[client] Resolve cold mgmt cache domains synchronously before DNS takeover

Async server-domain resolution risked blackholing infra DNS at bootstrap:
if OS DNS was reconfigured to route through a dead exit node before the
background resolve ran, the cache never populated.

Resolve cold domains (no cached record) synchronously while NetBird has not
yet taken over the system resolver and will serve DNS (dnsWillBeServed), so
the cache is primed via the working OS resolver before takeover. Stale and
post-takeover resolves stay async to keep the engine sync lock unblocked.

client/internal/dns/mgmt/export_test.go

@@ -0,0 +1,23 @@
+package mgmt
+
+import "time"
+
+// pendingCount returns how many initial resolves are still in flight. Test-only.
+func (m *Resolver) pendingCount() int {
+	m.mutex.RLock()
+	defer m.mutex.RUnlock()
+	return len(m.pending)
+}
+
+// waitForPendingResolves blocks until all pending resolves settle or the
+// timeout elapses, returning true if all settled. Test-only.
+func (m *Resolver) waitForPendingResolves(timeout time.Duration) bool {
+	deadline := time.Now().Add(timeout)
+	for m.pendingCount() > 0 {
+		if time.Now().After(deadline) {
+			return false
+		}
+		time.Sleep(10 * time.Millisecond)
+	}
+	return true
+}

client/internal/dns/mgmt/mgmt.go

@@ -50,17 +50,31 @@ type cachedRecord struct {
 	consecFailures    int
 }
 
+// pendingEntry marks a domain whose initial resolve is in flight, so ServeDNS
+// can wait on it instead of falling through to upstream.
+type pendingEntry struct{}
+
 // Resolver caches critical NetBird infrastructure domains.
-// records, refreshing, mgmtDomain and serverDomains are all guarded by mutex.
+// records, refreshing, pending, mgmtDomain and serverDomains are all guarded by mutex.
 type Resolver struct {
+	// ctx is the server-lifetime context for background resolves.
+	ctx context.Context
+
 	records       map[dns.Question]*cachedRecord
 	mgmtDomain    *domain.Domain
 	serverDomains *dnsconfig.ServerDomains
 	mutex         sync.RWMutex
 
+	// pending holds domains whose initial resolve is in flight, keyed by
+	// punycode FQDN (trailing dot).
+	pending map[string]pendingEntry
+
 	chain            ChainResolver
 	chainMaxPriority int
 	refreshGroup     singleflight.Group
+	// resolveGroup dedups initial (cold-cache) resolves; kept separate from
+	// refreshGroup so initial and stale-refresh flights don't collapse.
+	resolveGroup singleflight.Group
 
 	// refreshing tracks questions whose refresh is running via the OS
 	// fallback path. A ServeDNS hit for a question in this map indicates
@@ -74,10 +88,12 @@ type Resolver struct {
 }
 
 // NewResolver creates a new management domains cache resolver.
-func NewResolver() *Resolver {
+func NewResolver(ctx context.Context) *Resolver {
 	return &Resolver{
+		ctx:        ctx,
 		records:    make(map[dns.Question]*cachedRecord),
 		refreshing: make(map[dns.Question]*atomic.Bool),
+		pending:    make(map[string]pendingEntry),
 		cacheTTL:   resolveCacheTTL(),
 	}
 }
@@ -117,6 +133,7 @@ func (m *Resolver) ServeDNS(w dns.ResponseWriter, r *dns.Msg) {
 	m.mutex.RLock()
 	cached, found := m.records[question]
 	inflight := m.refreshing[question]
+	_, isPending := m.pending[question.Name]
 	var shouldRefresh bool
 	if found {
 		stale := time.Since(cached.cachedAt) > m.cacheTTL
@@ -126,8 +143,17 @@ func (m *Resolver) ServeDNS(w dns.ResponseWriter, r *dns.Msg) {
 	m.mutex.RUnlock()
 
 	if !found {
-		m.continueToNext(w, r)
-		return
+		// Registered but not resolved yet: wait on the in-flight resolve
+		// rather than falling through to (possibly dead) upstream.
+		if isPending && m.awaitPendingResolve(question.Name) {
+			m.mutex.RLock()
+			cached, found = m.records[question]
+			m.mutex.RUnlock()
+		}
+		if !found {
+			m.continueToNext(w, r)
+			return
+		}
 	}
 
 	if inflight != nil && inflight.CompareAndSwap(false, true) {
@@ -467,6 +493,13 @@ func (m *Resolver) RemoveDomain(d domain.Domain) error {
 	return nil
 }
 
+// RequestedDomains returns the cacheable infrastructure domains (signal, relay,
+// STUN, TURN; flow excluded) so the cache handler can be registered for them
+// before resolution completes.
+func (m *Resolver) RequestedDomains(serverDomains dnsconfig.ServerDomains) domain.List {
+	return m.extractDomainsFromServerDomains(serverDomains)
+}
+
 // GetCachedDomains returns a list of all cached domains.
 func (m *Resolver) GetCachedDomains() domain.List {
 	m.mutex.RLock()
@@ -486,10 +519,12 @@ func (m *Resolver) GetCachedDomains() domain.List {
 	return domains
 }
 
-// UpdateFromServerDomains updates the cache with server domains from network configuration.
-// It merges new domains with existing ones, replacing entire domain types when updated.
-// Empty updates are ignored to prevent clearing infrastructure domains during partial updates.
-func (m *Resolver) UpdateFromServerDomains(ctx context.Context, serverDomains dnsconfig.ServerDomains) (domain.List, error) {
+// UpdateFromServerDomains merges server domains into the cache and resolves
+// them. New types replace whole types; empty updates are ignored. Resolution is
+// async (off the caller's sync lock) except for cold domains when dnsWillBeServed
+// and takeover is pending, which kickoffResolve primes synchronously. ctx is the
+// server lifetime, so a fast sync won't cancel resolves but Stop will.
+func (m *Resolver) UpdateFromServerDomains(ctx context.Context, serverDomains dnsconfig.ServerDomains, dnsWillBeServed bool) (domain.List, error) {
 	newDomains := m.extractDomainsFromServerDomains(serverDomains)
 	var removedDomains domain.List
 
@@ -507,11 +542,136 @@ func (m *Resolver) UpdateFromServerDomains(ctx context.Context, serverDomains dn
 		removedDomains = m.removeStaleDomains(currentDomains, allDomains)
 	}
 
-	m.addNewDomains(ctx, newDomains)
+	m.kickoffResolve(ctx, newDomains, dnsWillBeServed)
 
