reduce hashing on relevant groups calc

## Describe your changes

Restore the management gRPC client's backoff reset after a stream
connects.

This is a regression: PR #4935 originally added this reset; it was then
lost in
commit `58daa674e` during the `withMgmtStream` refactor (Sync/Job
unification).

Compared to before the reset was done AFTER receiveEvents, whereas 
now it's done before. 

Now should cover for Sync and Job (which didn't exist in #4935. Hold for
long living stream that breaks.

Reset prevents two things

1. long living conns going wrong from waiting a long backoff time window
before to drive reconn
2. past MaxElapsedTime (3months) long lived conns failures from being
treated as "unrecoverable" (hence triggering an full restart of the
engine)

## Issue ticket number and link

Internal support case (customer agents lost data-plane connectivity
during a
management maintenance/release window). No public issue. Regression
introduced
in commit `58daa674e`, which removed the `backOff.Reset()` added by PR
#4935.

## Stack

<!-- branch-stack -->

### Checklist
- [x] Is it a bug fix
- [ ] Is a typo/documentation fix
- [ ] Is a feature enhancement
- [ ] It is a refactor
- [ ] Created tests that fail without the change (if possible)

> By submitting this pull request, you confirm that you have read and
agree to the terms of the [Contributor License
Agreement](https://github.com/netbirdio/netbird/blob/main/CONTRIBUTOR_LICENSE_AGREEMENT.md).

## Documentation
Select exactly one:

- [ ] I added/updated documentation for this change
- [x] Documentation is **not needed** for this change (explain why)

Internal reconnection-behavior fix in the management gRPC client. No
public
NetBird CLI, API, or configuration surface changes.

### Docs PR URL (required if "docs added" is checked)
Paste the PR link from https://github.com/netbirdio/docs here:

N/A

<!-- codesmith:footer -->
---
<a
href="https://app.blacksmith.sh/netbirdio/codesmith/netbird/pr/6883"><picture><source
media="(prefers-color-scheme: dark)"
srcset="https://pr-comments-assets.blacksmith.sh/codesmith/view-with-codesmith-dark-v2.svg"><source
media="(prefers-color-scheme: light)"
srcset="https://pr-comments-assets.blacksmith.sh/codesmith/view-with-codesmith-light-v2.svg"><img
alt="View with [code]smith"
src="https://pr-comments-assets.blacksmith.sh/codesmith/view-with-codesmith-dark-v2.svg"></picture></a>
<a
href="https://backend.blacksmith.sh/track/enable-autofix?expires=1787484622&installation_model_id=427504&pr_number=6883&repository=netbirdio%2Fnetbird&return_to=https%3A%2F%2Fgithub.com%2Fnetbirdio%2Fnetbird%2Fpull%2F6883&signature=525b08edbcb94ee6f4ae4226fb1fe8ddb829caa7d4489e8efa639770af9dccdf"><picture><source
media="(prefers-color-scheme: dark)"
srcset="https://pr-comments-assets.blacksmith.sh/codesmith/autofix-with-codesmith-dark.svg"><source
media="(prefers-color-scheme: light)"
srcset="https://pr-comments-assets.blacksmith.sh/codesmith/autofix-with-codesmith-light.svg"><img
alt="Autofix with [code]smith"
src="https://pr-comments-assets.blacksmith.sh/codesmith/autofix-with-codesmith-dark.svg"></picture></a>
<sup>Need help on this PR? Tag <code>@codesmith-bot</code> with what you
need. Autofix is disabled.</sup>

<!-- codesmith:autofix:disabled -->
<!-- /codesmith:footer -->

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

* **Bug Fixes**
  * Improved management stream retry behavior.
* Retry delays now reset after a stream is successfully established,
helping subsequent connection attempts recover more quickly.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->

client/internal/conn_mgr.go

@@ -49,11 +49,21 @@ type ConnMgr struct {
 	// engine.syncMsgMux; all other reads stay under engine.syncMsgMux only.
 	lazyConnMgrMu sync.RWMutex
 
+	// reconcileRoutedIPs re-applies a peer's routed allowed IPs after its lazy wake endpoint is
+	// (re)armed (Mode A at arm time). Injected by the engine; nil disables the reconcile.
+	reconcileRoutedIPs func(peerKey string) error
+
 	wg            sync.WaitGroup
 	lazyCtx       context.Context
 	lazyCtxCancel context.CancelFunc
 }
 
