Compare commits

..

1 Commits

Author SHA1 Message Date
Zoltan Papp
9efa3c6579 [client] Start the restarted UI with the user's environment block (#7245)
The updater runs as LocalSystem and started netbird-ui via
CreateProcessAsUser with a nil environment, so the UI inherited the
SYSTEM environment (USERPROFILE, APPDATA pointing at systemprofile)
while running under the user's token. The WebView2-based UI exits
immediately in that state, so the UI never came back after an update.

Build the environment from the user's token with CreateEnvironmentBlock
and pass it to CreateProcessAsUser.
2026-08-19 12:19:10 +02:00
27 changed files with 264 additions and 389 deletions

View File

@@ -26,7 +26,8 @@ import (
"github.com/netbirdio/netbird/client/internal/routemanager"
"github.com/netbirdio/netbird/client/internal/stdnet"
"github.com/netbirdio/netbird/client/net"
"github.com/netbirdio/netbird/client/netevents"
"github.com/netbirdio/netbird/client/netstate"
"github.com/netbirdio/netbird/client/netsweep"
"github.com/netbirdio/netbird/client/system"
"github.com/netbirdio/netbird/formatter"
"github.com/netbirdio/netbird/route"
@@ -81,10 +82,13 @@ type Client struct {
deviceName string
uiVersion string
networkChangeListener listener.NetworkChangeListener
// netMgr outlives engine restarts: it mirrors the OS connectivity, not
// the engine lifecycle. Run and RunWithoutLogin inject its state and
// sweeper into each new ConnectClient.
netMgr *netevents.Manager
// netState outlives engine restarts: it mirrors the OS connectivity, not
// the engine lifecycle. Run and RunWithoutLogin inject it into each new
// ConnectClient, which distributes it to every reconnection loop.
netState *netstate.State
// sweeper also outlives engine restarts; NotifyNetworkChange sweeps it.
sweeper *netsweep.Sweeper
stateMu sync.RWMutex
connectClient *internal.ConnectClient
@@ -149,16 +153,16 @@ func NewClient(androidSDKVersion int, deviceName string, uiVersion string, tunAd
net.SetAndroidProtectSocketFn(tunAdapter.ProtectSocket)
system.SetIFaceDiscover(iFaceDiscover)
recorder := peer.NewRecorder("")
return &Client{
deviceName: deviceName,
uiVersion: uiVersion,
tunAdapter: tunAdapter,
iFaceDiscover: iFaceDiscover,
recorder: recorder,
recorder: peer.NewRecorder(""),
ctxCancelLock: &sync.Mutex{},
networkChangeListener: networkChangeListener,
netMgr: netevents.NewManager(recorder),
netState: netstate.New(),
sweeper: netsweep.New(),
}
}
@@ -199,9 +203,8 @@ func (c *Client) Run(platformFiles PlatformFiles, urlOpener URLOpener, isAndroid
}
// todo do not throw error in case of cancelled context
ctx = internal.CtxInitState(ctx)
connectClient := internal.NewConnectClient(ctx, cfg, c.recorder,
internal.WithNetEvents(c.netMgr))
internal.WithNetworkState(c.netState), internal.WithSweeper(c.sweeper))
c.setState(cfg, cacheDir, cfgFile, connectClient)
// This path runs the interactive SSO flow, so reaching here means the peer
// is authenticated again — release the latch Status() reports from. Clear
@@ -243,7 +246,7 @@ func (c *Client) RunWithoutLogin(platformFiles PlatformFiles, dns *DNSList, dnsR
// todo do not throw error in case of cancelled context
ctx = internal.CtxInitState(ctx)
connectClient := internal.NewConnectClient(ctx, cfg, c.recorder,
internal.WithNetEvents(c.netMgr))
internal.WithNetworkState(c.netState), internal.WithSweeper(c.sweeper))
c.setState(cfg, cacheDir, cfgFile, connectClient)
return connectClient.RunOnAndroid(c.tunAdapter, c.iFaceDiscover, c.networkChangeListener, slices.Clone(dns.items), dnsReadyListener, stateFile, cacheDir)
}
@@ -295,12 +298,9 @@ func (c *Client) GetTunSettings() (*TunSettings, error) {
// While unavailable, the internal reconnect loops suspend their attempts and
// the connection listener reports NoNetwork instead of Connecting; when
// availability returns, the loops resume immediately with a fresh backoff.
// Losing the last network also sweeps the registered connections: nothing can
// redial while offline, so the stale sockets would otherwise stay silently
// "connected" until their own timeouts and the client would keep reporting
// Connected with no network at all.
func (c *Client) SetNetworkAvailable(available bool) {
c.netMgr.SetNetworkAvailable(available)
c.netState.Set(available)
c.recorder.SetNetworkAvailable(available)
}
// NotifyNetworkChange marks the management, signal and relay connections
@@ -308,7 +308,8 @@ func (c *Client) SetNetworkAvailable(available bool) {
// whatever has not redialed on the new network by then. The engine and the
// TUN device stay untouched.
func (c *Client) NotifyNetworkChange() {
c.netMgr.NotifyNetworkChange()
c.sweeper.MarkNetworkChange()
log.Infof("network change: connections marked stale")
}
// DebugBundle generates a debug bundle, uploads it, and returns the upload key.

View File

@@ -16,14 +16,9 @@ import (
"google.golang.org/grpc"
nbnet "github.com/netbirdio/netbird/client/net"
"github.com/netbirdio/netbird/client/netevents/sweep"
"github.com/netbirdio/netbird/client/netsweep"
)
// Sweeper registers in-flight dials for the network change sweep.
type Sweeper interface {
StartDial(ctx context.Context) *sweep.Dial
}
func WithCustomDialer(_ bool, _ string) grpc.DialOption {
return grpc.WithContextDialer(dialContext)
}
@@ -31,7 +26,7 @@ func WithCustomDialer(_ bool, _ string) grpc.DialOption {
// WithSweeper dials like WithCustomDialer but registers connections and
// dials with the sweeper. Append it after WithCustomDialer: gRPC applies
// dial options in order, so the later context dialer wins.
func WithSweeper(sweeper Sweeper) grpc.DialOption {
func WithSweeper(sweeper *netsweep.Sweeper) grpc.DialOption {
return grpc.WithContextDialer(func(ctx context.Context, addr string) (net.Conn, error) {
dial := sweeper.StartDial(ctx)
defer dial.Release()