 	return removedDomains, nil
 }
 
+// kickoffResolve resolves each unresolved domain, skipping fresh/in-flight ones.
+// Cold domains resolve synchronously only before takeover (no upstream root
+// handler) and when dnsWillBeServed, to prime the cache via the working OS
+// resolver before OS DNS routes through the tunnel; otherwise async.
+func (m *Resolver) kickoffResolve(ctx context.Context, domains domain.List, dnsWillBeServed bool) {
+	m.mutex.RLock()
+	chain := m.chain
+	maxPriority := m.chainMaxPriority
+	m.mutex.RUnlock()
+	preTakeover := chain == nil || !chain.HasRootHandlerAtOrBelow(maxPriority)
+
+	for _, d := range domains {
+		dnsName := strings.ToLower(dns.Fqdn(d.PunycodeString()))
+
+		m.mutex.Lock()
+		_, hasPending := m.pending[dnsName]
+		fresh := m.hasFreshRecordLocked(dnsName)
+		cold := !m.hasAnyRecordLocked(dnsName)
+		if !hasPending && !fresh {
+			m.pending[dnsName] = pendingEntry{}
+		}
+		m.mutex.Unlock()
+
+		if hasPending || fresh {
+			continue
+		}
+
+		if cold && preTakeover && dnsWillBeServed {
+			m.resolveInitial(ctx, d, dnsName)
+			continue
+		}
+
+		m.scheduleInitialResolve(ctx, d, dnsName)
+	}
+}
+
+// resolveInitial resolves a cold domain synchronously, deduped via resolveGroup
+// so a concurrent ServeDNS await joins the same flight. Clears pending when done.
+func (m *Resolver) resolveInitial(ctx context.Context, d domain.Domain, dnsName string) {
+	key := "initial|" + dnsName
+	_, _, _ = m.resolveGroup.Do(key, func() (any, error) {
+		defer m.clearPending(dnsName)
+		if err := m.AddDomain(ctx, d); err != nil {
+			log.Warnf("initial resolve mgmt domain=%s: %v", d.SafeString(), err)
+			return struct{}{}, err
+		}
+		log.Debugf("added/updated management cache domain=%s", d.SafeString())
+		return struct{}{}, nil
+	})
+}
+
+// scheduleInitialResolve runs AddDomain in the background, deduped per domain
+// by resolveGroup, clearing the pending marker when it finishes. ctx is the
+// server-lifetime context so a Stop cancels in-flight resolves.
+func (m *Resolver) scheduleInitialResolve(ctx context.Context, d domain.Domain, dnsName string) {
+	key := "initial|" + dnsName
+	_ = m.resolveGroup.DoChan(key, func() (any, error) {
+		defer m.clearPending(dnsName)
+		if err := m.AddDomain(ctx, d); err != nil {
+			log.Warnf("failed to add/update domain=%s: %v", d.SafeString(), err)
+			return struct{}{}, err
+		}
+		log.Debugf("added/updated management cache domain=%s", d.SafeString())
+		return struct{}{}, nil
+	})
+}
+
+// hasFreshRecordLocked reports whether a non-stale A or AAAA record exists for
+// the name. Caller holds m.mutex.
+func (m *Resolver) hasFreshRecordLocked(dnsName string) bool {
+	for _, qtype := range []uint16{dns.TypeA, dns.TypeAAAA} {
+		q := dns.Question{Name: dnsName, Qtype: qtype, Qclass: dns.ClassINET}
+		if c, ok := m.records[q]; ok && time.Since(c.cachedAt) <= m.cacheTTL {
+			return true
+		}
+	}
+	return false
+}
+
+// hasAnyRecordLocked reports whether any A or AAAA record exists for the name,
+// fresh or stale. Caller holds m.mutex.
+func (m *Resolver) hasAnyRecordLocked(dnsName string) bool {
+	for _, qtype := range []uint16{dns.TypeA, dns.TypeAAAA} {
+		q := dns.Question{Name: dnsName, Qtype: qtype, Qclass: dns.ClassINET}
+		if _, ok := m.records[q]; ok {
+			return true
+		}
+	}
+	return false
+}
+
+func (m *Resolver) clearPending(dnsName string) {
+	m.mutex.Lock()
+	delete(m.pending, dnsName)
+	m.mutex.Unlock()
+}
+
+// awaitPendingResolve joins the in-flight resolve for dnsName (bounded by
+// dnsTimeout) and reports whether a record became available.
+func (m *Resolver) awaitPendingResolve(dnsName string) bool {
+	key := "initial|" + dnsName
+	d, err := domain.FromString(strings.TrimSuffix(dnsName, "."))
+	if err != nil {
+		return false
+	}
+
+	ch := m.resolveGroup.DoChan(key, func() (any, error) {
+		defer m.clearPending(dnsName)
+		if err := m.AddDomain(m.ctx, d); err != nil {
+			return struct{}{}, err
+		}
+		return struct{}{}, nil
+	})
+
+	select {
+	case <-ch:
+	case <-time.After(dnsTimeout):
+		return false
+	}
+
+	m.mutex.RLock()
+	defer m.mutex.RUnlock()
+	return m.hasFreshRecordLocked(dnsName)
+}
+
 // removeStaleDomains removes cached domains not present in the target domain list.
 // Management domains are preserved and never removed during server domain updates.
 func (m *Resolver) removeStaleDomains(currentDomains, newDomains domain.List) domain.List {
@@ -577,17 +737,6 @@ func (m *Resolver) isManagementDomain(domain domain.Domain) bool {
 	return m.mgmtDomain != nil && domain == *m.mgmtDomain
 }
 
-// addNewDomains resolves and caches all domains from the update
-func (m *Resolver) addNewDomains(ctx context.Context, newDomains domain.List) {
-	for _, newDomain := range newDomains {
-		if err := m.AddDomain(ctx, newDomain); err != nil {
-			log.Warnf("failed to add/update domain=%s: %v", newDomain.SafeString(), err)
-		} else {
-			log.Debugf("added/updated management cache domain=%s", newDomain.SafeString())
-		}
-	}
-}
-
 func (m *Resolver) extractDomainsFromServerDomains(serverDomains dnsconfig.ServerDomains) domain.List {
 	var domains domain.List
 

client/internal/dns/mgmt/mgmt_refresh_test.go

@@ -130,7 +130,7 @@ func TestResolver_CacheTTLGatesRefresh(t *testing.T) {
 	q := dns.Question{Name: "mgmt.example.com.", Qtype: dns.TypeA, Qclass: dns.ClassINET}
 
 	t.Run("short TTL treats entry as stale and refreshes", func(t *testing.T) {
-		r := NewResolver()
+		r := NewResolver(context.Background())
 		r.cacheTTL = 10 * time.Millisecond
 		chain := newFakeChain()
 		chain.setAnswer(q.Name, dns.TypeA, "10.0.0.2")
@@ -146,7 +146,7 @@ func TestResolver_CacheTTLGatesRefresh(t *testing.T) {
 	})
 
 	t.Run("long TTL keeps entry fresh and skips refresh", func(t *testing.T) {
-		r := NewResolver()
+		r := NewResolver(context.Background())
 		r.cacheTTL = time.Hour
 		chain := newFakeChain()
 		chain.setAnswer(q.Name, dns.TypeA, "10.0.0.2")
@@ -162,7 +162,7 @@ func TestResolver_CacheTTLGatesRefresh(t *testing.T) {
 }
 
 func TestResolver_ServeFresh_NoRefresh(t *testing.T) {
-	r := NewResolver()
+	r := NewResolver(context.Background())
 	chain := newFakeChain()
 	chain.setAnswer("mgmt.example.com.", dns.TypeA, "10.0.0.2")
 	r.SetChainResolver(chain, 50)
@@ -183,7 +183,7 @@ func TestResolver_ServeFresh_NoRefresh(t *testing.T) {
 }
 
 func TestResolver_StaleTriggersAsyncRefresh(t *testing.T) {
-	r := NewResolver()
+	r := NewResolver(context.Background())
 	chain := newFakeChain()
 	chain.setAnswer("mgmt.example.com.", dns.TypeA, "10.0.0.2")
 	r.SetChainResolver(chain, 50)
@@ -213,7 +213,7 @@ func TestResolver_StaleTriggersAsyncRefresh(t *testing.T) {
 }
 
 func TestResolver_ConcurrentStaleHitsCollapseRefresh(t *testing.T) {
-	r := NewResolver()
+	r := NewResolver(context.Background())
 	chain := newFakeChain()
 	chain.setAnswer("mgmt.example.com.", dns.TypeA, "10.0.0.2")
 
@@ -262,7 +262,7 @@ func TestResolver_ConcurrentStaleHitsCollapseRefresh(t *testing.T) {
 }
 
 func TestResolver_RefreshFailureArmsBackoff(t *testing.T) {
-	r := NewResolver()
+	r := NewResolver(context.Background())
 	chain := newFakeChain()
 	chain.err = errors.New("boom")
 	r.SetChainResolver(chain, 50)
@@ -299,7 +299,7 @@ func TestResolver_RefreshFailureArmsBackoff(t *testing.T) {
 }
 
 func TestResolver_NoRootHandler_SkipsChain(t *testing.T) {
-	r := NewResolver()
+	r := NewResolver(context.Background())
 	chain := newFakeChain()
 	chain.hasRoot = false
 	chain.setAnswer("mgmt.example.com.", dns.TypeA, "10.0.0.2")
@@ -320,7 +320,7 @@ func TestResolver_ServeDuringRefreshSetsLoopFlag(t *testing.T) {
 	// ServeDNS being invoked for a question while a refresh for that question
 	// is inflight indicates a resolver loop (OS resolver sent the recursive
 	// query back to us). The inflightRefresh.loopLoggedOnce flag must be set.
-	r := NewResolver()
+	r := NewResolver(context.Background())
 
 	q := dns.Question{Name: "mgmt.example.com.", Qtype: dns.TypeA, Qclass: dns.ClassINET}
 	r.records[q] = &cachedRecord{
@@ -346,7 +346,7 @@ func TestResolver_ServeDuringRefreshSetsLoopFlag(t *testing.T) {
 }
 
 func TestResolver_LoopFlagOnlyTrippedOncePerRefresh(t *testing.T) {
-	r := NewResolver()
+	r := NewResolver(context.Background())
 
 	q := dns.Question{Name: "mgmt.example.com.", Qtype: dns.TypeA, Qclass: dns.ClassINET}
 	r.records[q] = &cachedRecord{
@@ -373,7 +373,7 @@ func TestResolver_LoopFlagOnlyTrippedOncePerRefresh(t *testing.T) {
 }
 
 func TestResolver_NoLoopFlagWhenNotRefreshing(t *testing.T) {
-	r := NewResolver()
+	r := NewResolver(context.Background())
 
 	q := dns.Question{Name: "mgmt.example.com.", Qtype: dns.TypeA, Qclass: dns.ClassINET}
 	r.records[q] = &cachedRecord{
@@ -393,7 +393,7 @@ func TestResolver_NoLoopFlagWhenNotRefreshing(t *testing.T) {
 }
 
 func TestResolver_AddDomain_UsesChainWhenRootRegistered(t *testing.T) {
-	r := NewResolver()
+	r := NewResolver(context.Background())
 	chain := newFakeChain()
 	chain.setAnswer("mgmt.example.com.", dns.TypeA, "10.0.0.2")
 	chain.setAnswer("mgmt.example.com.", dns.TypeAAAA, "fd00::2")

client/internal/dns/mgmt/mgmt_test.go

@@ -17,7 +17,7 @@ import (
 )
 
 func TestResolver_NewResolver(t *testing.T) {
-	resolver := NewResolver()
+	resolver := NewResolver(context.Background())
 
 	assert.NotNil(t, resolver)
 	assert.NotNil(t, resolver.records)
@@ -49,7 +49,7 @@ func TestResolveCacheTTL(t *testing.T) {
 
 func TestNewResolver_CacheTTLFromEnv(t *testing.T) {
 	t.Setenv(envMgmtCacheTTL, "7s")
-	r := NewResolver()
+	r := NewResolver(context.Background())
 	assert.Equal(t, 7*time.Second, r.cacheTTL, "NewResolver should evaluate cacheTTL once from env")
 }
 
