Files
netbird/client/internal/conn_mgr.go
Riccardo Manfrin 46568f7af8 [client] Reconcile routed allowed IPs when a lazy connection goes idle (#6863)
## Describe your changes

Under lazy connections, when a routing peer goes idle its WireGuard peer
is torn down and re-created with a wake endpoint by
the activity listener, carrying only the overlay /32
(`peerCfg.AllowedIPs`). The routed subnet prefixes are dropped from
  the device on the Connected→Idle transition.

They are meant to be restored by the route watcher, which reacts to the
peer's status change and calls `recalculateRoutes` →
  `AddAllowedIP`. Two things prevent that from healing the peer:

- `AddAllowedIP` uses `update_only`, which is a silent no-op (no error)
when the peer does not exist. While the peer is
being torn down and re-armed with its wake endpoint, it is briefly
absent, so a re-add that lands in that window is lost.
- The allowed-IP refcounter only calls its add function on a prefix's
0→1 transition. The routed prefix stays referenced
across the idle cycle, so once the device entry is gone the refcounter
does not re-push it on its own, and nothing retries.

As a result, traffic to the routed subnet is black-holed while the peer
is idle. Because the wake endpoint only fires when a
packet matches the peer's AllowedIPs, a packet to the subnet is dropped
before reaching the wake endpoint, so it cannot
wake the peer. The peer only recovers when woken by other means (e.g. a
ping to its overlay IP).

  ## Approach

This change keeps the existing Connected→Idle transition as-is and
reconciles the AllowedIPs afterwards, avoiding any
additional locking on the transition path. The peer is torn down and
re-armed with its wake endpoint as today; the routed
prefixes are then re-applied from the route manager's allowed-IP
refcounter once the wake endpoint has been (re)armed.

A single add-only method, `ReconcilePeerAllowedIPs(peerKey)`, re-applies
every routed prefix currently tracked for the peer
in the refcounter (the authoritative store; it already covers static,
dynamic and dnsinterceptor routes). It runs whenever
the peer's wake endpoint is (re)created in the lazy manager — every
point where the activity listener builds it with the
  overlay /32 only:

- **initial registration** (`AddPeer`, cold start): the route manager
may have already pushed the peer's routes before the
wake endpoint existed, so those `AddAllowedIP` calls no-op'd; the
reconcile installs them on the freshly created wake
  endpoint.
- **the two paths into idle** (`DeactivatePeer` on a remote GOAWAY,
`onPeerInactivityTimedOut` on local inactivity): the
peer is torn down and re-armed, so the routed prefixes must be
re-applied.

In every case the routed prefixes end up on the wake endpoint, so
traffic to a routed subnet can wake the peer. Arming the
wake endpoint and reconciling are wrapped in a single
`armActivityListener` helper so the two always happen together.

New helper: `refcounter.Counter.KeysMatching(pred)` to enumerate a
peer's prefixes under the counter lock.

Note on scope: the reconcile restores what the refcounter tracks. All
routed AllowedIPs currently go through it, so this
covers the routed-prefix case; it does not attempt to reconcile
AllowedIPs installed outside the refcounter. The
Idle→Connected (wake) path does not need this: the peer is not removed
there (the listener close leaves it in place and only
the endpoint is updated), so a concurrent `AddAllowedIP` lands normally.

  ## Testing

Reproduced deterministically in a local dev setup (userspace client,
`NB_WG_KERNEL_DISABLED=true`, `B_LAZY_CONN_INACTIVITY_THRESHOLD=1`
inactivity threshold 1
min). A temporary 30s sleep in the tear-down → re-arm window widens the
race so the route watcher's async `AddAllowedIP`
reliably lands while the peer is absent and no-ops (the sleep is a test
aid, not part of the change):

- **without the reconcile:** after the peer goes idle, a ping to any
routed IP — both a pre-existing route and one added
  during the window — black-holes; the peer never wakes.
  - **with the reconcile:** the same ping wakes the peer and passes.

Added unit tests: `ReconcilePeerAllowedIPs` (re-applies all of a peer's
tracked prefixes, scoped to that peer) and
  `refcounter.Counter.KeysMatching`.