View File

@@ -1,19 +1,12 @@
package grpc
import (
"context"
"google.golang.org/grpc"
"github.com/netbirdio/netbird/client/netevents/sweep"
"github.com/netbirdio/netbird/client/netsweep"
"github.com/netbirdio/netbird/util/wsproxy/client"
)
// Sweeper registers in-flight dials for the network change sweep.
type Sweeper interface {
StartDial(ctx context.Context) *sweep.Dial
}
// WithCustomDialer returns a gRPC dial option that uses WebSocket transport for WASM/JS environments.
// The component parameter specifies the WebSocket proxy component path (e.g., "/management", "/signal").
func WithCustomDialer(tlsEnabled bool, component string) grpc.DialOption {
@@ -21,6 +14,6 @@ func WithCustomDialer(tlsEnabled bool, component string) grpc.DialOption {
}
// WithSweeper is a no-op on WASM/JS: there is no network change signal.
func WithSweeper(_ Sweeper) grpc.DialOption {
func WithSweeper(_ *netsweep.Sweeper) grpc.DialOption {
return grpc.EmptyDialOption{}
}

View File

@@ -6,19 +6,16 @@ import (
"time"
"github.com/cenkalti/backoff/v4"
)
// ChangeWatcher exposes OS network availability transitions.
type ChangeWatcher interface {
Changed() <-chan struct{}
}
"github.com/netbirdio/netbird/client/netstate"
)
// Retry mirrors backoff.Retry, but the sleep between attempts also wakes on
// OS network availability transitions: an operation cut down by a network
// change retries the moment the network settles instead of sleeping through
// the recovery. A nil watcher never fires, leaving plain backoff.Retry
// the recovery. A nil netState never fires, leaving plain backoff.Retry
// behavior.
func Retry(ctx context.Context, operation backoff.Operation, bo backoff.BackOff, watcher ChangeWatcher) error {
func Retry(ctx context.Context, operation backoff.Operation, bo backoff.BackOff, netState *netstate.State) error {
bo.Reset()
for {
err := operation()
@@ -39,14 +36,10 @@ func Retry(ctx context.Context, operation backoff.Operation, bo backoff.BackOff,
return err
}
var changed <-chan struct{}
if watcher != nil {
changed = watcher.Changed()
}
timer := time.NewTimer(next)
select {
case <-timer.C:
case <-changed:
case <-netState.Changed():
timer.Stop()
case <-ctx.Done():
timer.Stop()

View File

@@ -10,7 +10,7 @@ import (
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
"github.com/netbirdio/netbird/client/netevents/netstate"
"github.com/netbirdio/netbird/client/netstate"
)
func TestRetryWakesOnNetworkChange(t *testing.T) {

View File

@@ -38,7 +38,8 @@ import (
"github.com/netbirdio/netbird/client/internal/updater"
"github.com/netbirdio/netbird/client/internal/updater/installer"
nbnet "github.com/netbirdio/netbird/client/net"
"github.com/netbirdio/netbird/client/netevents"
"github.com/netbirdio/netbird/client/netstate"
"github.com/netbirdio/netbird/client/netsweep"
cProto "github.com/netbirdio/netbird/client/proto"
"github.com/netbirdio/netbird/client/ssh"
sshconfig "github.com/netbirdio/netbird/client/ssh/config"
@@ -72,17 +73,28 @@ type ConnectClient struct {
persistSyncResponse bool
// netEvents gates every reconnection loop on OS-reported network
// availability and sweeps connections on network change.
netEvents *netevents.Manager
// netState gates every reconnection loop on OS-reported network
// availability. Nil (the default) disables gating; mobile platforms
// inject it via WithNetworkState.
netState *netstate.State
// sweeper cuts the management, signal and relay connections on network
// change; nil disables it.
sweeper *netsweep.Sweeper
}
// ConnectClientOption configures optional ConnectClient behavior.
type ConnectClientOption func(*ConnectClient)
// WithNetEvents injects the OS network event handling.
func WithNetEvents(events *netevents.Manager) ConnectClientOption {
return func(c *ConnectClient) { c.netEvents = events }
// WithNetworkState injects the OS network availability state that gates every
// reconnection loop; without it gating is disabled.
func WithNetworkState(netState *netstate.State) ConnectClientOption {
return func(c *ConnectClient) { c.netState = netState }
}
// WithSweeper injects the network change sweeper.
func WithSweeper(sweeper *netsweep.Sweeper) ConnectClientOption {
return func(c *ConnectClient) { c.sweeper = sweeper }
}
func NewConnectClient(
@@ -293,7 +305,7 @@ func (c *ConnectClient) run(mobileDependency MobileDependency, runningChan chan
}
// suspend connection attempts while the OS reports no usable network
if waited, err := c.netEvents.Wait(c.ctx); err != nil {
if waited, err := c.netState.Wait(c.ctx); err != nil {
return nil
} else if waited {
backOff.Reset()
@@ -311,7 +323,7 @@ func (c *ConnectClient) run(mobileDependency MobileDependency, runningChan chan
log.Debugf("connecting to the Management service %s", c.config.ManagementURL.Host)
mgmClient, err := mgm.NewClient(engineCtx, c.config.ManagementURL.Host, myPrivateKey, mgmTlsEnabled,
mgm.WithNetEvents(c.netEvents))
mgm.WithNetworkState(c.netState), mgm.WithSweeper(c.sweeper))
if err != nil {
// On daemon shutdown / Down() the parent context is cancelled
// and the dial fails with "context canceled". Wrapping that
@@ -386,7 +398,7 @@ func (c *ConnectClient) run(mobileDependency MobileDependency, runningChan chan
}()
// with the global Netbird config in hand connect (just a connection, no stream yet) Signal
signalClient, err := connectToSignal(engineCtx, loginResp.GetNetbirdConfig(), myPrivateKey, c.netEvents)
signalClient, err := connectToSignal(engineCtx, loginResp.GetNetbirdConfig(), myPrivateKey, c.netState, c.sweeper)
if err != nil {
log.Error(err)
return wrapErr(err)
@@ -423,7 +435,7 @@ func (c *ConnectClient) run(mobileDependency MobileDependency, runningChan chan
}
relayManager := relayClient.NewManager(engineCtx, relayURLs, myPrivateKey.PublicKey().String(), engineConfig.MTU,
relayClient.WithNetEvents(c.netEvents))
relayClient.WithNetworkState(c.netState), relayClient.WithSweeper(c.sweeper))
c.statusRecorder.SetRelayMgr(relayManager)
if len(relayURLs) > 0 {
if token != nil {
@@ -451,7 +463,7 @@ func (c *ConnectClient) run(mobileDependency MobileDependency, runningChan chan
UpdateManager: c.updateManager,
ClientMetrics: c.clientMetrics,
MetricsCtx: c.ctx,
NetState: c.netEvents,
NetState: c.netState,
}, mobileDependency)
engine.SetSyncResponsePersistence(c.persistSyncResponse)
c.engine = engine
@@ -711,7 +723,7 @@ func selectMTU(localMTU uint16, peerMTU int32) uint16 {
}
// connectToSignal creates Signal Service client and established a connection
func connectToSignal(ctx context.Context, wtConfig *mgmProto.NetbirdConfig, ourPrivateKey wgtypes.Key, netEvents *netevents.Manager) (*signal.GrpcClient, error) {
func connectToSignal(ctx context.Context, wtConfig *mgmProto.NetbirdConfig, ourPrivateKey wgtypes.Key, netState *netstate.State, sweeper *netsweep.Sweeper) (*signal.GrpcClient, error) {
var sigTLSEnabled bool
if wtConfig.Signal.Protocol == mgmProto.HostConfig_HTTPS {
sigTLSEnabled = true
@@ -720,7 +732,7 @@ func connectToSignal(ctx context.Context, wtConfig *mgmProto.NetbirdConfig, ourP
}
signalClient, err := signal.NewClient(ctx, wtConfig.Signal.Uri, ourPrivateKey, sigTLSEnabled,
signal.WithNetEvents(netEvents))
signal.WithNetworkState(netState), signal.WithSweeper(sweeper))
if err != nil {
log.Errorf("error while connecting to the Signal Exchange Service %s: %s", wtConfig.Signal.Uri, err)
return nil, gstatus.Errorf(codes.FailedPrecondition, "failed connecting to Signal Service : %s", err)