@@ -169,7 +169,7 @@ func TestResolver_PopulateFromConfig(t *testing.T) {
 	ctx, cancel := context.WithCancel(context.Background())
 	defer cancel()
 
-	resolver := NewResolver()
+	resolver := NewResolver(context.Background())
 
 	// Test with IP address - should return error since IP addresses are rejected
 	mgmtURL, _ := url.Parse("https://127.0.0.1")
@@ -184,7 +184,7 @@ func TestResolver_PopulateFromConfig(t *testing.T) {
 }
 
 func TestResolver_ServeDNS(t *testing.T) {
-	resolver := NewResolver()
+	resolver := NewResolver(context.Background())
 	ctx := context.Background()
 
 	// Add a test domain to the cache - use example.org which is reserved for testing
@@ -284,7 +284,7 @@ func TestResolver_ServeDNS(t *testing.T) {
 }
 
 func TestResolver_GetCachedDomains(t *testing.T) {
-	resolver := NewResolver()
+	resolver := NewResolver(context.Background())
 	ctx := context.Background()
 
 	testDomain, err := domain.FromString("example.org")
@@ -304,7 +304,7 @@ func TestResolver_GetCachedDomains(t *testing.T) {
 }
 
 func TestResolver_ManagementDomainProtection(t *testing.T) {
-	resolver := NewResolver()
+	resolver := NewResolver(context.Background())
 	ctx := context.Background()
 
 	mgmtURL, _ := url.Parse("https://example.org")
@@ -325,10 +325,11 @@ func TestResolver_ManagementDomainProtection(t *testing.T) {
 		Relay:  []domain.Domain{"cloudflare.com"},
 	}
 
-	_, err = resolver.UpdateFromServerDomains(ctx, serverDomains)
+	_, err = resolver.UpdateFromServerDomains(ctx, serverDomains, true)
 	if err != nil {
 		t.Logf("Server domains update failed: %v", err)
 	}
+	resolver.waitForPendingResolves(10 * time.Second)
 
 	finalDomains := resolver.GetCachedDomains()
 
@@ -351,7 +352,7 @@ func extractDomainFromURL(u *url.URL) (domain.Domain, error) {
 }
 
 func TestResolver_EmptyUpdateDoesNotRemoveDomains(t *testing.T) {
-	resolver := NewResolver()
+	resolver := NewResolver(context.Background())
 	ctx := context.Background()
 
 	// Set up initial domains using resolvable domains
@@ -362,10 +363,11 @@ func TestResolver_EmptyUpdateDoesNotRemoveDomains(t *testing.T) {
 	}
 
 	// Add initial domains
-	_, err := resolver.UpdateFromServerDomains(ctx, initialDomains)
+	_, err := resolver.UpdateFromServerDomains(ctx, initialDomains, true)
 	if err != nil {
 		t.Skipf("Skipping test due to DNS resolution failure: %v", err)
 	}
+	resolver.waitForPendingResolves(10 * time.Second)
 
 	// Verify domains were added
 	cachedDomains := resolver.GetCachedDomains()
@@ -373,7 +375,7 @@ func TestResolver_EmptyUpdateDoesNotRemoveDomains(t *testing.T) {
 
 	// Update with empty ServerDomains (simulating partial network map update)
 	emptyDomains := dnsconfig.ServerDomains{}
-	removedDomains, err := resolver.UpdateFromServerDomains(ctx, emptyDomains)
+	removedDomains, err := resolver.UpdateFromServerDomains(ctx, emptyDomains, true)
 	assert.NoError(t, err)
 
 	// Verify no domains were removed
@@ -385,7 +387,7 @@ func TestResolver_EmptyUpdateDoesNotRemoveDomains(t *testing.T) {
 }
 
 func TestResolver_PartialUpdateReplacesOnlyUpdatedTypes(t *testing.T) {
-	resolver := NewResolver()
+	resolver := NewResolver(context.Background())
 	ctx := context.Background()
 
 	// Set up initial complete domains using resolvable domains
@@ -396,20 +398,22 @@ func TestResolver_PartialUpdateReplacesOnlyUpdatedTypes(t *testing.T) {
 	}
 
 	// Add initial domains
-	_, err := resolver.UpdateFromServerDomains(ctx, initialDomains)
+	_, err := resolver.UpdateFromServerDomains(ctx, initialDomains, true)
 	if err != nil {
 		t.Skipf("Skipping test due to DNS resolution failure: %v", err)
 	}
+	resolver.waitForPendingResolves(10 * time.Second)
 	assert.Len(t, resolver.GetCachedDomains(), 3)
 
 	// Update with partial ServerDomains (only signal domain - this should replace signal but preserve stun/turn)
 	partialDomains := dnsconfig.ServerDomains{
 		Signal: "github.com",
 	}
-	removedDomains, err := resolver.UpdateFromServerDomains(ctx, partialDomains)
+	removedDomains, err := resolver.UpdateFromServerDomains(ctx, partialDomains, true)
 	if err != nil {
 		t.Skipf("Skipping test due to DNS resolution failure: %v", err)
 	}
+	resolver.waitForPendingResolves(10 * time.Second)
 
 	// Should remove only the old signal domain
 	assert.Len(t, removedDomains, 1, "Should remove only the old signal domain")
@@ -429,7 +433,7 @@ func TestResolver_PartialUpdateReplacesOnlyUpdatedTypes(t *testing.T) {
 }
 
 func TestResolver_PartialUpdateAddsNewTypePreservesExisting(t *testing.T) {
-	resolver := NewResolver()
+	resolver := NewResolver(context.Background())
 	ctx := context.Background()
 
 	// Set up initial complete domains using resolvable domains
@@ -440,10 +444,11 @@ func TestResolver_PartialUpdateAddsNewTypePreservesExisting(t *testing.T) {
 	}
 
 	// Add initial domains
-	_, err := resolver.UpdateFromServerDomains(ctx, initialDomains)
+	_, err := resolver.UpdateFromServerDomains(ctx, initialDomains, true)
 	if err != nil {
 		t.Skipf("Skipping test due to DNS resolution failure: %v", err)
 	}
+	resolver.waitForPendingResolves(10 * time.Second)
 	assert.Len(t, resolver.GetCachedDomains(), 3)
 
 	// Update with partial ServerDomains (only flow domain - flow is intentionally excluded from
@@ -451,10 +456,11 @@ func TestResolver_PartialUpdateAddsNewTypePreservesExisting(t *testing.T) {
 	partialDomains := dnsconfig.ServerDomains{
 		Flow: "github.com",
 	}
-	removedDomains, err := resolver.UpdateFromServerDomains(ctx, partialDomains)
+	removedDomains, err := resolver.UpdateFromServerDomains(ctx, partialDomains, true)
 	if err != nil {
 		t.Skipf("Skipping test due to DNS resolution failure: %v", err)
 	}
+	resolver.waitForPendingResolves(10 * time.Second)
 
 	assert.Len(t, removedDomains, 0, "Should not remove any domains when only flow domain is provided")
 

client/internal/dns/server.go

@@ -282,7 +282,7 @@ func newDefaultServer(
 	handlerChain := NewHandlerChain()
 	ctx, stop := context.WithCancel(ctx)
 
-	mgmtCacheResolver := mgmt.NewResolver()
+	mgmtCacheResolver := mgmt.NewResolver(ctx)
 	mgmtCacheResolver.SetChainResolver(handlerChain, PriorityUpstream)
 
 	defaultServer := &DefaultServer{
@@ -613,7 +613,11 @@ func (s *DefaultServer) UpdateServerConfig(domains dnsconfig.ServerDomains) erro
 	defer s.mux.Unlock()
 
 	if s.mgmtCacheResolver != nil {
-		removedDomains, err := s.mgmtCacheResolver.UpdateFromServerDomains(s.ctx, domains)
+		// Mirrors the Initialize guard: without it NetBird never becomes the
+		// system resolver, so the mgmt cache is never queried and need not be
+		// primed synchronously.
+		dnsWillBeServed := !s.disableSys && !netstack.IsEnabled()
+		removedDomains, err := s.mgmtCacheResolver.UpdateFromServerDomains(s.ctx, domains, dnsWillBeServed)
 		if err != nil {
 			return fmt.Errorf("update management cache resolver: %w", err)
 		}
@@ -622,7 +626,9 @@ func (s *DefaultServer) UpdateServerConfig(domains dnsconfig.ServerDomains) erro
 			s.deregisterHandler(removedDomains.ToPunycodeList(), PriorityMgmtCache)
 		}
 