+// SetRoutedIPsReconciler injects the callback used to re-apply a peer's routed allowed IPs when
+// its lazy wake endpoint is (re)armed. Must be called before the lazy manager starts.
+func (e *ConnMgr) SetRoutedIPsReconciler(fn func(peerKey string) error) {
+	e.reconcileRoutedIPs = fn
+}
+
 func NewConnMgr(engineConfig *EngineConfig, statusRecorder *peer.Status, peerStore *peerstore.Store, iface lazyconn.WGIface) *ConnMgr {
 	e := &ConnMgr{
 		peerStore:        peerStore,
@@ -291,6 +301,7 @@ func (e *ConnMgr) Close() {
 func (e *ConnMgr) initLazyManager(engineCtx context.Context) {
 	cfg := manager.Config{
 		InactivityThreshold: inactivityThresholdEnv(),
+		ReconcileAllowedIPs: e.reconcileRoutedIPs,
 	}
 
 	e.lazyConnMgrMu.Lock()

client/internal/engine.go

@@ -663,6 +663,12 @@ func (e *Engine) Start(netbirdConfig *mgmProto.NetbirdConfig, mgmtURL *url.URL)
 	iceCfg := e.createICEConfig()
 
 	e.connMgr = NewConnMgr(e.config, e.statusRecorder, e.peerStore, wgIface)
+	e.connMgr.SetRoutedIPsReconciler(func(peerKey string) error {
+		if e.routeManager == nil {
+			return nil
+		}
+		return e.routeManager.ReconcilePeerAllowedIPs(peerKey)
+	})
 	e.connMgr.Start(e.ctx)
 
 	// Wire DNS-time lazy-connection warm-up now that the connection manager

client/internal/lazyconn/manager/manager.go

@@ -29,6 +29,11 @@ type managedPeer struct {
 
 type Config struct {
 	InactivityThreshold *time.Duration
+	// ReconcileAllowedIPs re-applies a peer's routed allowed IPs after its wake endpoint is
+	// armed. The activity listener creates the wake peer with the overlay /32 only; without the
+	// routed prefixes WireGuard would not steer subnet-bound traffic to the wake endpoint, so an
+	// idle routing peer could never be woken by that traffic. Optional; nil disables the reconcile.
+	ReconcileAllowedIPs func(peerKey string) error
 }
 
 // Manager manages lazy connections
@@ -56,6 +61,9 @@ type Manager struct {
 	peerToHAGroups map[string][]route.HAUniqueID // peer ID -> HA groups they belong to
 	haGroupToPeers map[route.HAUniqueID][]string // HA group -> peer IDs in the group
 	routesMu       sync.RWMutex
+
+	// reconcileAllowedIPs re-applies a peer's routed allowed IPs after its wake endpoint is armed.
+	reconcileAllowedIPs func(peerKey string) error
 }
 
 // NewManager creates a new lazy connection manager
@@ -73,6 +81,7 @@ func NewManager(config Config, engineCtx context.Context, peerStore *peerstore.S
 		activityManager:      activity.NewManager(wgIface),
 		peerToHAGroups:       make(map[string][]route.HAUniqueID),
 		haGroupToPeers:       make(map[route.HAUniqueID][]string),
+		reconcileAllowedIPs:  config.ReconcileAllowedIPs,
 	}
 
 	if wgIface.IsUserspaceBind() {
@@ -201,7 +210,7 @@ func (m *Manager) AddPeer(peerCfg lazyconn.PeerConfig) (bool, error) {
 		return false, nil
 	}
 
-	if err := m.activityManager.MonitorPeerActivity(peerCfg); err != nil {
+	if err := m.armActivityListener(peerCfg); err != nil {
 		return false, err
 	}
 
@@ -288,7 +297,7 @@ func (m *Manager) DeactivatePeer(peerID peerid.ConnID) {
 
 	m.inactivityManager.RemovePeer(mp.peerCfg.PublicKey)
 