View File

@@ -59,7 +59,7 @@ import (
"github.com/netbirdio/netbird/client/internal/syncstore"
"github.com/netbirdio/netbird/client/internal/updater"
"github.com/netbirdio/netbird/client/jobexec"
"github.com/netbirdio/netbird/client/netevents"
"github.com/netbirdio/netbird/client/netstate"
cProto "github.com/netbirdio/netbird/client/proto"
"github.com/netbirdio/netbird/client/system"
nbdns "github.com/netbirdio/netbird/dns"
@@ -184,7 +184,7 @@ type EngineServices struct {
MetricsCtx context.Context
// NetState gates the reconnection loops on OS-reported network
// availability; nil disables gating.
NetState *netevents.Manager
NetState *netstate.State
}
// Engine is a mechanism responsible for reacting on Signal and Management stream events and managing connections to the remote peers.
@@ -210,7 +210,7 @@ type Engine struct {
// netState gates the peer reconnection guards on OS-reported network
// availability; nil disables gating.
netState *netevents.Manager
netState *netstate.State
// STUNs is a list of STUN servers used by ICE
STUNs []*stun.URI

View File

@@ -26,7 +26,7 @@ import (
"github.com/netbirdio/netbird/client/internal/portforward"
"github.com/netbirdio/netbird/client/internal/rosenpass"
"github.com/netbirdio/netbird/client/internal/stdnet"
"github.com/netbirdio/netbird/client/netevents"
"github.com/netbirdio/netbird/client/netstate"
"github.com/netbirdio/netbird/route"
relayClient "github.com/netbirdio/netbird/shared/relay/client"
)
@@ -97,7 +97,7 @@ type ConnConfig struct {
// NetworkState gates the reconnection guard on OS-reported network
// availability; nil disables gating.
NetworkState *netevents.Manager
NetworkState *netstate.State
}
type Conn struct {

View File

@@ -6,6 +6,8 @@ import (
"github.com/cenkalti/backoff/v4"
log "github.com/sirupsen/logrus"
"github.com/netbirdio/netbird/client/netstate"
)
// ConnStatus represents the connection state as seen by the guard.
@@ -22,12 +24,6 @@ const (
type connStatusFunc func() ConnStatus
// NetworkWatcher is the availability view the guard gates reconnects on.
type NetworkWatcher interface {
IsOnline() bool
Changed() <-chan struct{}
}
// Guard is responsible for the reconnection logic.
// It will trigger to send an offer to the peer then has connection issues.
// Watch these events:
@@ -43,14 +39,14 @@ type Guard struct {
srWatcher *SRWatcher
// netState gates reconnect attempts on OS-reported network availability;
// nil disables gating.
netState NetworkWatcher
netState *netstate.State
relayedConnDisconnected chan struct{}
iCEConnDisconnected chan struct{}
}
// NewGuard creates a reconnection guard for a peer connection. A nil netState
// disables network availability gating.
func NewGuard(log *log.Entry, isConnectedFn connStatusFunc, timeout time.Duration, srWatcher *SRWatcher, netState NetworkWatcher) *Guard {
func NewGuard(log *log.Entry, isConnectedFn connStatusFunc, timeout time.Duration, srWatcher *SRWatcher, netState *netstate.State) *Guard {
return &Guard{
log: log,
isConnectedOnAllWay: isConnectedFn,
@@ -108,17 +104,14 @@ func (g *Guard) reconnectLoopWithRetry(ctx context.Context, callback func()) {
iceState := &iceRetryState{log: g.log}
defer iceState.reset()
var netChanged <-chan struct{}
if g.netState != nil {
netChanged = g.netState.Changed()
}
netChanged := g.netState.Changed()
for {
select {
case <-tickerChannel:
// skip attempts while the OS reports no usable network; the
// netChanged case below resumes the loop once it returns
if g.netState != nil && !g.netState.IsOnline() {
if !g.netState.IsOnline() {
continue
}
switch g.isConnectedOnAllWay() {