-		newDomains := s.mgmtCacheResolver.GetCachedDomains()
+		// Register for the requested domains, not just resolved ones: resolution
+		// now runs in the background, so the cache may still be empty here.
+		newDomains := s.mgmtCacheResolver.RequestedDomains(domains)
 		if len(newDomains) > 0 {
 			s.registerHandler(newDomains.ToPunycodeList(), s.mgmtCacheResolver, PriorityMgmtCache)
 		}

management/internals/shared/grpc/server.go

@@ -1205,7 +1205,7 @@ func (s *Server) SyncMeta(ctx context.Context, req *proto.EncryptedMessage) (*pr
 		return nil, msg
 	}
 
-	err = s.accountManager.SyncPeerMeta(ctx, peerKey.String(), extractPeerMeta(ctx, syncMetaReq.GetMeta()), realIP)
+	err = s.accountManager.SyncPeerMeta(ctx, peerKey.String(), extractPeerMeta(ctx, syncMetaReq.GetMeta()))
 	if err != nil {
 		return nil, mapError(ctx, err)
 	}
@@ -1254,10 +1254,7 @@ func (s *Server) Logout(ctx context.Context, req *proto.EncryptedMessage) (*prot
 func toProtocolChecks(ctx context.Context, postureChecks []*posture.Checks) []*proto.Checks {
 	protoChecks := make([]*proto.Checks, 0, len(postureChecks))
 	for _, postureCheck := range postureChecks {
-		check := toProtocolCheck(postureCheck)
-		if check != nil {
-			protoChecks = append(protoChecks, check)
-		}
+		protoChecks = append(protoChecks, toProtocolCheck(postureCheck))
 	}
 
 	return protoChecks
@@ -1281,9 +1278,5 @@ func toProtocolCheck(postureCheck *posture.Checks) *proto.Checks {
 		}
 	}
 
-	if len(protoCheck.Files) == 0 {
-		return nil
-	}
-
 	return protoCheck
 }

management/server/account.go

@@ -1889,12 +1889,12 @@ func domainIsUpToDate(domain string, domainCategory string, userAuth auth.UserAu
 // concurrent stream that started earlier loses the optimistic-lock race
 // in MarkPeerConnected and bails without writing.
 func (am *DefaultAccountManager) SyncAndMarkPeer(ctx context.Context, accountID string, peerPubKey string, meta nbpeer.PeerSystemMeta, realIP net.IP, syncTime time.Time) (*nbpeer.Peer, *types.NetworkMap, []*posture.Checks, int64, error) {
-	peer, netMap, postureChecks, dnsfwdPort, err := am.SyncPeer(ctx, types.PeerSync{WireGuardPubKey: peerPubKey, Meta: meta, RealIP: realIP}, accountID)
+	peer, netMap, postureChecks, dnsfwdPort, err := am.SyncPeer(ctx, types.PeerSync{WireGuardPubKey: peerPubKey, Meta: meta}, accountID)
 	if err != nil {
 		return nil, nil, nil, 0, fmt.Errorf("error syncing peer: %w", err)
 	}
 
-	if err := am.MarkPeerConnected(ctx, peerPubKey, accountID, syncTime.UnixNano(), netMap); err != nil {
+	if err := am.MarkPeerConnected(ctx, peerPubKey, realIP, accountID, syncTime.UnixNano(), netMap); err != nil {
 		log.WithContext(ctx).Warnf("failed marking peer as connected %s %v", peerPubKey, err)
 	}
 
@@ -1914,13 +1914,13 @@ func (am *DefaultAccountManager) OnPeerDisconnected(ctx context.Context, account
 	return nil
 }
 
-func (am *DefaultAccountManager) SyncPeerMeta(ctx context.Context, peerPubKey string, meta nbpeer.PeerSystemMeta, realIP net.IP) error {
+func (am *DefaultAccountManager) SyncPeerMeta(ctx context.Context, peerPubKey string, meta nbpeer.PeerSystemMeta) error {
 	accountID, err := am.Store.GetAccountIDByPeerPubKey(ctx, peerPubKey)
 	if err != nil {
 		return err
 	}
 
-	_, _, _, _, err = am.SyncPeer(ctx, types.PeerSync{WireGuardPubKey: peerPubKey, Meta: meta, RealIP: realIP, UpdateAccountPeers: true}, accountID)
+	_, _, _, _, err = am.SyncPeer(ctx, types.PeerSync{WireGuardPubKey: peerPubKey, Meta: meta, UpdateAccountPeers: true}, accountID)
 	if err != nil {
 		return err
 	}

management/server/account/manager.go

@@ -62,7 +62,7 @@ type Manager interface {
 	GetUserFromUserAuth(ctx context.Context, userAuth auth.UserAuth) (*types.User, error)
 	ListUsers(ctx context.Context, accountID string) ([]*types.User, error)
 	GetPeers(ctx context.Context, accountID, userID, nameFilter, ipFilter string) ([]*nbpeer.Peer, error)
-	MarkPeerConnected(ctx context.Context, peerKey string, accountID string, sessionStartedAt int64, nmap *types.NetworkMap) error
+	MarkPeerConnected(ctx context.Context, peerKey string, realIP net.IP, accountID string, sessionStartedAt int64, nmap *types.NetworkMap) error
 	MarkPeerDisconnected(ctx context.Context, peerKey string, accountID string, sessionStartedAt int64) error
 	DeletePeer(ctx context.Context, accountID, peerID, userID string) error
 	UpdatePeer(ctx context.Context, accountID, userID string, p *nbpeer.Peer) (*nbpeer.Peer, error)
@@ -123,7 +123,7 @@ type Manager interface {
 	GetValidatedPeers(ctx context.Context, accountID string) (map[string]struct{}, map[string]string, error)
 	SyncAndMarkPeer(ctx context.Context, accountID string, peerPubKey string, meta nbpeer.PeerSystemMeta, realIP net.IP, syncTime time.Time) (*nbpeer.Peer, *types.NetworkMap, []*posture.Checks, int64, error)
 	OnPeerDisconnected(ctx context.Context, accountID string, peerPubKey string, streamStartTime time.Time) error
-	SyncPeerMeta(ctx context.Context, peerPubKey string, meta nbpeer.PeerSystemMeta, realIP net.IP) error
+	SyncPeerMeta(ctx context.Context, peerPubKey string, meta nbpeer.PeerSystemMeta) error
 	FindExistingPostureCheck(accountID string, checks *posture.ChecksDefinition) (*posture.Checks, error)
 	GetAccountIDForPeerKey(ctx context.Context, peerKey string) (string, error)
 	GetAccountSettings(ctx context.Context, accountID string, userID string) (*types.Settings, error)

management/server/account/manager_mock.go

@@ -1323,17 +1323,17 @@ func (mr *MockManagerMockRecorder) ExtendPeerSession(ctx, peerPubKey, userID int
 }
 
 // MarkPeerConnected mocks base method.
-func (m *MockManager) MarkPeerConnected(ctx context.Context, peerKey string, accountID string, sessionStartedAt int64, nmap *types.NetworkMap) error {
+func (m *MockManager) MarkPeerConnected(ctx context.Context, peerKey string, realIP net.IP, accountID string, sessionStartedAt int64, nmap *types.NetworkMap) error {
 	m.ctrl.T.Helper()
-	ret := m.ctrl.Call(m, "MarkPeerConnected", ctx, peerKey, accountID, sessionStartedAt, nmap)
+	ret := m.ctrl.Call(m, "MarkPeerConnected", ctx, peerKey, realIP, accountID, sessionStartedAt, nmap)
 	ret0, _ := ret[0].(error)
 	return ret0
 }
 
 // MarkPeerConnected indicates an expected call of MarkPeerConnected.
-func (mr *MockManagerMockRecorder) MarkPeerConnected(ctx, peerKey, accountID, sessionStartedAt, nmap interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) MarkPeerConnected(ctx, peerKey, realIP, accountID, sessionStartedAt, nmap interface{}) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
-	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "MarkPeerConnected", reflect.TypeOf((*MockManager)(nil).MarkPeerConnected), ctx, peerKey, accountID, sessionStartedAt, nmap)
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "MarkPeerConnected", reflect.TypeOf((*MockManager)(nil).MarkPeerConnected), ctx, peerKey, realIP, accountID, sessionStartedAt, nmap)
 }
 
 // MarkPeerDisconnected mocks base method.
@@ -1586,17 +1586,17 @@ func (mr *MockManagerMockRecorder) SyncPeer(ctx, sync, accountID interface{}) *g
 }
 
 // SyncPeerMeta mocks base method.
-func (m *MockManager) SyncPeerMeta(ctx context.Context, peerPubKey string, meta peer.PeerSystemMeta, realIP net.IP) error {
+func (m *MockManager) SyncPeerMeta(ctx context.Context, peerPubKey string, meta peer.PeerSystemMeta) error {
 	m.ctrl.T.Helper()
-	ret := m.ctrl.Call(m, "SyncPeerMeta", ctx, peerPubKey, meta, realIP)
+	ret := m.ctrl.Call(m, "SyncPeerMeta", ctx, peerPubKey, meta)
 	ret0, _ := ret[0].(error)
 	return ret0
 }
 