-	if err := m.activityManager.MonitorPeerActivity(*mp.peerCfg); err != nil {
+	if err := m.armActivityListener(*mp.peerCfg); err != nil {
 		mp.peerCfg.Log.Errorf("failed to create activity monitor: %v", err)
 		return
 	}
@@ -465,6 +474,31 @@ func (m *Manager) close() {
 }
 
 // shouldDeferIdleForHA checks if peer should stay connected due to HA group requirements
+// armRoutedAllowedIPs re-applies the peer's routed allowed IPs onto its freshly armed wake
+// endpoint. The activity listener creates the wake peer with the overlay /32 only, so without
+// this the routed prefixes would be missing and traffic to a routed subnet could not wake the
+// idle routing peer. It is a no-op when no reconciler is configured.
+// armActivityListener (re)arms the peer's wake endpoint via the activity manager and then
+// re-applies its routed allowed IPs, so traffic to a routed subnet can wake an idle routing
+// peer. The routed prefixes must be re-applied after the wake endpoint exists because the
+// listener creates it with the overlay /32 only.
+func (m *Manager) armActivityListener(peerCfg lazyconn.PeerConfig) error {
+	if err := m.activityManager.MonitorPeerActivity(peerCfg); err != nil {
+		return err
+	}
+	m.armRoutedAllowedIPs(&peerCfg)
+	return nil
+}
+
+func (m *Manager) armRoutedAllowedIPs(peerCfg *lazyconn.PeerConfig) {
+	if m.reconcileAllowedIPs == nil {
+		return
+	}
+	if err := m.reconcileAllowedIPs(peerCfg.PublicKey); err != nil {
+		peerCfg.Log.Errorf("failed to reconcile routed allowed IPs on wake endpoint: %v", err)
+	}
+}
+
 func (m *Manager) shouldDeferIdleForHA(inactivePeers map[string]struct{}, peerID string) bool {
 	m.routesMu.RLock()
 	defer m.routesMu.RUnlock()
@@ -577,7 +611,7 @@ func (m *Manager) onPeerInactivityTimedOut(peerIDs map[string]struct{}) {
 
 		mp.peerCfg.Log.Infof("start activity monitor")
 
-		if err := m.activityManager.MonitorPeerActivity(*mp.peerCfg); err != nil {
+		if err := m.armActivityListener(*mp.peerCfg); err != nil {
 			mp.peerCfg.Log.Errorf("failed to create activity monitor: %v", err)
 			continue
 		}

client/internal/routemanager/manager.go

@@ -61,6 +61,7 @@ type Manager interface {
 	InitialRouteRange() []string
 	SetFirewall(firewall.Manager) error
 	SetDNSForwarderPort(port uint16)
+	ReconcilePeerAllowedIPs(peerKey string) error
 	Stop(stateManager *statemanager.Manager)
 }
 
@@ -232,6 +233,30 @@ func (m *DefaultManager) setupRefCounters(useNoop bool) {
 	)
 }
 
+// ReconcilePeerAllowedIPs re-applies every routed allowed IP currently tracked for the peer
+// onto the WireGuard device. The allowed-IP refcounter only calls its AddFunc (which pushes to
+// the device) on a prefix's 0->1 transition, so a peer whose device entry was rebuilt without a
+// matching refcounter change — e.g. a lazy connection cycling through idle->wake, which recreates
+// the WireGuard peer with the overlay /32 only — ends up missing routed prefixes the refcounter
+// still considers installed, and nothing retries. Calling this when the peer's WireGuard entry is
+// (re)created restores convergence. It is add-only and idempotent: AddAllowedIP is update-only, so
+// prefixes are re-added to an existing peer and an absent peer is left untouched.
+func (m *DefaultManager) ReconcilePeerAllowedIPs(peerKey string) error {
+	if m.allowedIPsRefCounter == nil {
+		return nil
+	}
+
+	return m.allowedIPsRefCounter.ReapplyMatching(
+		func(out string) bool { return out == peerKey },
+		func(prefix netip.Prefix) error {
+			if err := m.wgInterface.AddAllowedIP(peerKey, prefix); err != nil {
+				return fmt.Errorf("add allowed IP %s for peer %s: %w", prefix, peerKey, err)
+			}
+			return nil
+		},
+	)
+}
+
 // Init sets up the routing
 func (m *DefaultManager) Init() error {
 	m.routeSelector = m.initSelector()

client/internal/routemanager/mock.go

@@ -112,6 +112,11 @@ func (m *MockManager) SetFirewall(firewall.Manager) error {
 func (m *MockManager) SetDNSForwarderPort(port uint16) {
 }
 