View File

@@ -9,7 +9,7 @@ import (
log "github.com/sirupsen/logrus"
"github.com/netbirdio/netbird/client/internal/peer/ice"
"github.com/netbirdio/netbird/client/netevents/netstate"
"github.com/netbirdio/netbird/client/netstate"
)
// newTestGuardWithNetState builds a guard with a realistic MaxInterval: the

View File

@@ -307,6 +307,16 @@ func startUIInSession(uiPath string, sessionID uint32) error {
}
}()
var env *uint16
if err := windows.CreateEnvironmentBlock(&env, primaryToken, false); err != nil {
return fmt.Errorf("create environment block: %w", err)
}
defer func() {
if err := windows.DestroyEnvironmentBlock(env); err != nil {
log.Warnf("failed to destroy environment block: %v", err)
}
}()
// Prepare startup info
var si windows.StartupInfo
si.Cb = uint32(unsafe.Sizeof(si))
@@ -329,7 +339,7 @@ func startUIInSession(uiPath string, sessionID uint32) error {
nil,
false,
creationFlags,
nil,
env,
nil,
&si,
&pi,

View File

@@ -22,7 +22,8 @@ import (
"github.com/netbirdio/netbird/client/internal/listener"
"github.com/netbirdio/netbird/client/internal/peer"
"github.com/netbirdio/netbird/client/internal/profilemanager"
"github.com/netbirdio/netbird/client/netevents"
"github.com/netbirdio/netbird/client/netstate"
"github.com/netbirdio/netbird/client/netsweep"
"github.com/netbirdio/netbird/client/system"
"github.com/netbirdio/netbird/formatter"
"github.com/netbirdio/netbird/route"
@@ -83,10 +84,12 @@ type Client struct {
onHostDnsFn func([]string)
dnsManager dns.IosDnsManager
loginComplete bool
// netMgr outlives engine restarts: it mirrors the OS connectivity, not
// the engine lifecycle. Run injects its state and sweeper into each new
// ConnectClient.
netMgr *netevents.Manager
// netState outlives engine restarts: it mirrors the OS connectivity, not
// the engine lifecycle. Run injects it into each new ConnectClient, which
// distributes it to every reconnection loop.
netState *netstate.State
// sweeper also outlives engine restarts; NotifyNetworkChange sweeps it.
sweeper *netsweep.Sweeper
// preloadedConfig holds config loaded from JSON (used on tvOS where file writes are blocked)
preloadedConfig *profilemanager.Config
@@ -97,7 +100,6 @@ type Client struct {
// NewClient instantiate a new Client
func NewClient(cfgFile, stateFile, cacheDir, logFilePath, deviceName string, osVersion string, osName string, networkChangeListener NetworkChangeListener, dnsManager DnsManager) *Client {
recorder := peer.NewRecorder("")
return &Client{
cfgFile: cfgFile,
stateFile: stateFile,
@@ -106,11 +108,12 @@ func NewClient(cfgFile, stateFile, cacheDir, logFilePath, deviceName string, osV
deviceName: deviceName,
osName: osName,
osVersion: osVersion,
recorder: recorder,
recorder: peer.NewRecorder(""),
ctxCancelLock: &sync.Mutex{},
networkChangeListener: networkChangeListener,
dnsManager: dnsManager,
netMgr: netevents.NewManager(recorder),
netState: netstate.New(),
sweeper: netsweep.New(),
}
}
@@ -187,7 +190,7 @@ func (c *Client) Run(fd int32, interfaceName string, envList *EnvList) error {
cfg.WgIface = interfaceName
connectClient := internal.NewConnectClient(ctx, cfg, c.recorder,
internal.WithNetEvents(c.netMgr))
internal.WithNetworkState(c.netState), internal.WithSweeper(c.sweeper))
c.setState(cfg, connectClient)
// Persist the latest sync response so DebugBundle can include the network
// map. On iOS this is backed by disk to keep it out of the constrained
@@ -200,11 +203,10 @@ func (c *Client) Run(fd int32, interfaceName string, envList *EnvList) error {
// (e.g. from NWPathMonitor). While unavailable, the internal reconnect loops
// suspend their attempts and the connection listener reports NoNetwork
// instead of Connecting; when availability returns, the loops resume
// immediately with a fresh backoff. Losing the last network also sweeps the
// registered connections, so the client does not keep reporting Connected
// over stale sockets with no network at all.
// immediately with a fresh backoff.
func (c *Client) SetNetworkAvailable(available bool) {
c.netMgr.SetNetworkAvailable(available)
c.netState.Set(available)
c.recorder.SetNetworkAvailable(available)
}
// NotifyNetworkChange marks the management, signal and relay connections
@@ -212,7 +214,8 @@ func (c *Client) SetNetworkAvailable(available bool) {
// whatever has not redialed on the new network by then. The engine and the
// TUN device stay untouched.
func (c *Client) NotifyNetworkChange() {
c.netMgr.NotifyNetworkChange()
c.sweeper.MarkNetworkChange()
log.Infof("network change: connections marked stale")
}
// Stop the internal client and free the resources