 // SyncPeerMeta indicates an expected call of SyncPeerMeta.
-func (mr *MockManagerMockRecorder) SyncPeerMeta(ctx, peerPubKey, meta, realIP interface{}) *gomock.Call {
+func (mr *MockManagerMockRecorder) SyncPeerMeta(ctx, peerPubKey, meta interface{}) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
-	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "SyncPeerMeta", reflect.TypeOf((*MockManager)(nil).SyncPeerMeta), ctx, peerPubKey, meta, realIP)
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "SyncPeerMeta", reflect.TypeOf((*MockManager)(nil).SyncPeerMeta), ctx, peerPubKey, meta)
 }
 
 // SyncUserJWTGroups mocks base method.

management/server/account_test.go

@@ -1836,7 +1836,7 @@ func TestDefaultAccountManager_UpdatePeer_PeerLoginExpiration(t *testing.T) {
 	accountID, err := manager.GetAccountIDByUserID(context.Background(), auth.UserAuth{UserId: userID})
 	require.NoError(t, err, "unable to get the account")
 
-	err = manager.MarkPeerConnected(context.Background(), key.PublicKey().String(), accountID, time.Now().UTC().UnixNano(), nil)
+	err = manager.MarkPeerConnected(context.Background(), key.PublicKey().String(), nil, accountID, time.Now().UTC().UnixNano(), nil)
 	require.NoError(t, err, "unable to mark peer connected")
 
 	_, err = manager.UpdateAccountSettings(context.Background(), accountID, userID, &types.Settings{
@@ -1907,7 +1907,7 @@ func TestDefaultAccountManager_MarkPeerConnected_PeerLoginExpiration(t *testing.
 	require.NoError(t, err, "unable to get the account")
 
 	// when we mark peer as connected, the peer login expiration routine should trigger
-	err = manager.MarkPeerConnected(context.Background(), key.PublicKey().String(), accountID, time.Now().UTC().UnixNano(), nil)
+	err = manager.MarkPeerConnected(context.Background(), key.PublicKey().String(), nil, accountID, time.Now().UTC().UnixNano(), nil)
 	require.NoError(t, err, "unable to mark peer connected")
 
 	failed := waitTimeout(wg, time.Second)
@@ -1935,7 +1935,7 @@ func TestDefaultAccountManager_OnPeerDisconnected_LastSeenCheck(t *testing.T) {
 
 	t.Run("disconnect peer when session token matches", func(t *testing.T) {
 		streamStartTime := time.Now().UTC()
-		err = manager.MarkPeerConnected(context.Background(), peerPubKey, accountID, streamStartTime.UnixNano(), nil)
+		err = manager.MarkPeerConnected(context.Background(), peerPubKey, nil, accountID, streamStartTime.UnixNano(), nil)
 		require.NoError(t, err, "unable to mark peer connected")
 
 		peer, err := manager.Store.GetPeerByPeerPubKey(context.Background(), store.LockingStrengthNone, peerPubKey)
@@ -1956,7 +1956,7 @@ func TestDefaultAccountManager_OnPeerDisconnected_LastSeenCheck(t *testing.T) {
 	t.Run("skip disconnect when stored session is newer (zombie stream protection)", func(t *testing.T) {
 		// Newer stream wins on connect (sets SessionStartedAt = now ns).
 		streamStartTime := time.Now().UTC()
-		err = manager.MarkPeerConnected(context.Background(), peerPubKey, accountID, streamStartTime.UnixNano(), nil)
+		err = manager.MarkPeerConnected(context.Background(), peerPubKey, nil, accountID, streamStartTime.UnixNano(), nil)
 		require.NoError(t, err, "unable to mark peer connected")
 
 		peer, err := manager.Store.GetPeerByPeerPubKey(context.Background(), store.LockingStrengthNone, peerPubKey)
@@ -1980,7 +1980,7 @@ func TestDefaultAccountManager_OnPeerDisconnected_LastSeenCheck(t *testing.T) {
 
 	t.Run("skip stale connect when stored session is newer (blocked goroutine protection)", func(t *testing.T) {
 		node2SyncTime := time.Now().UTC()
-		err = manager.MarkPeerConnected(context.Background(), peerPubKey, accountID, node2SyncTime.UnixNano(), nil)
+		err = manager.MarkPeerConnected(context.Background(), peerPubKey, nil, accountID, node2SyncTime.UnixNano(), nil)
 		require.NoError(t, err, "node 2 should connect peer")
 
 		peer, err := manager.Store.GetPeerByPeerPubKey(context.Background(), store.LockingStrengthNone, peerPubKey)
@@ -1990,7 +1990,7 @@ func TestDefaultAccountManager_OnPeerDisconnected_LastSeenCheck(t *testing.T) {
 			"SessionStartedAt should equal node2SyncTime token")
 
 		node1StaleSyncTime := node2SyncTime.Add(-1 * time.Minute)
-		err = manager.MarkPeerConnected(context.Background(), peerPubKey, accountID, node1StaleSyncTime.UnixNano(), nil)
+		err = manager.MarkPeerConnected(context.Background(), peerPubKey, nil, accountID, node1StaleSyncTime.UnixNano(), nil)
 		require.NoError(t, err, "stale connect should not return error")
 
 		peer, err = manager.Store.GetPeerByPeerPubKey(context.Background(), store.LockingStrengthNone, peerPubKey)
@@ -2052,7 +2052,7 @@ func TestDefaultAccountManager_MarkPeerConnected_ConcurrentRace(t *testing.T) {
 			defer done.Done()
 			ready.Done()
 			start.Wait()
-			errs <- manager.MarkPeerConnected(context.Background(), peerPubKey, accountID, token, nil)
+			errs <- manager.MarkPeerConnected(context.Background(), peerPubKey, nil, accountID, token, nil)
 		}()
 	}
 
@@ -2093,7 +2093,7 @@ func TestDefaultAccountManager_UpdateAccountSettings_PeerLoginExpiration(t *test
 	account, err := manager.Store.GetAccount(context.Background(), accountID)
 	require.NoError(t, err, "unable to get the account")
 
-	err = manager.MarkPeerConnected(context.Background(), key.PublicKey().String(), accountID, time.Now().UTC().UnixNano(), nil)
+	err = manager.MarkPeerConnected(context.Background(), key.PublicKey().String(), nil, accountID, time.Now().UTC().UnixNano(), nil)
 	require.NoError(t, err, "unable to mark peer connected")
 
 	wg := &sync.WaitGroup{}

management/server/mock_server/account_mock.go

@@ -39,7 +39,7 @@ type MockAccountManager struct {
 	GetUserFromUserAuthFunc               func(ctx context.Context, userAuth auth.UserAuth) (*types.User, error)
 	ListUsersFunc                         func(ctx context.Context, accountID string) ([]*types.User, error)
 	GetPeersFunc                          func(ctx context.Context, accountID, userID, nameFilter, ipFilter string) ([]*nbpeer.Peer, error)
-	MarkPeerConnectedFunc                 func(ctx context.Context, peerKey string, accountID string, sessionStartedAt int64, nmap *types.NetworkMap) error
+	MarkPeerConnectedFunc                 func(ctx context.Context, peerKey string, realIP net.IP, accountID string, sessionStartedAt int64, nmap *types.NetworkMap) error
 	MarkPeerDisconnectedFunc              func(ctx context.Context, peerKey string, accountID string, sessionStartedAt int64) error
 	SyncAndMarkPeerFunc                   func(ctx context.Context, accountID string, peerPubKey string, meta nbpeer.PeerSystemMeta, realIP net.IP, syncTime time.Time) (*nbpeer.Peer, *types.NetworkMap, []*posture.Checks, int64, error)
 	DeletePeerFunc                        func(ctx context.Context, accountID, peerKey, userID string) error
@@ -114,7 +114,7 @@ type MockAccountManager struct {
 	GetIdpManagerFunc                     func() idp.Manager
 	UpdateIntegratedValidatorFunc         func(ctx context.Context, accountID, userID, validator string, groups []string) error
 	GroupValidationFunc                   func(ctx context.Context, accountId string, groups []string) (bool, error)
-	SyncPeerMetaFunc                      func(ctx context.Context, peerPubKey string, meta nbpeer.PeerSystemMeta, realIP net.IP) error
+	SyncPeerMetaFunc                      func(ctx context.Context, peerPubKey string, meta nbpeer.PeerSystemMeta) error
 	FindExistingPostureCheckFunc          func(accountID string, checks *posture.ChecksDefinition) (*posture.Checks, error)
 	GetAccountIDForPeerKeyFunc            func(ctx context.Context, peerKey string) (string, error)
 	GetAccountByIDFunc                    func(ctx context.Context, accountID string, userID string) (*types.Account, error)
@@ -345,9 +345,9 @@ func (am *MockAccountManager) GetAccountIDByUserID(ctx context.Context, userAuth
 }
 