Note: `netbird status -d` is not a reliable signal for this —
`AddPeerStateRoute` records the route regardless of whether
the underlying `AddAllowedIP` no-op'd, so it reflects the route
manager's intent rather than device state. The reliable
  signal is functional (ping the subnet from idle).
  
  ## Checklist

  - [x] Is it a bug fix
  - [ ] Is a typo/documentation fix
  - [ ] Is a feature enhancement
  - [ ] It is a refactor
- [x] Created tests that fail without the change (unit tests for the
reconcile + `KeysMatching`)

  ## Documentation

- [x] Documentation is **not needed** for this change (internal client
behavior, no API / gRPC / CLI / flag change)

<!-- codesmith:footer -->
---
<a
href="https://app.blacksmith.sh/netbirdio/codesmith/netbird/pr/6863"><picture><source
media="(prefers-color-scheme: dark)"
srcset="https://pr-comments-assets.blacksmith.sh/codesmith/view-with-codesmith-dark-v2.svg"><source
media="(prefers-color-scheme: light)"
srcset="https://pr-comments-assets.blacksmith.sh/codesmith/view-with-codesmith-light-v2.svg"><img
alt="View with Codesmith"
src="https://pr-comments-assets.blacksmith.sh/codesmith/view-with-codesmith-dark-v2.svg"></picture></a>
<a
href="https://backend.blacksmith.sh/track/enable-autofix?expires=1787330960&installation_model_id=427504&pr_number=6863&repository=netbirdio%2Fnetbird&return_to=https%3A%2F%2Fgithub.com%2Fnetbirdio%2Fnetbird%2Fpull%2F6863&signature=f3d6a97d7db82e92b3939fdd0f159c5ee74913ff88f4eb82e41e88fcb787aff4"><picture><source
media="(prefers-color-scheme: dark)"
srcset="https://pr-comments-assets.blacksmith.sh/codesmith/autofix-with-codesmith-dark.svg"><source
media="(prefers-color-scheme: light)"
srcset="https://pr-comments-assets.blacksmith.sh/codesmith/autofix-with-codesmith-light.svg"><img
alt="Autofix with Codesmith"
src="https://pr-comments-assets.blacksmith.sh/codesmith/autofix-with-codesmith-dark.svg"></picture></a>
<sup>Need help on this PR? Tag <code>/codesmith</code> with what you
need. Autofix is disabled.</sup>

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

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

* **Bug Fixes**
* Routed IP assignments are automatically reconciled and restored
whenever a peer’s lazy wake endpoint is armed or re-armed.
* Routed allowed IPs are re-applied after inactivity transitions and
monitoring re-initialization.
* If reconciliation can’t be performed, the client safely skips it; if
reconciliation encounters issues, failures are logged without stopping
connection monitoring.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
2026-07-23 18:40:54 +02:00