View File

@@ -1,158 +0,0 @@
// Package netevents owns the OS network event handling shared by the mobile
// bindings: availability changes park or wake the reconnection loops and drive
// the NoNetwork listener state, and both losing the last network and switching
// networks sweep the stale connections so their owners redial immediately.
package netevents
import (
"context"
"time"
"github.com/cenkalti/backoff/v4"
log "github.com/sirupsen/logrus"
"github.com/netbirdio/netbird/client/netevents/netstate"
"github.com/netbirdio/netbird/client/netevents/sweep"
)
// Recorder receives the availability changes for listener state reporting.
type Recorder interface {
SetNetworkAvailable(available bool)
}
// Manager ties the network availability state, the connection sweeper and the
// status recorder together; it outlives engine restarts. A nil *Manager is
// the valid no-events value: the read methods report always-online and never
// sweep.
type Manager struct {
netState *netstate.State
sweeper *sweep.Sweeper
recorder Recorder
}
// NewManager creates a Manager reporting into recorder, starting online.
func NewManager(recorder Recorder) *Manager {
return &Manager{
netState: netstate.New(),
sweeper: sweep.New(),
recorder: recorder,
}
}
// SetNetworkAvailable records OS-reported network availability. While
// unavailable, the reconnection loops suspend their attempts and the
// connection listener reports NoNetwork instead of Connecting; when
// availability returns, the loops resume immediately with a fresh backoff.
// Losing the last network also sweeps the registered connections: nothing can
// redial while offline, so the stale sockets would otherwise stay silently
// "connected" until their own timeouts and the client would keep reporting
// Connected with no network at all.
func (m *Manager) SetNetworkAvailable(available bool) {
if !available && m.netState.IsOnline() {
m.sweeper.MarkNetworkChange()
}
m.netState.Set(available)
m.recorder.SetNetworkAvailable(available)
}
// NotifyNetworkChange marks the management, signal and relay connections
// stale after the OS switched networks and schedules a sweep that cuts
// whatever has not redialed on the new network by then. The engine and the
// TUN device stay untouched.
func (m *Manager) NotifyNetworkChange() {
m.sweeper.MarkNetworkChange()
log.Infof("network change: connections marked stale")
}
// IsOnline reports whether the OS reports at least one usable network.
func (m *Manager) IsOnline() bool {
if m == nil {
return true
}
return m.netState.IsOnline()
}
// Changed returns a channel closed on the next availability transition.
func (m *Manager) Changed() <-chan struct{} {
if m == nil {
return nil
}
return m.netState.Changed()
}
// Wait blocks while the network is offline; see netstate.State.Wait.
func (m *Manager) Wait(ctx context.Context) (bool, error) {
if m == nil {
return false, nil
}
return m.netState.Wait(ctx)
}
// WaitSettled waits until an online verdict holds for a full settleWindow, or
// while offline until the budget runs out. Returns false when ctx is
// cancelled. The settle window exists because a disconnect often precedes the
// OS offline flag by a few milliseconds, so a fresh online verdict cannot be
// trusted immediately. A nil Manager has no events to watch: it degrades to a
// fixed budget-long sleep.
func (m *Manager) WaitSettled(ctx context.Context, budget, settleWindow time.Duration) bool {
if m == nil {
select {
case <-time.After(budget):
return true
case <-ctx.Done():
return false
}
}
budgetTimer := time.NewTimer(budget)
defer budgetTimer.Stop()
settle := time.NewTimer(settleWindow)
defer settle.Stop()
for {
// Channel first, flag second: a flip in between still fires the channel.
changedCh := m.netState.Changed()
if m.netState.IsOnline() {
select {
case <-settle.C:
return true
case <-changedCh:
case <-ctx.Done():
return false
}
} else {
select {
case <-budgetTimer.C:
return true
case <-changedCh:
case <-ctx.Done():
return false
}
}
if !settle.Stop() {
select {
case <-settle.C:
default:
}
}
settle.Reset(settleWindow)
}
}
// StartDial registers an in-flight dial with the sweeper; see sweep.Sweeper.StartDial.
func (m *Manager) StartDial(ctx context.Context) *sweep.Dial {
if m == nil {
return (*sweep.Sweeper)(nil).StartDial(ctx)
}
return m.sweeper.StartDial(ctx)
}
// QuickRetryBackoff wraps bo for a quick retry after a network change; see
// sweep.Sweeper.QuickRetryBackoff.
func (m *Manager) QuickRetryBackoff(ctx context.Context, bo backoff.BackOff) backoff.BackOff {
if m == nil {
return bo
}
return m.sweeper.QuickRetryBackoff(ctx, bo, m.netState)
}

View File

@@ -1,34 +0,0 @@
package netevents
import (
"context"
"testing"
"time"
"github.com/stretchr/testify/assert"
)
type recorderStub struct{}
func (recorderStub) SetNetworkAvailable(bool) {}
func TestWaitSettledAfterOutage(t *testing.T) {
const budget = 1500 * time.Millisecond
const settleWindow = 200 * time.Millisecond
const outage = 2 * settleWindow
m := NewManager(recorderStub{})
m.SetNetworkAvailable(false)
start := time.Now()
go func() {
time.Sleep(outage)
m.SetNetworkAvailable(true)
}()
ok := m.WaitSettled(context.Background(), budget, settleWindow)
elapsed := time.Since(start)
assert.True(t, ok, "recovered network must let the caller proceed")
assert.GreaterOrEqual(t, elapsed, outage+settleWindow, "an online verdict must hold a full settle window before it is trusted")
}

View File

@@ -1,10 +1,10 @@
// Package sweep cuts network-bound activity when the OS switches networks:
// Package netsweep cuts network-bound activity when the OS switches networks:
// a sweep closes the registered connections and aborts the in-flight dials, so
// their owners redial immediately instead of waiting for the old sockets to
// time out.
//
// A nil *Sweeper disables everything: all methods are nil-safe no-ops.
package sweep
package netsweep
import (
"context"
@@ -16,7 +16,7 @@ import (
"github.com/cenkalti/backoff/v4"
log "github.com/sirupsen/logrus"
"github.com/netbirdio/netbird/client/netevents/netstate"
"github.com/netbirdio/netbird/client/netstate"
)
// DefaultSweepDelay absorbs network flapping while the OS settles on a
@@ -34,7 +34,7 @@ type Config struct {
// ErrSwept reports that a dial finished after a network change swept its
// registration. The connection is already closed; the caller must treat it
// as a failed dial and redial on the new network.
var ErrSwept = errors.New("sweep: connection swept by network change")
var ErrSwept = errors.New("netsweep: connection swept by network change")
// sweepID identifies one registration in a sweeper. Connections and dials
// draw from the same counter, so an id is unique across both registries.