 // MarkPeerConnected mock implementation of MarkPeerConnected from server.AccountManager interface
-func (am *MockAccountManager) MarkPeerConnected(ctx context.Context, peerKey string, accountID string, sessionStartedAt int64, nmap *types.NetworkMap) error {
+func (am *MockAccountManager) MarkPeerConnected(ctx context.Context, peerKey string, realIP net.IP, accountID string, sessionStartedAt int64, nmap *types.NetworkMap) error {
 	if am.MarkPeerConnectedFunc != nil {
-		return am.MarkPeerConnectedFunc(ctx, peerKey, accountID, sessionStartedAt, nmap)
+		return am.MarkPeerConnectedFunc(ctx, peerKey, realIP, accountID, sessionStartedAt, nmap)
 	}
 	return status.Errorf(codes.Unimplemented, "method MarkPeerConnected is not implemented")
 }
@@ -975,9 +975,9 @@ func (am *MockAccountManager) GroupValidation(ctx context.Context, accountId str
 }
 
 // SyncPeerMeta mocks SyncPeerMeta of the AccountManager interface
-func (am *MockAccountManager) SyncPeerMeta(ctx context.Context, peerPubKey string, meta nbpeer.PeerSystemMeta, realIP net.IP) error {
+func (am *MockAccountManager) SyncPeerMeta(ctx context.Context, peerPubKey string, meta nbpeer.PeerSystemMeta) error {
 	if am.SyncPeerMetaFunc != nil {
-		return am.SyncPeerMetaFunc(ctx, peerPubKey, meta, realIP)
+		return am.SyncPeerMetaFunc(ctx, peerPubKey, meta)
 	}
 	return status.Errorf(codes.Unimplemented, "method SyncPeerMeta is not implemented")
 }

management/server/peer.go

@@ -74,7 +74,7 @@ func (am *DefaultAccountManager) GetPeers(ctx context.Context, accountID, userID
 //
 // Disconnects use MarkPeerDisconnected and require the session to match
 // exactly; see PeerStatus.SessionStartedAt for the protocol.
-func (am *DefaultAccountManager) MarkPeerConnected(ctx context.Context, peerPubKey string, accountID string, sessionStartedAt int64, nmap *types.NetworkMap) error {
+func (am *DefaultAccountManager) MarkPeerConnected(ctx context.Context, peerPubKey string, realIP net.IP, accountID string, sessionStartedAt int64, nmap *types.NetworkMap) error {
 	start := time.Now()
 	defer func() {
 		am.metrics.AccountManagerMetrics().RecordPeerStatusUpdateDuration(telemetry.PeerStatusConnect, time.Since(start))
@@ -102,6 +102,10 @@ func (am *DefaultAccountManager) MarkPeerConnected(ctx context.Context, peerPubK
 	}
 	am.metrics.AccountManagerMetrics().CountPeerStatusUpdate(telemetry.PeerStatusConnect, telemetry.PeerStatusApplied)
 
+	if am.geo != nil && realIP != nil {
+		am.updatePeerLocationIfChanged(ctx, accountID, peer, realIP)
+	}
+
 	if err = am.schedulePeerExpirations(ctx, accountID, peer); err != nil {
 		return err
 	}
@@ -191,28 +195,24 @@ func (am *DefaultAccountManager) MarkPeerDisconnected(ctx context.Context, peerP
 	return nil
 }
 
-// resolvePeerLocation looks up the geo location for realIP, returning nil when
-// there is nothing to apply: geo disabled, no real IP, the IP is unchanged from
-// what the peer already has, or the lookup failed. Geo lookups are skipped on
-// same-IP reconnects since they are comparatively expensive. The returned value
-// is applied by Peer.UpdateMetaIfNew so the change is persisted by its peer save.
-func (am *DefaultAccountManager) resolvePeerLocation(ctx context.Context, peer *nbpeer.Peer, realIP net.IP) *nbpeer.Location {
-	if am.geo == nil || realIP == nil {
-		return nil
-	}
+// updatePeerLocationIfChanged refreshes the geolocation on a separate
+// row update, only when the connection IP actually changed. Geo lookups
+// are expensive so we skip same-IP reconnects.
+func (am *DefaultAccountManager) updatePeerLocationIfChanged(ctx context.Context, accountID string, peer *nbpeer.Peer, realIP net.IP) {
 	if peer.Location.ConnectionIP != nil && peer.Location.ConnectionIP.Equal(realIP) {
-		return nil
+		return
 	}
 	location, err := am.geo.Lookup(realIP)
 	if err != nil {
 		log.WithContext(ctx).Warnf("failed to get location for peer %s realip: [%s]: %v", peer.ID, realIP.String(), err)
-		return nil
+		return
 	}
-	return &nbpeer.Location{
-		ConnectionIP: realIP,
-		CountryCode:  location.Country.ISOCode,
-		CityName:     location.City.Names.En,
-		GeoNameID:    location.City.GeonameID,
+	peer.Location.ConnectionIP = realIP
+	peer.Location.CountryCode = location.Country.ISOCode
+	peer.Location.CityName = location.City.Names.En
+	peer.Location.GeoNameID = location.City.GeonameID
+	if err := am.Store.SavePeerLocation(ctx, accountID, peer); err != nil {
+		log.WithContext(ctx).Warnf("could not store location for peer %s: %s", peer.ID, err)
 	}
 }
 
@@ -980,8 +980,7 @@ func getPeerIPDNSLabel(ip netip.Addr, peerHostName string) (string, error) {
 // SyncPeer checks whether peer is eligible for receiving NetworkMap (authenticated) and returns its NetworkMap if eligible
 func (am *DefaultAccountManager) SyncPeer(ctx context.Context, sync types.PeerSync, accountID string) (*nbpeer.Peer, *types.NetworkMap, []*posture.Checks, int64, error) {
 	var peer *nbpeer.Peer
-	var ipv6CapabilityChanged bool
-	var metaDiff nbpeer.MetaDiff
+	var updated, versionChanged, ipv6CapabilityChanged bool
 	var err error
 
 	settings, err := am.Store.GetAccountSettings(ctx, store.LockingStrengthNone, accountID)
@@ -1011,10 +1010,9 @@ func (am *DefaultAccountManager) SyncPeer(ctx context.Context, sync types.PeerSy
 		}
 
 		oldHasIPv6Cap := peer.HasCapability(nbpeer.PeerCapabilityIPv6Overlay)
-		newLocation := am.resolvePeerLocation(ctx, peer, sync.RealIP)
-		metaDiff = peer.UpdateMetaIfNew(ctx, sync.Meta, newLocation)
+		updated, versionChanged = peer.UpdateMetaIfNew(ctx, sync.Meta)
 		ipv6CapabilityChanged = oldHasIPv6Cap != peer.HasCapability(nbpeer.PeerCapabilityIPv6Overlay)
-		if metaDiff.Updated() {
+		if updated {
 			am.metrics.AccountManagerMetrics().CountPeerMetUpdate()
 			log.WithContext(ctx).Tracef("peer %s metadata updated", peer.ID)
 			if err = transaction.SavePeer(ctx, accountID, peer); err != nil {
@@ -1042,10 +1040,9 @@ func (am *DefaultAccountManager) SyncPeer(ctx context.Context, sync types.PeerSy
 		return nil, nil, nil, 0, err
 	}
 
-	metaDiffAffectsPosture := posture.AffectsPosture(&metaDiff, resPostureChecks)
-	if isStatusChanged || sync.UpdateAccountPeers || ipv6CapabilityChanged || metaDiffAffectsPosture || metaDiff.VersionChanged || metaDiff.Hostname {
+	if isStatusChanged || sync.UpdateAccountPeers || ipv6CapabilityChanged || (updated && (len(resPostureChecks) > 0 || versionChanged)) {
 		changedPeerIDs := []string{peer.ID}
-		affectedPeerIDs := am.syncPeerAffectedPeers(ctx, accountID, peer.ID, nmap, peerNotValid, metaDiffAffectsPosture)
+		affectedPeerIDs := am.syncPeerAffectedPeers(ctx, accountID, peer.ID, nmap, peerNotValid, updated, len(resPostureChecks) > 0)
 		if err = am.networkMapController.OnPeersUpdated(ctx, accountID, changedPeerIDs, affectedPeerIDs); err != nil {
 			return nil, nil, nil, 0, fmt.Errorf("notify network map controller of peer update: %w", err)
 		}
@@ -1062,8 +1059,8 @@ func (am *DefaultAccountManager) SyncPeer(ctx context.Context, sync types.PeerSy
 // metadata change that flips a posture result removes this peer from others'
 // maps asymmetrically; that case (and an invalid peer, whose map is empty) falls
 // back to the resolver.
-func (am *DefaultAccountManager) syncPeerAffectedPeers(ctx context.Context, accountID, peerID string, nmap *types.NetworkMap, peerNotValid, metaChangeAffectedPosture bool) []string {
-	if peerNotValid || metaChangeAffectedPosture {
+func (am *DefaultAccountManager) syncPeerAffectedPeers(ctx context.Context, accountID, peerID string, nmap *types.NetworkMap, peerNotValid, metaUpdated, hasPostureChecks bool) []string {
+	if peerNotValid || (metaUpdated && hasPostureChecks) {
 		return am.resolveAffectedPeersForPeerChanges(ctx, am.Store, accountID, []string{peerID})
 	}
 	return affectedPeerIDsFromNetworkMap(nmap, peerID)
@@ -1173,7 +1170,7 @@ func (am *DefaultAccountManager) LoginPeer(ctx context.Context, login types.Peer
 	}
 