+// ReconcilePeerAllowedIPs mock implementation of ReconcilePeerAllowedIPs from Manager interface
+func (m *MockManager) ReconcilePeerAllowedIPs(peerKey string) error {
+	return nil
+}
+
 // Stop mock implementation of Stop from Manager interface
 func (m *MockManager) Stop(stateManager *statemanager.Manager) {
 	if m.StopFunc != nil {

client/internal/routemanager/reconcile_test.go

@@ -0,0 +1,90 @@
+//go:build !windows
+
+package routemanager
+
+import (
+	"net"
+	"net/netip"
+	"sync"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	"golang.zx2c4.com/wireguard/tun/netstack"
+
+	"github.com/netbirdio/netbird/client/iface/device"
+	"github.com/netbirdio/netbird/client/iface/wgaddr"
+	"github.com/netbirdio/netbird/client/internal/routemanager/refcounter"
+)
+
+// reconcileWGMock is a minimal iface.WGIface that only records AddAllowedIP calls; every other
+// method is an inert stub because ReconcilePeerAllowedIPs exercises none of them.
+type reconcileWGMock struct {
+	mu   sync.Mutex
+	adds map[string][]netip.Prefix
+}
+
+func (m *reconcileWGMock) AddAllowedIP(peerKey string, allowedIP netip.Prefix) error {
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	if m.adds == nil {
+		m.adds = map[string][]netip.Prefix{}
+	}
+	m.adds[peerKey] = append(m.adds[peerKey], allowedIP)
+	return nil
+}
+
+func (m *reconcileWGMock) added(peerKey string) []netip.Prefix {
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	return m.adds[peerKey]
+}
+
+func (m *reconcileWGMock) RemoveAllowedIP(string, netip.Prefix) error { return nil }
+func (m *reconcileWGMock) Name() string                               { return "utun-test" }
+func (m *reconcileWGMock) Address() wgaddr.Address                    { return wgaddr.Address{} }
+func (m *reconcileWGMock) ToInterface() *net.Interface                { return nil }
+func (m *reconcileWGMock) IsUserspaceBind() bool                      { return false }
+func (m *reconcileWGMock) GetFilter() device.PacketFilter             { return nil }
+func (m *reconcileWGMock) GetDevice() *device.FilteredDevice          { return nil }
+func (m *reconcileWGMock) GetNet() *netstack.Net                      { return nil }
+
+// TestReconcilePeerAllowedIPs verifies the declarative reconcile re-applies every routed prefix
+// tracked for the peer (self-heal, independent of refcount level) and stays scoped to that peer.
+func TestReconcilePeerAllowedIPs(t *testing.T) {
+	wg := &reconcileWGMock{}
+	m := &DefaultManager{wgInterface: wg}
+	m.allowedIPsRefCounter = refcounter.New[netip.Prefix, string, string](
+		func(_ netip.Prefix, peerKey string) (string, error) { return peerKey, nil },
+		func(netip.Prefix, string) error { return nil },
+	)
+
+	peerA1 := netip.MustParsePrefix("10.0.0.0/24")
+	peerA2 := netip.MustParsePrefix("10.1.0.0/24")
+	peerB1 := netip.MustParsePrefix("10.2.0.0/24")
+
+	for prefix, peer := range map[netip.Prefix]string{peerA1: "peerA", peerA2: "peerA", peerB1: "peerB"} {
+		_, err := m.allowedIPsRefCounter.Increment(prefix, peer)
+		require.NoError(t, err)
+	}
+	// Extra reference: reconcile must still re-apply the prefix even though its refcount never
+	// hit 0 again (the exact case the plain incremental path skips).
+	_, err := m.allowedIPsRefCounter.Increment(peerA1, "peerA")
+	require.NoError(t, err)
+
+	require.NoError(t, m.ReconcilePeerAllowedIPs("peerA"))
+
+	assert.ElementsMatch(t, []netip.Prefix{peerA1, peerA2}, wg.added("peerA"),
+		"reconcile must re-apply all routed prefixes of the peer")
+	assert.Empty(t, wg.added("peerB"), "reconcile must not touch another peer's prefixes")
+}
+
+// TestReconcilePeerAllowedIPsNoCounter verifies reconcile is a safe no-op before the refcounter is
+// set up.
+func TestReconcilePeerAllowedIPsNoCounter(t *testing.T) {
+	wg := &reconcileWGMock{}
+	m := &DefaultManager{wgInterface: wg}
+
+	require.NoError(t, m.ReconcilePeerAllowedIPs("peerA"))
+	assert.Empty(t, wg.added("peerA"))
+}