View File

@@ -1,4 +1,4 @@
package sweep
package netsweep
import (
"context"

View File

@@ -1,11 +1,11 @@
package sweep
package netsweep
import (
"time"
"github.com/cenkalti/backoff/v4"
"github.com/netbirdio/netbird/client/netevents/netstate"
"github.com/netbirdio/netbird/client/netstate"
)
const quickRetryDelay = 200 * time.Millisecond

View File

@@ -1,4 +1,4 @@
package sweep
package netsweep
import (
"context"

View File

@@ -21,7 +21,8 @@ import (
"google.golang.org/grpc/connectivity"
nbgrpc "github.com/netbirdio/netbird/client/grpc"
"github.com/netbirdio/netbird/client/netevents"
"github.com/netbirdio/netbird/client/netstate"
"github.com/netbirdio/netbird/client/netsweep"
"github.com/netbirdio/netbird/client/system"
"github.com/netbirdio/netbird/encryption"
"github.com/netbirdio/netbird/shared/management/domain"
@@ -63,9 +64,12 @@ type GrpcClient struct {
connStateCallbackLock sync.RWMutex
serverURL string
// netEvents gates the stream retry loop on OS-reported network
// availability and sweeps the transport on network change.
netEvents *netevents.Manager
// netState gates the stream retry loop on OS-reported network
// availability; nil (the default) disables gating.
netState *netstate.State
// sweeper cuts the transport connections on network change; nil disables it.
sweeper *netsweep.Sweeper
// syncStreamErr holds the last Sync stream error, or nil while the stream
// is established and healthy. GetServerKey succeeds even when the peer
@@ -119,9 +123,15 @@ func MaxRecvMsgSize() int {
// Option configures optional GrpcClient behavior.
type Option func(*GrpcClient)
// WithNetEvents injects the OS network event handling.
func WithNetEvents(events *netevents.Manager) Option {
return func(c *GrpcClient) { c.netEvents = events }
// WithNetworkState injects the OS network availability state that gates the
// stream retry loop; without it gating is disabled.
func WithNetworkState(netState *netstate.State) Option {
return func(c *GrpcClient) { c.netState = netState }
}
// WithSweeper injects the network change sweeper.
func WithSweeper(sweeper *netsweep.Sweeper) Option {
return func(c *GrpcClient) { c.sweeper = sweeper }
}
// NewClient creates a new client to Management service
@@ -142,7 +152,9 @@ func NewClient(ctx context.Context, addr string, ourPrivateKey wgtypes.Key, tlsE
extraOpts = append(extraOpts, grpc.WithDefaultCallOptions(grpc.MaxCallRecvMsgSize(maxSize)))
log.Infof("management gRPC max receive message size set to %d bytes", maxSize)
}
extraOpts = append(extraOpts, nbgrpc.WithSweeper(c.netEvents))
if c.sweeper != nil {
extraOpts = append(extraOpts, nbgrpc.WithSweeper(c.sweeper))
}
var conn *grpc.ClientConn
operation := func() error {
@@ -223,19 +235,16 @@ func (c *GrpcClient) withMgmtStream(
ctx context.Context,
handler func(ctx context.Context, serverPubKey wgtypes.Key, backOff backoff.BackOff) error,
) error {
backOff := c.netEvents.QuickRetryBackoff(ctx, defaultBackoff(ctx))
backOff := c.sweeper.QuickRetryBackoff(ctx, defaultBackoff(ctx), c.netState)
operation := func() error {
// suspend reconnect attempts while the OS reports no usable network.
// Wait only errors on a cancelled context, which means shutdown, so
// stop the loop without reporting a failure.
if waited, err := c.netEvents.Wait(ctx); err != nil {
if waited, err := c.netState.Wait(ctx); err != nil {
log.Debugf("management connection context has been canceled while offline, this usually indicates shutdown")
return nil //nolint:nilerr // a cancelled context means shutdown, not a retryable failure
} else if waited {
backOff.Reset()
// dials attempted while offline grew the channel's internal backoff;
// reset it too, or the reconnect waits out that timer first
c.conn.ResetConnectBackoff()
}
connState := c.conn.GetState()
@@ -264,7 +273,7 @@ func (c *GrpcClient) withMgmtStream(
return handler(ctx, *serverPubKey, backOff)
}
err := nbgrpc.Retry(ctx, operation, backOff, c.netEvents)
err := nbgrpc.Retry(ctx, operation, backOff, c.netState)
if err != nil {
log.Warnf("exiting the Management service connection retry loop due to the unrecoverable error: %s", err)
}

View File

@@ -14,7 +14,7 @@ import (
log "github.com/sirupsen/logrus"
"github.com/netbirdio/netbird/client/netevents/sweep"
"github.com/netbirdio/netbird/client/netsweep"
auth "github.com/netbirdio/netbird/shared/relay/auth/hmac"
"github.com/netbirdio/netbird/shared/relay/client/dialer"
netErr "github.com/netbirdio/netbird/shared/relay/client/dialer/net"
@@ -151,14 +151,6 @@ type transportConn interface {
Protocol() string
}
// NetEvents is the OS network event view the relay consumes: availability
// gating for the reconnect guard and dial registration for the network change
// sweep.
type NetEvents interface {
NetworkWatcher
StartDial(ctx context.Context) *sweep.Dial
}
// Client is a client for the relay server. It is responsible for establishing a connection to the relay server and
// managing connections to other peers. All exported functions are safe to call concurrently. After close the connection,
// the client can be reused by calling Connect again. When the client is closed, all connections are closed too.
@@ -194,10 +186,9 @@ type Client struct {
// the manager.
transportFallback *transportFallback
// netEvents registers the relay dial for the network change sweep; the
// read loop reports the disconnect and the guard reconnects. Shared via
// the manager.
netEvents NetEvents
// sweeper cuts the relay connection on network change; the read loop
// reports the disconnect and the guard reconnects. Shared via the manager.
sweeper *netsweep.Sweeper
// datagramFallbackTriggered guards a single fallback per connection so a
// burst of oversized datagrams triggers one reconnect, not many.
datagramFallbackTriggered atomic.Bool
@@ -409,12 +400,7 @@ func (c *Client) Close() error {
func (c *Client) connect(ctx context.Context) (*RelayAddr, error) {
// A sweep cancels this context, so a dial started on the old network
// aborts instead of waiting out its handshake timeout.
var dial *sweep.Dial
if c.netEvents != nil {
dial = c.netEvents.StartDial(ctx)
} else {
dial = (*sweep.Sweeper)(nil).StartDial(ctx)
}
dial := c.sweeper.StartDial(ctx)
defer dial.Release()
ctx = dial.Ctx()

