcher(t, []string{"*.p.example.com", "exact.example.com"}),
+	}
+
+	cases := []struct {
+		domain  string
+		covered bool
+	}{
+		{"svc.p.example.com", true},
+		{"exact.example.com", true},
+		{"a.b.p.example.com", false}, // wildcard does not span labels
+		{"p.example.com", false},
+		{"other.example.com", false},
+	}
+	for _, tc := range cases {
+		t.Run(tc.domain, func(t *testing.T) {
+			assert.Equal(t, tc.covered, s.staticCertCovers(domain.Domain(tc.domain)))
+		})
+	}
+}
+
+func TestStaticCertCoversNoWatcher(t *testing.T) {
+	s := &Server{Logger: quietLifecycleLogger()}
+	assert.False(t, s.staticCertCovers(domain.Domain("svc.p.example.com")))
+}

shared/management/client/client_test.go

@@ -322,15 +322,21 @@ func TestClient_Sync(t *testing.T) {
 		if resp.GetNetbirdConfig() == nil {
 			t.Error("expecting non nil NetbirdConfig got nil")
 		}
-		if len(resp.GetRemotePeers()) != 1 {
-			t.Errorf("expecting RemotePeers size %d got %d", 1, len(resp.GetRemotePeers()))
+		// we test network map peers from 0.29.3 and dev builds
+		if len(resp.GetRemotePeers()) != 0 {
+			t.Error("expecting top-level RemotePeers to be empty for v0.29.3+ clients")
+		}
+		networkMap := resp.GetNetworkMap()
+		if len(networkMap.GetRemotePeers()) != 1 {
+			t.Errorf("expecting RemotePeers size %d got %d", 1, len(networkMap.GetRemotePeers()))
 			return
 		}
-		if resp.GetRemotePeersIsEmpty() == true {
+
+		if networkMap.GetRemotePeersIsEmpty() {
 			t.Error("expecting RemotePeers property to be false, got true")
 		}
-		if resp.GetRemotePeers()[0].GetWgPubKey() != remoteKey.PublicKey().String() {
-			t.Errorf("expecting RemotePeer public key %s got %s", remoteKey.PublicKey().String(), resp.GetRemotePeers()[0].GetWgPubKey())
+		if networkMap.GetRemotePeers()[0].GetWgPubKey() != remoteKey.PublicKey().String() {
+			t.Errorf("expecting RemotePeer public key %s got %s", remoteKey.PublicKey().String(), networkMap.GetRemotePeers()[0].GetWgPubKey())
 		}
 	case <-time.After(3 * time.Second):
 		t.Error("timeout waiting for test to finish")