client/internal/routemanager/refcounter/refcounter.go

@@ -94,6 +94,26 @@ func (rm *Counter[Key, I, O]) Get(key Key) (Ref[O], bool) {
 	return ref, ok
 }
 
+// ReapplyMatching calls apply for every key whose stored Out satisfies pred, holding the
+// counter lock for the whole pass. Running apply under the lock keeps it atomic with respect
+// to Increment/Decrement: a prefix dropped to zero is removed from the map (and had its
+// RemoveFunc called) before this pass observes it, so a stale key can never be re-applied.
+// pred and apply are invoked under the lock, so they must not call back into the counter.
+func (rm *Counter[Key, I, O]) ReapplyMatching(pred func(out O) bool, apply func(key Key) error) error {
+	rm.mu.Lock()
+	defer rm.mu.Unlock()
+
+	var merr *multierror.Error
+	for key, ref := range rm.refCountMap {
+		if pred(ref.Out) {
+			if err := apply(key); err != nil {
+				merr = multierror.Append(merr, err)
+			}
+		}
+	}
+	return nberrors.FormatErrorOrNil(merr)
+}
+
 // Increment increments the reference count for the given key.
 // If this is the first reference to the key, the AddFunc is called.
 func (rm *Counter[Key, I, O]) Increment(key Key, in I) (Ref[O], error) {

client/internal/routemanager/refcounter/refcounter_test.go

@@ -0,0 +1,47 @@
+package refcounter
+
+import (
+	"net/netip"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+// TestReapplyMatching verifies ReapplyMatching invokes apply for exactly the keys whose stored
+// Out satisfies the predicate (no duplicates for multiply-referenced keys) — the primitive
+// ReconcilePeerAllowedIPs relies on to re-apply a single peer's routed prefixes.
+func TestReapplyMatching(t *testing.T) {
+	rc := New[netip.Prefix, string, string](
+		func(_ netip.Prefix, peerKey string) (string, error) { return peerKey, nil },
+		func(netip.Prefix, string) error { return nil },
+	)
+
+	peerA1 := netip.MustParsePrefix("10.0.0.0/24")
+	peerA2 := netip.MustParsePrefix("10.1.0.0/24")
+	peerB1 := netip.MustParsePrefix("10.2.0.0/24")
+
+	for prefix, peer := range map[netip.Prefix]string{peerA1: "peerA", peerA2: "peerA", peerB1: "peerB"} {
+		_, err := rc.Increment(prefix, peer)
+		require.NoError(t, err)
+	}
+	// a second reference must not make the key applied twice
+	_, err := rc.Increment(peerA1, "peerA")
+	require.NoError(t, err)
+
+	var applied []netip.Prefix
+	err = rc.ReapplyMatching(
+		func(out string) bool { return out == "peerA" },
+		func(key netip.Prefix) error { applied = append(applied, key); return nil },
+	)
+	require.NoError(t, err)
+	assert.ElementsMatch(t, []netip.Prefix{peerA1, peerA2}, applied)
+
+	var none []netip.Prefix
+	err = rc.ReapplyMatching(
+		func(out string) bool { return out == "missing" },
+		func(key netip.Prefix) error { none = append(none, key); return nil },
+	)
+	require.NoError(t, err)
+	assert.Empty(t, none)
+}

management/server/types/account_components.go

@@ -321,10 +321,19 @@ func (a *Account) getPeersGroupsPoliciesRoutes(
 
 	relevantPeerIDs[peerID] = a.GetPeer(peerID).ToComponent()
 