View File

@@ -7,6 +7,8 @@ import (
"github.com/cenkalti/backoff/v4"
log "github.com/sirupsen/logrus"
"github.com/netbirdio/netbird/client/netstate"
)
const (
@@ -22,13 +24,6 @@ const (
verdictSettleWindow = 200 * time.Millisecond
)
// NetworkWatcher is the availability view the guard gates reconnects on.
type NetworkWatcher interface {
Wait(ctx context.Context) (bool, error)
IsOnline() bool
WaitSettled(ctx context.Context, budget, settleWindow time.Duration) bool
}
// Guard manage the reconnection tries to the Relay server in case of disconnection event.
type Guard struct {
// OnNewRelayClient is a channel that is used to notify the relay manager about a new relay client instance.
@@ -40,8 +35,9 @@ type Guard struct {
// attempts.
maxBackoffInterval time.Duration
// netState gates reconnect attempts on OS-reported network availability.
netState NetworkWatcher
// netState gates reconnect attempts on OS-reported network availability;
// nil disables gating.
netState *netstate.State
// lastErr is the error from the most recent failed reconnect attempt,
// surfaced as the home relay status while disconnected.
@@ -49,8 +45,9 @@ type Guard struct {
}
// NewGuard creates a new guard for the relay client. A non-positive
// maxBackoffInterval falls back to defaultMaxBackoffInterval.
func NewGuard(sp *ServerPicker, maxBackoffInterval time.Duration, netState NetworkWatcher) *Guard {
// maxBackoffInterval falls back to defaultMaxBackoffInterval. A nil netState
// disables network availability gating.
func NewGuard(sp *ServerPicker, maxBackoffInterval time.Duration, netState *netstate.State) *Guard {
if maxBackoffInterval <= 0 {
maxBackoffInterval = defaultMaxBackoffInterval
}
@@ -100,14 +97,12 @@ func (g *Guard) StartReconnectTrys(ctx context.Context, relayClient *Client) {
select {
case <-ticker.C:
// suspend reconnect attempts while the OS reports no usable network
if g.netState != nil {
if waited, err := g.netState.Wait(ctx); err != nil {
return
} else if waited {
ticker.Stop()
ticker = g.exponentTicker(ctx)
continue
}
if waited, err := g.netState.Wait(ctx); err != nil {
return
} else if waited {
ticker.Stop()
ticker = g.exponentTicker(ctx)
continue
}
if err := g.retry(ctx); err != nil {
log.Errorf("failed to pick new Relay server: %s", err)
@@ -134,18 +129,13 @@ func (g *Guard) tryToQuickReconnect(parentCtx context.Context, rc *Client) bool
return false
}
if g.netState != nil {
if ok := g.netState.WaitSettled(parentCtx, quickReconnectBudget, verdictSettleWindow); !ok {
return false
}
// Still offline after the budget: leave the retry to the ticker.
if !g.netState.IsOnline() {
return false
}
} else {
if cancelled := waiteBeforeRetry(parentCtx); !cancelled {
return false
}
if ok := g.waitForNetwork(parentCtx); !ok {
return false
}
// Still offline after the budget: leave the retry to the ticker.
if !g.netState.IsOnline() {
return false
}
log.Infof("try to reconnect to Relay server: %s", rc.connectionURL)
@@ -210,14 +200,47 @@ func (g *Guard) exponentTicker(ctx context.Context) *backoff.Ticker {
return backoff.NewTicker(bo)
}
func waiteBeforeRetry(ctx context.Context) bool {
timer := time.NewTimer(quickReconnectBudget)
defer timer.Stop()
// waitForNetwork waits out the settle window while online, or waits for the
// network to return while offline, within the budget. Returns false when ctx
// is cancelled. Without an injected netState it degrades to a fixed
// budget-long sleep, the pre-netstate behavior.
func (g *Guard) waitForNetwork(ctx context.Context) bool {
budget := time.NewTimer(quickReconnectBudget)
defer budget.Stop()
select {
case <-timer.C:
return true
case <-ctx.Done():
return false
settleWindow := verdictSettleWindow
if g.netState == nil {
settleWindow = quickReconnectBudget
}
settle := time.NewTimer(settleWindow)
defer settle.Stop()
for {
// Channel first, flag second: a flip in between still fires the channel.
changedCh := g.netState.Changed()
if g.netState.IsOnline() {
select {
case <-settle.C:
return true
case <-changedCh:
case <-ctx.Done():
return false
}
} else {
select {
case <-budget.C:
return true
case <-changedCh:
case <-ctx.Done():
return false
}
}
if !settle.Stop() {
select {
case <-settle.C:
default:
}
}
settle.Reset(settleWindow)
}
}

View File

@@ -0,0 +1,30 @@
package client
import (
"context"
"testing"
"time"
"github.com/stretchr/testify/assert"
"github.com/netbirdio/netbird/client/netstate"
)
func TestWaitForNetworkSettlesAfterOutage(t *testing.T) {
ns := netstate.New()
ns.Set(false)
g := NewGuard(nil, 0, ns)
const outage = 2 * verdictSettleWindow
start := time.Now()
go func() {
time.Sleep(outage)
ns.Set(true)
}()
ok := g.waitForNetwork(context.Background())
elapsed := time.Since(start)
assert.True(t, ok, "recovered network must let the quick reconnect proceed")
assert.GreaterOrEqual(t, elapsed, outage+verdictSettleWindow, "reconnect must wait a full settle window after the network returns")
}