 	// This is needed to keep in memory for the peer config. Otherwise browser client will end in a retry loop
-	peer.Meta = login.Meta
+	peer.UpdateMetaIfNew(ctx, login.Meta)
 
 	peerGroupIDs, err = getPeerGroupIDs(ctx, am.Store, accountID, peer.ID)
 	if err != nil {

management/server/peer/peer.go

@@ -256,18 +256,14 @@ func (p *Peer) Copy() *Peer {
 	}
 }
 
-// UpdateMetaIfNew updates peer's system metadata and connection geo location if
-// new information is provided. newLocation is the geo location resolved from the
-// peer's current connection IP, or nil when there is nothing to apply (geo
-// disabled, no real IP, or the IP is unchanged); the caller owns the expensive
-// lookup and the same-IP guard. It returns a MetaDiff describing what changed;
-// diff.Updated() reports whether the peer needs to be persisted.
-func (p *Peer) UpdateMetaIfNew(ctx context.Context, meta PeerSystemMeta, newLocation *Location) MetaDiff {
+// UpdateMetaIfNew updates peer's system metadata if new information is provided
+// returns true if meta was updated, false otherwise
+func (p *Peer) UpdateMetaIfNew(ctx context.Context, meta PeerSystemMeta) (updated, versionChanged bool) {
 	if meta.isEmpty() {
-		return MetaDiff{}
+		return updated, versionChanged
 	}
 
-	versionChanged := p.Meta.WtVersion != meta.WtVersion
+	versionChanged = p.Meta.WtVersion != meta.WtVersion
 
 	// Avoid overwriting UIVersion if the update was triggered sole by the CLI client
 	if meta.UIVersion == "" {
@@ -276,177 +272,97 @@ func (p *Peer) UpdateMetaIfNew(ctx context.Context, meta PeerSystemMeta, newLoca
 
 	oldVersion := p.Meta.WtVersion
 
-	diff := diffMeta(p.Meta, meta)
-	if diff.Any() {
+	diff := metaDiff(p.Meta, meta)
+	if len(diff) != 0 {
 		p.Meta = meta
-	}
-	diff.VersionChanged = versionChanged
-
-	locationInfo := ""
-	if newLocation != nil {
-		p.Location = *newLocation
-		diff.LocationChanged = true
-		locationInfo = fmt.Sprintf("location changed to %s, ", newLocation.ConnectionIP)
+		updated = true
 	}
 
 	versionInfo := ""
-	if diff.VersionChanged {
+	if versionChanged {
 		versionInfo = fmt.Sprintf("version changed: %s -> %s, ", oldVersion, meta.WtVersion)
 	}
 
-	if diff.Any() || diff.VersionChanged || diff.LocationChanged {
+	if len(diff) > 0 || versionChanged {
 		log.WithContext(ctx).
-			Debugf("peer meta updated, %s%s%d field(s) changed: %s", versionInfo, locationInfo, len(diff.Changed), strings.Join(diff.Changed, ", "))
+			Debugf("peer meta updated, %s%d field(s) changed: %s", versionInfo, len(diff), strings.Join(diff, ", "))
 	}
 
-	return diff
-}
-
-// MetaDiff records which PeerSystemMeta fields differ between two metas. Each bool
-// maps to a single struct field, except Environment, which is split into Cloud and
-// Platform. Changed holds the human-readable `field: <old> -> <new>` entries so the
-// existing log line and isEqual can be derived from the same comparison.
-//
-// VersionChanged and LocationChanged sit outside the per-meta-field set:
-// VersionChanged tracks the WireGuard client version specifically (compared before
-// the UIVersion fixup, to signal client upgrades) and LocationChanged tracks the
-// peer's connection geo location, which lives on Peer rather than PeerSystemMeta.
-// Neither contributes an entry to Changed, so the field-coverage accounting stays
-// driven purely by the PeerSystemMeta comparison.
-type MetaDiff struct {
-	Hostname            bool
-	GoOS                bool
-	Kernel              bool
-	KernelVersion       bool
-	Core                bool
-	Platform            bool
-	OS                  bool
-	OSVersion           bool
-	WtVersion           bool
-	UIVersion           bool
-	SystemSerialNumber  bool
-	SystemProductName   bool
-	SystemManufacturer  bool
-	EnvironmentCloud    bool
-	EnvironmentPlatform bool
-	Flags               bool
-	Capabilities        bool
-	NetworkAddresses    bool
-	Files               bool
-
-	VersionChanged  bool
-	LocationChanged bool
-
-	Changed []string
-}
-
-// Any reports whether any PeerSystemMeta field changed.
-func (d MetaDiff) Any() bool {
-	return len(d.Changed) != 0
-}
-
-// Updated reports whether the peer needs to be persisted: any meta field changed
-// or the geo location changed. The version flag alone does not imply a write,
-// since a version change is also reflected in the WtVersion meta field.
-func (d MetaDiff) Updated() bool {
-	return d.Any() || d.LocationChanged || d.VersionChanged
+	return updated, versionChanged
 }
 
+// metaDiff returns a human-readable list of the fields that differ between the
+// old and new meta, each formatted as `field: <old> -> <new>`. It is the single
+// source of truth for meta comparison: isEqual reports equality as an empty
+// diff, so the log line can never disagree with the change decision. Slices are
+// cloned before sorting, so callers' meta is not mutated.
 func metaDiff(oldMeta, newMeta PeerSystemMeta) []string {
-	return diffMeta(oldMeta, newMeta).Changed
-}
-
-// diffMeta compares two metas field by field, returning both a per-field flag set
-// (for callers that need to know exactly what changed, e.g. matching against
-// posture checks) and the human-readable Changed list. It is the single source of
-// truth for meta comparison: isEqual reports equality as an empty diff, so the log
-// line, the change decision, and the flags can never disagree.
-func diffMeta(oldMeta, newMeta PeerSystemMeta) MetaDiff {
-	var d MetaDiff
+	var diff []string
 	add := func(field string, oldVal, newVal any) {
-		d.Changed = append(d.Changed, fmt.Sprintf("%s: %v -> %v", field, oldVal, newVal))
+		diff = append(diff, fmt.Sprintf("%s: %v -> %v", field, oldVal, newVal))
 	}
 
 	if oldMeta.Hostname != newMeta.Hostname {
-		d.Hostname = true
 		add("hostname", oldMeta.Hostname, newMeta.Hostname)
 	}
 	if oldMeta.GoOS != newMeta.GoOS {
-		d.GoOS = true
 		add("goos", oldMeta.GoOS, newMeta.GoOS)
 	}
 	if oldMeta.Kernel != newMeta.Kernel {
-		d.Kernel = true
 		add("kernel", oldMeta.Kernel, newMeta.Kernel)
 	}
 	if oldMeta.KernelVersion != newMeta.KernelVersion {
-		d.KernelVersion = true
 		add("kernel_version", oldMeta.KernelVersion, newMeta.KernelVersion)
 	}
 	if oldMeta.Core != newMeta.Core {
-		d.Core = true
 		add("core", oldMeta.Core, newMeta.Core)
 	}
 	if oldMeta.Platform != newMeta.Platform {
-		d.Platform = true
 		add("platform", oldMeta.Platform, newMeta.Platform)
 	}
 	if oldMeta.OS != newMeta.OS {
-		d.OS = true
 		add("os", oldMeta.OS, newMeta.OS)
 	}
 	if oldMeta.OSVersion != newMeta.OSVersion {
-		d.OSVersion = true
 		add("os_version", oldMeta.OSVersion, newMeta.OSVersion)
 	}
 	if oldMeta.WtVersion != newMeta.WtVersion {
-		d.WtVersion = true
 		add("wt_version", oldMeta.WtVersion, newMeta.WtVersion)
 	}
 	if oldMeta.UIVersion != newMeta.UIVersion {
-		d.UIVersion = true
 		add("ui_version", oldMeta.UIVersion, newMeta.UIVersion)
 	}
 	if oldMeta.SystemSerialNumber != newMeta.SystemSerialNumber {
-		d.SystemSerialNumber = true
 		add("system_serial_number", oldMeta.SystemSerialNumber, newMeta.SystemSerialNumber)
 	}
 	if oldMeta.SystemProductName != newMeta.SystemProductName {
-		d.SystemProductName = true
 		add("system_product_name", oldMeta.SystemProductName, newMeta.SystemProductName)
 	}
 	if oldMeta.SystemManufacturer != newMeta.SystemManufacturer {
-		d.SystemManufacturer = true
 		add("system_manufacturer", oldMeta.SystemManufacturer, newMeta.SystemManufacturer)
 	}
 	if oldMeta.Environment.Cloud != newMeta.Environment.Cloud {
-		d.EnvironmentCloud = true
 		add("environment_cloud", oldMeta.Environment.Cloud, newMeta.Environment.Cloud)
 	}
 	if oldMeta.Environment.Platform != newMeta.Environment.Platform {
-		d.EnvironmentPlatform = true
 		add("environment_platform", oldMeta.Environment.Platform, newMeta.Environment.Platform)
 	}
 	if !oldMeta.Flags.isEqual(newMeta.Flags) {
-		d.Flags = true
 		add("flags", fmt.Sprintf("%+v", oldMeta.Flags), fmt.Sprintf("%+v", newMeta.Flags))
 	}
 	if !capabilitiesEqual(oldMeta.Capabilities, newMeta.Capabilities) {
-		d.Capabilities = true
 		add("capabilities", oldMeta.Capabilities, newMeta.Capabilities)
 	}
 
 	if !sameMultiset(oldMeta.NetworkAddresses, newMeta.NetworkAddresses) {
-		d.NetworkAddresses = true
 		add("network_addresses", fmt.Sprintf("%v", oldMeta.NetworkAddresses), fmt.Sprintf("%v", newMeta.NetworkAddresses))
 	}
 
 	if !sameMultiset(oldMeta.Files, newMeta.Files) {
-		d.Files = true
 		add("files", fmt.Sprintf("%v", oldMeta.Files), fmt.Sprintf("%v", newMeta.Files))
 	}
 