+	addRelevantGroup := func(groupID string) *Group {
+		if g, ok := relevantGroupIDs[groupID]; ok {
+			return g
+		}
+		g := a.GetGroup(groupID)
+		relevantGroupIDs[groupID] = g
+		return g
+	}
+
 	peerGroupSet := make(map[string]struct{}, 8)
 	for groupID, group := range a.Groups {
 		if slices.Contains(group.Peers, peerID) {
-			relevantGroupIDs[groupID] = a.GetGroup(groupID)
+			relevantGroupIDs[groupID] = group
 			peerGroupSet[groupID] = struct{}{}
 		}
 	}
@@ -355,15 +364,12 @@ func (a *Account) getPeersGroupsPoliciesRoutes(
 			continue
 		}
 
-		for _, groupID := range r.PeerGroups {
-			relevantGroupIDs[groupID] = a.GetGroup(groupID)
-		}
 		for _, groupID := range r.Groups {
-			relevantGroupIDs[groupID] = a.GetGroup(groupID)
+			addRelevantGroup(groupID)
 		}
 		if r.Enabled {
 			for _, groupID := range r.AccessControlGroups {
-				relevantGroupIDs[groupID] = a.GetGroup(groupID)
+				addRelevantGroup(groupID)
 				routeAccessControlGroups[groupID] = struct{}{}
 			}
 		}
@@ -389,7 +395,7 @@ func (a *Account) getPeersGroupsPoliciesRoutes(
 			}
 		}
 		for _, groupID := range r.PeerGroups {
-			g := a.GetGroup(groupID)
+			g := addRelevantGroup(groupID)
 			if g == nil {
 				continue
 			}
@@ -424,10 +430,10 @@ func (a *Account) getPeersGroupsPoliciesRoutes(
 					if _, needed := routeAccessControlGroups[destGroupID]; needed {
 						policyRelevant = true
 						for _, srcGroupID := range rule.Sources {
-							relevantGroupIDs[srcGroupID] = a.GetGroup(srcGroupID)
+							addRelevantGroup(srcGroupID)
 						}
 						for _, dstGroupID := range rule.Destinations {
-							relevantGroupIDs[dstGroupID] = a.GetGroup(dstGroupID)
+							addRelevantGroup(dstGroupID)
 						}
 						break
 					}
@@ -463,7 +469,7 @@ func (a *Account) getPeersGroupsPoliciesRoutes(
 					}
 				}
 				for _, dstGroupID := range rule.Destinations {
-					relevantGroupIDs[dstGroupID] = a.GetGroup(dstGroupID)
+					addRelevantGroup(dstGroupID)
 				}
 			}
 
@@ -475,7 +481,7 @@ func (a *Account) getPeersGroupsPoliciesRoutes(
 					}
 				}
 				for _, srcGroupID := range rule.Sources {
-					relevantGroupIDs[srcGroupID] = a.GetGroup(srcGroupID)
+					addRelevantGroup(srcGroupID)
 				}
 
 				if rule.Protocol == PolicyRuleProtocolNetbirdSSH {

release_files/freebsd-port-diff.sh

@@ -3,8 +3,8 @@
 # FreeBSD Port Diff Generator for NetBird
 #
 # This script generates the diff file required for submitting a FreeBSD port update.
-# It works on macOS, Linux, and FreeBSD by fetching files from FreeBSD cgit and
-# computing checksums from the Go module proxy.
+# It works on macOS, Linux, and FreeBSD by fetching files from the FreeBSD ports
+# GitHub mirror and computing checksums from the Go module proxy.
 #
 # Usage: ./freebsd-port-diff.sh [new_version]
 # Example: ./freebsd-port-diff.sh 0.60.7
@@ -14,7 +14,7 @@
 set -e
 
 GITHUB_REPO="netbirdio/netbird"
-PORTS_CGIT_BASE="https://cgit.freebsd.org/ports/plain/security/netbird"
+PORTS_MIRROR_BASE="https://raw.githubusercontent.com/freebsd/freebsd-ports/main/security/netbird"
 GO_PROXY="https://proxy.golang.org/github.com/netbirdio/netbird/@v"
 OUTPUT_DIR="${OUTPUT_DIR:-.}"
 AWK_FIRST_FIELD='{print $1}'
@@ -30,10 +30,17 @@ fetch_all_tags() {
 
 fetch_current_ports_version() {
     echo "Fetching current version from FreeBSD ports..." >&2
-    curl -sL "${PORTS_CGIT_BASE}/Makefile" 2>/dev/null | \
+    local makefile version
+    makefile=$(fetch_ports_file "Makefile") || return 1
+    version=$(echo "$makefile" | \
         grep -E "^DISTVERSION=" | \
         sed 's/DISTVERSION=[[:space:]]*//' | \
-        tr -d '\t '
+        tr -d '\t ')
+    if [[ -z "$version" ]]; then
+        echo "Error: Could not extract DISTVERSION from ports Makefile" >&2
+        return 1
+    fi
+    echo "$version"
     return 0
 }
 