View File

@@ -12,6 +12,8 @@ import (
log "github.com/sirupsen/logrus"
"github.com/netbirdio/netbird/client/netstate"
"github.com/netbirdio/netbird/client/netsweep"
relayAuth "github.com/netbirdio/netbird/shared/relay/auth/hmac"
)
@@ -65,9 +67,15 @@ func WithMaxBackoffInterval(d time.Duration) ManagerOption {
return func(m *Manager) { m.maxBackoffInterval = d }
}
// WithNetEvents injects the OS network event handling.
func WithNetEvents(events NetEvents) ManagerOption {
return func(m *Manager) { m.netEvents = events }
// WithNetworkState injects the OS network availability state that gates the
// reconnect guard; without it reconnect attempts are not gated.
func WithNetworkState(netState *netstate.State) ManagerOption {
return func(m *Manager) { m.netState = netState }
}
// WithSweeper injects the network change sweeper.
func WithSweeper(sweeper *netsweep.Sweeper) ManagerOption {
return func(m *Manager) { m.sweeper = sweeper }
}
// Manager is a manager for the relay client instances. It establishes one persistent connection to the given relay URL
@@ -97,7 +105,8 @@ type Manager struct {
mtu uint16
maxBackoffInterval time.Duration
netEvents NetEvents
netState *netstate.State
sweeper *netsweep.Sweeper
cleanupInterval time.Duration
keepUnusedServerTime time.Duration
@@ -134,9 +143,9 @@ func NewManager(ctx context.Context, serverURLs []string, peerID string, mtu uin
for _, opt := range opts {
opt(m)
}
m.serverPicker.NetEvents = m.netEvents
m.serverPicker.Sweeper = m.sweeper
m.serverPicker.ServerURLs.Store(serverURLs)
m.reconnectGuard = NewGuard(m.serverPicker, m.maxBackoffInterval, m.netEvents)
m.reconnectGuard = NewGuard(m.serverPicker, m.maxBackoffInterval, m.netState)
return m
}
@@ -361,7 +370,7 @@ func (m *Manager) openConnVia(ctx context.Context, serverAddress, peerKey string
relayClient := NewClientWithServerIP(serverAddress, serverIP, m.tokenStore, m.peerID, m.mtu)
relayClient.SetTransportFallback(m.transportFallback)
relayClient.netEvents = m.netEvents
relayClient.sweeper = m.sweeper
err := relayClient.Connect(m.ctx)
if err != nil {
rt.Lock()

View File

@@ -9,6 +9,7 @@ import (
log "github.com/sirupsen/logrus"
"github.com/netbirdio/netbird/client/netsweep"
auth "github.com/netbirdio/netbird/shared/relay/auth/hmac"
)
@@ -30,7 +31,7 @@ type ServerPicker struct {
MTU uint16
ConnectionTimeout time.Duration
TransportFallback *transportFallback
NetEvents NetEvents
Sweeper *netsweep.Sweeper
}
func (sp *ServerPicker) PickServer(parentCtx context.Context) (*Client, error) {
@@ -74,7 +75,7 @@ func (sp *ServerPicker) startConnection(ctx context.Context, resultChan chan con
log.Infof("try to connecting to relay server: %s", url)
relayClient := NewClient(url, sp.TokenStore, sp.PeerID, sp.MTU)
relayClient.SetTransportFallback(sp.TransportFallback)
relayClient.netEvents = sp.NetEvents
relayClient.sweeper = sp.Sweeper
err := relayClient.Connect(ctx)
resultChan <- connResult{
RelayClient: relayClient,

View File

@@ -19,7 +19,8 @@ import (
"google.golang.org/grpc/status"
nbgrpc "github.com/netbirdio/netbird/client/grpc"
"github.com/netbirdio/netbird/client/netevents"
"github.com/netbirdio/netbird/client/netstate"
"github.com/netbirdio/netbird/client/netsweep"
"github.com/netbirdio/netbird/encryption"
"github.com/netbirdio/netbird/shared/management/client"
"github.com/netbirdio/netbird/shared/signal/proto"
@@ -66,9 +67,12 @@ type GrpcClient struct {
connStateCallback ConnStateNotifier
connStateCallbackLock sync.RWMutex
// netEvents gates the Receive retry loop on OS-reported network
// availability and sweeps the transport on network change.
netEvents *netevents.Manager
// netState gates the Receive retry loop on OS-reported network
// availability; nil (the default) disables gating.
netState *netstate.State
// sweeper cuts the transport connections on network change; nil disables it.
sweeper *netsweep.Sweeper
onReconnectedListenerFn func()
@@ -96,9 +100,15 @@ type GrpcClient struct {
// Option configures optional GrpcClient behavior.
type Option func(*GrpcClient)
// WithNetEvents injects the OS network event handling.
func WithNetEvents(events *netevents.Manager) Option {
return func(c *GrpcClient) { c.netEvents = events }
// WithNetworkState injects the OS network availability state that gates the
// Receive retry loop; without it gating is disabled.
func WithNetworkState(netState *netstate.State) Option {
return func(c *GrpcClient) { c.netState = netState }
}
// WithSweeper injects the network change sweeper.
func WithSweeper(sweeper *netsweep.Sweeper) Option {
return func(c *GrpcClient) { c.sweeper = sweeper }
}
// NewClient creates a new Signal client
@@ -116,7 +126,9 @@ func NewClient(ctx context.Context, addr string, key wgtypes.Key, tlsEnabled boo
}
var extraOpts []grpc.DialOption
extraOpts = append(extraOpts, nbgrpc.WithSweeper(c.netEvents))
if c.sweeper != nil {
extraOpts = append(extraOpts, nbgrpc.WithSweeper(c.sweeper))
}
var conn *grpc.ClientConn
operation := func() error {
@@ -186,20 +198,17 @@ func defaultBackoff(ctx context.Context) backoff.BackOff {
// The connection retry logic will try to reconnect for 30 min and if wasn't successful will propagate the error to the function caller.
func (c *GrpcClient) Receive(ctx context.Context, msgHandler func(msg *proto.Message) error) error {
backOff := c.netEvents.QuickRetryBackoff(ctx, defaultBackoff(ctx))
backOff := c.sweeper.QuickRetryBackoff(ctx, defaultBackoff(ctx), c.netState)
operation := func() error {
// suspend reconnect attempts while the OS reports no usable network.
// Wait only errors on a cancelled context, which means shutdown, so
// stop the loop without reporting a failure.
if waited, err := c.netEvents.Wait(ctx); err != nil {
if waited, err := c.netState.Wait(ctx); err != nil {
log.Debugf("signal connection context has been canceled while offline, this usually indicates shutdown")
return nil
} else if waited {
backOff.Reset()
// dials attempted while offline grew the channel's internal backoff;
// reset it too, or the reconnect waits out that timer first
c.signalConn.ResetConnectBackoff()
}
c.notifyStreamDisconnected()
@@ -272,7 +281,7 @@ func (c *GrpcClient) Receive(ctx context.Context, msgHandler func(msg *proto.Mes
return nil
}
err := nbgrpc.Retry(ctx, operation, backOff, c.netEvents)
err := nbgrpc.Retry(ctx, operation, backOff, c.netState)
if err != nil {
log.Errorf("exiting the Signal service connection retry loop due to the unrecoverable error: %v", err)
return err