-	return d
+	return diff
 }
 
 // sameMultiset reports whether two slices contain the same elements with the

management/server/posture/checks.go

@@ -7,7 +7,6 @@ import (
 	"regexp"
 
 	"github.com/hashicorp/go-version"
-
 	nbpeer "github.com/netbirdio/netbird/management/server/peer"
 	"github.com/netbirdio/netbird/shared/management/http/api"
 	"github.com/netbirdio/netbird/shared/management/status"
@@ -52,34 +51,6 @@ type Checks struct {
 	Checks ChecksDefinition `gorm:"serializer:json"`
 }
 
-// AffectsPosture reports whether the peer metadata changes described by diff can
-// alter the outcome of any of the given posture checks. It maps each check kind to
-// the metadata fields it inspects, so an unrelated change (e.g. a hostname update)
-// does not force a posture re-evaluation.
-func AffectsPosture(diff *nbpeer.MetaDiff, checks []*Checks) bool {
-	if diff == nil {
-		return false
-	}
-	for _, c := range checks {
-		if c.Checks.ProcessCheck != nil && diff.Files {
-			return true
-		}
-		if c.Checks.OSVersionCheck != nil && (diff.OSVersion || diff.OS || diff.KernelVersion) {
-			return true
-		}
-		if c.Checks.NBVersionCheck != nil && diff.WtVersion {
-			return true
-		}
-		if c.Checks.GeoLocationCheck != nil && diff.LocationChanged {
-			return true
-		}
-		if c.Checks.PeerNetworkRangeCheck != nil && diff.NetworkAddresses {
-			return true
-		}
-	}
-	return false
-}
-
 // ChecksDefinition contains definition of actual check
 type ChecksDefinition struct {
 	NBVersionCheck        *NBVersionCheck        `json:",omitempty"`

management/server/store/sql_store.go

@@ -581,6 +581,28 @@ func (s *SqlStore) MarkPeerDisconnectedIfSameSession(ctx context.Context, accoun
 	return result.RowsAffected > 0, nil
 }
 
+func (s *SqlStore) SavePeerLocation(ctx context.Context, accountID string, peerWithLocation *nbpeer.Peer) error {
+	// To maintain data integrity, we create a copy of the peer's location to prevent unintended updates to other fields.
+	var peerCopy nbpeer.Peer
+	// Since the location field has been migrated to JSON serialization,
+	// updating the struct ensures the correct data format is inserted into the database.
+	peerCopy.Location = peerWithLocation.Location
+
+	result := s.db.Model(&nbpeer.Peer{}).
+		Where(accountAndIDQueryCondition, accountID, peerWithLocation.ID).
+		Updates(peerCopy)
+
+	if result.Error != nil {
+		return status.Errorf(status.Internal, "failed to save peer locations to store: %v", result.Error)
+	}
+
+	if result.RowsAffected == 0 {
+		return status.Errorf(status.NotFound, peerNotFoundFMT, peerWithLocation.ID)
+	}
+
+	return nil
+}
+
 // ApproveAccountPeers marks all peers that currently require approval in the given account as approved.
 func (s *SqlStore) ApproveAccountPeers(ctx context.Context, accountID string) (int, error) {
 	result := s.db.Model(&nbpeer.Peer{}).

management/server/store/sql_store_test.go

@@ -618,6 +618,56 @@ func TestSqlStore_SavePeerStatus(t *testing.T) {
 	assert.WithinDurationf(t, newStatus.LastSeen, actual.LastSeen.UTC(), time.Millisecond, "LastSeen should be equal")
 }
 
+func TestSqlStore_SavePeerLocation(t *testing.T) {
+	store, cleanUp, err := NewTestStoreFromSQL(context.Background(), "../testdata/store.sql", t.TempDir())
+	t.Cleanup(cleanUp)
+	assert.NoError(t, err)
+
+	account, err := store.GetAccount(context.Background(), "bf1c8084-ba50-4ce7-9439-34653001fc3b")
+	require.NoError(t, err)
+
+	peer := &nbpeer.Peer{
+		AccountID: account.Id,
+		ID:        "testpeer",
+		Location: nbpeer.Location{
+			ConnectionIP: net.ParseIP("0.0.0.0"),
+			CountryCode:  "YY",
+			CityName:     "City",
+			GeoNameID:    1,
+		},
+		CreatedAt: time.Now().UTC(),
+		Meta:      nbpeer.PeerSystemMeta{},
+	}
+	// error is expected as peer is not in store yet
+	err = store.SavePeerLocation(context.Background(), account.Id, peer)
+	assert.Error(t, err)
+
+	account.Peers[peer.ID] = peer
+	err = store.SaveAccount(context.Background(), account)
+	require.NoError(t, err)
+
+	peer.Location.ConnectionIP = net.ParseIP("35.1.1.1")
+	peer.Location.CountryCode = "DE"
+	peer.Location.CityName = "Berlin"
+	peer.Location.GeoNameID = 2950159
+
+	err = store.SavePeerLocation(context.Background(), account.Id, account.Peers[peer.ID])
+	assert.NoError(t, err)
+
+	account, err = store.GetAccount(context.Background(), account.Id)
+	require.NoError(t, err)
+
+	actual := account.Peers[peer.ID].Location
+	assert.Equal(t, peer.Location, actual)
+
+	peer.ID = "non-existing-peer"
+	err = store.SavePeerLocation(context.Background(), account.Id, peer)
+	assert.Error(t, err)
+	parsedErr, ok := status.FromError(err)
+	require.True(t, ok)
+	require.Equal(t, status.NotFound, parsedErr.Type(), "should return not found error")
+}
+
 func Test_TestGetAccountByPrivateDomain(t *testing.T) {
 	if runtime.GOOS == "windows" {
 		t.Skip("The SQLite store is not properly supported by Windows yet")

management/server/store/store.go

@@ -185,6 +185,7 @@ type Store interface {
 	// recorded by the database. Returns true when the update happened,
 	// false when a newer session has taken over.
 	MarkPeerDisconnectedIfSameSession(ctx context.Context, accountID, peerID string, sessionStartedAt int64) (bool, error)
+	SavePeerLocation(ctx context.Context, accountID string, peer *nbpeer.Peer) error
 	ApproveAccountPeers(ctx context.Context, accountID string) (int, error)
 	DeletePeer(ctx context.Context, accountID string, peerID string) error
 

management/server/store/store_mock.go

@@ -2968,6 +2968,20 @@ func (mr *MockStoreMockRecorder) SavePeer(ctx, accountID, peer interface{}) *gom
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "SavePeer", reflect.TypeOf((*MockStore)(nil).SavePeer), ctx, accountID, peer)
 }
 
+// SavePeerLocation mocks base method.
+func (m *MockStore) SavePeerLocation(ctx context.Context, accountID string, peer *peer.Peer) error {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "SavePeerLocation", ctx, accountID, peer)
+	ret0, _ := ret[0].(error)
+	return ret0
+}
+
+// SavePeerLocation indicates an expected call of SavePeerLocation.
+func (mr *MockStoreMockRecorder) SavePeerLocation(ctx, accountID, peer interface{}) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "SavePeerLocation", reflect.TypeOf((*MockStore)(nil).SavePeerLocation), ctx, accountID, peer)
+}
+
 // SavePeerStatus mocks base method.
 func (m *MockStore) SavePeerStatus(ctx context.Context, accountID, peerID string, status peer.PeerStatus) error {
 	m.ctrl.T.Helper()

management/server/types/peer.go

@@ -12,9 +12,6 @@ type PeerSync struct {
 	WireGuardPubKey string
 	// Meta is the system information passed by peer, must be always present
 	Meta nbpeer.PeerSystemMeta
-	// RealIP is the peer's connection IP, used to refresh its geo location.
-	// May be nil when the request has no associated connection IP.
-	RealIP net.IP
 	// UpdateAccountPeers indicate updating account peers,
 	// which occurs when the peer's metadata is updated
 	UpdateAccountPeers bool