@@ -45,7 +52,16 @@ fetch_latest_github_release() {
 
 fetch_ports_file() {
     local filename="$1"
-    curl -sL "${PORTS_CGIT_BASE}/${filename}" 2>/dev/null
+    local content
+    if ! content=$(curl -fsL --proto '=https' --proto-redir '=https' --retry 3 "${PORTS_MIRROR_BASE}/${filename}" 2>/dev/null); then
+        echo "Error: Could not fetch ${filename} from ${PORTS_MIRROR_BASE}" >&2
+        return 1
+    fi
+    if [[ "$content" == \<* ]]; then
+        echo "Error: Received HTML instead of ${filename} from ${PORTS_MIRROR_BASE}" >&2
+        return 1
+    fi
+    printf '%s' "$content"
     return 0
 }
 

release_files/freebsd-port-issue-body.sh

@@ -9,18 +9,22 @@
 # Example: ./freebsd-port-issue-body.sh 0.56.0 0.59.1
 #
 # If no versions are provided, the script will:
-#   - Fetch OLD version from FreeBSD ports cgit (current version in ports tree)
+#   - Fetch OLD version from the FreeBSD ports GitHub mirror (current version in ports tree)
 #   - Fetch NEW version from latest NetBird GitHub release tag
 
 set -e
 
 GITHUB_REPO="netbirdio/netbird"
-PORTS_CGIT_URL="https://cgit.freebsd.org/ports/plain/security/netbird/Makefile"
+PORTS_MAKEFILE_URL="https://raw.githubusercontent.com/freebsd/freebsd-ports/main/security/netbird/Makefile"
 
 fetch_current_ports_version() {
     echo "Fetching current version from FreeBSD ports..." >&2
     local makefile_content
-    makefile_content=$(curl -sL "$PORTS_CGIT_URL" 2>/dev/null)
+    makefile_content=$(curl -fsL --proto '=https' --proto-redir '=https' --retry 3 "$PORTS_MAKEFILE_URL" 2>/dev/null) || makefile_content=""
+    if [[ "$makefile_content" == \<* ]]; then
+        echo "Error: Received HTML instead of Makefile from ${PORTS_MAKEFILE_URL}" >&2
+        return 1
+    fi
     if [[ -z "$makefile_content" ]]; then
         echo "Error: Could not fetch Makefile from FreeBSD ports" >&2
         return 1

shared/management/client/grpc.go

@@ -187,16 +187,16 @@ func (c *GrpcClient) ready() bool {
 // Sync wraps the real client's Sync endpoint call and takes care of retries and encryption/decryption of messages
 // Blocking request. The result will be sent via msgHandler callback function
 func (c *GrpcClient) Sync(ctx context.Context, sysInfo *system.Info, msgHandler func(msg *proto.SyncResponse) error) error {
-	return c.withMgmtStream(ctx, func(ctx context.Context, serverPubKey wgtypes.Key) error {
-		return c.handleSyncStream(ctx, serverPubKey, sysInfo, msgHandler)
+	return c.withMgmtStream(ctx, func(ctx context.Context, serverPubKey wgtypes.Key, backOff backoff.BackOff) error {
+		return c.handleSyncStream(ctx, serverPubKey, sysInfo, msgHandler, backOff)
 	})
 }
 
 // Job wraps the real client's Job endpoint call and takes care of retries and encryption/decryption of messages
 // Blocking request. The result will be sent via msgHandler callback function
 func (c *GrpcClient) Job(ctx context.Context, msgHandler func(msg *proto.JobRequest) *proto.JobResponse) error {
-	return c.withMgmtStream(ctx, func(ctx context.Context, serverPubKey wgtypes.Key) error {
-		return c.handleJobStream(ctx, serverPubKey, msgHandler)
+	return c.withMgmtStream(ctx, func(ctx context.Context, serverPubKey wgtypes.Key, backOff backoff.BackOff) error {
+		return c.handleJobStream(ctx, serverPubKey, msgHandler, backOff)
 	})
 }
 
@@ -204,7 +204,7 @@ func (c *GrpcClient) Job(ctx context.Context, msgHandler func(msg *proto.JobRequ
 // It takes care of retries, connection readiness, and fetching server public key.
 func (c *GrpcClient) withMgmtStream(
 	ctx context.Context,
-	handler func(ctx context.Context, serverPubKey wgtypes.Key) error,
+	handler func(ctx context.Context, serverPubKey wgtypes.Key, backOff backoff.BackOff) error,
 ) error {
 	backOff := defaultBackoff(ctx)
 	operation := func() error {
@@ -224,7 +224,7 @@ func (c *GrpcClient) withMgmtStream(
 			return err
 		}
 
-		return handler(ctx, *serverPubKey)
+		return handler(ctx, *serverPubKey, backOff)
 	}
 
 	err := backoff.Retry(operation, backOff)
@@ -239,6 +239,7 @@ func (c *GrpcClient) handleJobStream(
 	ctx context.Context,
 	serverPubKey wgtypes.Key,
 	msgHandler func(msg *proto.JobRequest) *proto.JobResponse,
+	backOff backoff.BackOff,
 ) error {
 	ctx, cancelStream := context.WithCancel(ctx)
 	defer cancelStream()
@@ -256,6 +257,19 @@ func (c *GrpcClient) handleJobStream(
 
 	log.Debug("job stream handshake sent successfully")
 
+	// The stream is up, so reset the backoff. This matters for two reasons,
+	// both caused by the backoff lib not resetting its state on a successful
+	// connection:
+	//  1. Without a reset, after a connect followed by an error the next retry
+	//     starts from the accumulated (large) interval instead of retrying
+	//     promptly, delaying reconnection.
+	//  2. Worse, once the accumulated elapsed time exceeds MaxElapsedTime, the
+	//     next stream error makes NextBackOff() return Stop, so the retry loop
+	//     exits immediately. That error is then mislabeled unrecoverable and
+	//     bubbles up to trigger a full engine restart / data-plane teardown
+	//     instead of a silent reconnection.
+	backOff.Reset()
+
 	// Main loop: receive, process, respond
 	for {
 		jobReq, err := c.receiveJobRequest(ctx, stream, serverPubKey)
@@ -371,7 +385,7 @@ func (c *GrpcClient) sendJobResponse(
 	return nil
 }
 
-func (c *GrpcClient) handleSyncStream(ctx context.Context, serverPubKey wgtypes.Key, sysInfo *system.Info, msgHandler func(msg *proto.SyncResponse) error) error {
+func (c *GrpcClient) handleSyncStream(ctx context.Context, serverPubKey wgtypes.Key, sysInfo *system.Info, msgHandler func(msg *proto.SyncResponse) error, backOff backoff.BackOff) error {
 	ctx, cancelStream := context.WithCancel(ctx)
 	defer cancelStream()
 
@@ -390,6 +404,19 @@ func (c *GrpcClient) handleSyncStream(ctx context.Context, serverPubKey wgtypes.
 	c.notifyConnected()
 	c.setSyncStreamConnected()
 
+	// The stream is up, so reset the backoff. This matters for two reasons,
+	// both caused by the backoff lib not resetting its state on a successful
+	// connection:
+	//  1. Without a reset, after a connect followed by an error the next retry
+	//     starts from the accumulated (large) interval instead of retrying
+	//     promptly, delaying reconnection.
+	//  2. Worse, once the accumulated elapsed time exceeds MaxElapsedTime, the
+	//     next stream error makes NextBackOff() return Stop, so the retry loop
+	//     exits immediately. That error is then mislabeled unrecoverable and
+	//     bubbles up to trigger a full engine restart / data-plane teardown
+	//     instead of a silent reconnection.
+	backOff.Reset()
+
 	// blocking until error
 	err = c.receiveUpdatesEvents(stream, serverPubKey, msgHandler)
 	if err != nil {