396 lines
12 KiB
Go

package internal
import (
"context"
"os"
"strconv"
"sync"
"time"
log "github.com/sirupsen/logrus"
"github.com/netbirdio/netbird/client/internal/lazyconn"
"github.com/netbirdio/netbird/client/internal/lazyconn/manager"
"github.com/netbirdio/netbird/client/internal/peer"
"github.com/netbirdio/netbird/client/internal/peerstore"
"github.com/netbirdio/netbird/route"
)
// lazyForce is the resolved local decision for lazy connections, layered above the
// management feature flag. lazyForceNone defers to management.
type lazyForce int
const (
lazyForceNone lazyForce = iota
lazyForceOn
lazyForceOff
)
// ConnMgr coordinates both lazy connections (established on-demand) and permanent peer connections.
//
// The connection manager is responsible for:
// - Managing lazy connections via the lazyConnManager
// - Maintaining a list of excluded peers that should always have permanent connections
// - Handling connection establishment based on peer signaling
//
// The implementation is not thread-safe; it is protected by engine.syncMsgMux.
// The only exception is ActivatePeer, which is safe for concurrent use so the
// DNS warm-up path can call it without contending on the engine mutex.
type ConnMgr struct {
peerStore *peerstore.Store
statusRecorder *peer.Status
iface lazyconn.WGIface
force lazyForce
rosenpassEnabled bool
lazyConnMgr *manager.Manager
// lazyConnMgrMu guards the lazyConnMgr pointer for readers outside the
// engine loop (ActivatePeer). Writers hold it in addition to
// engine.syncMsgMux; all other reads stay under engine.syncMsgMux only.
lazyConnMgrMu sync.RWMutex
// reconcileRoutedIPs re-applies a peer's routed allowed IPs after its lazy wake endpoint is
// (re)armed (Mode A at arm time). Injected by the engine; nil disables the reconcile.
reconcileRoutedIPs func(peerKey string) error
wg sync.WaitGroup
lazyCtx context.Context
lazyCtxCancel context.CancelFunc
}
// SetRoutedIPsReconciler injects the callback used to re-apply a peer's routed allowed IPs when
// its lazy wake endpoint is (re)armed. Must be called before the lazy manager starts.
func (e *ConnMgr) SetRoutedIPsReconciler(fn func(peerKey string) error) {
e.reconcileRoutedIPs = fn
}
func NewConnMgr(engineConfig *EngineConfig, statusRecorder *peer.Status, peerStore *peerstore.Store, iface lazyconn.WGIface) *ConnMgr {
e := &ConnMgr{
peerStore: peerStore,
statusRecorder: statusRecorder,
iface: iface,
force: resolveLazyForce(engineConfig.LazyConnection),
rosenpassEnabled: engineConfig.RosenpassEnabled,
}
return e
}
// Start initializes the connection manager. It starts the lazy connection manager when a
// local override forces it on; with no local override it waits for the management feature flag.
func (e *ConnMgr) Start(ctx context.Context) {
if e.lazyConnMgr != nil {
log.Errorf("lazy connection manager is already started")
return
}
switch e.force {
case lazyForceOff:
log.Infof("lazy connection manager is disabled by local override (%s or MDM policy)", lazyconn.EnvLazyConn)
e.statusRecorder.UpdateLazyConnection(false)
return
case lazyForceNone:
log.Infof("lazy connection manager is managed by the management feature flag")
e.statusRecorder.UpdateLazyConnection(false)
return
}
if e.rosenpassEnabled {
log.Warnf("rosenpass connection manager is enabled, lazy connection manager will not be started")
e.statusRecorder.UpdateLazyConnection(false)
return
}
e.initLazyManager(ctx)
e.statusRecorder.UpdateLazyConnection(true)
}
// UpdatedRemoteFeatureFlag is called when the remote feature flag is updated.
// If enabled, it initializes the lazy connection manager and start it. Do not need to call Start() again.
// If disabled, then it closes the lazy connection manager and open the connections to all peers.
func (e *ConnMgr) UpdatedRemoteFeatureFlag(ctx context.Context, enabled bool) error {
// a local override (NB_LAZY_CONN or local config) takes precedence over management
if e.force != lazyForceNone {
return nil
}
if enabled {
// if the lazy connection manager is already started, do not start it again
if e.lazyConnMgr != nil {
return nil
}
if e.rosenpassEnabled {
log.Infof("rosenpass connection manager is enabled, lazy connection manager will not be started")
e.statusRecorder.UpdateLazyConnection(false)
return nil
}
log.Infof("lazy connection manager is enabled by the management feature flag")
e.initLazyManager(ctx)
e.statusRecorder.UpdateLazyConnection(true)
return e.addPeersToLazyConnManager()
} else {
if e.lazyConnMgr == nil {
e.statusRecorder.UpdateLazyConnection(false)
return nil
}
log.Infof("lazy connection manager is disabled by management feature flag")
e.closeManager(ctx)
e.statusRecorder.UpdateLazyConnection(false)
return nil
}
}
// UpdateRouteHAMap updates the route HA mappings in the lazy connection manager
func (e *ConnMgr) UpdateRouteHAMap(haMap route.HAMap) {
if !e.isStartedWithLazyMgr() {
log.Debugf("lazy connection manager is not started, skipping UpdateRouteHAMap")
return
}
e.lazyConnMgr.UpdateRouteHAMap(haMap)
}
// SetExcludeList sets the list of peer IDs that should always have permanent connections.
func (e *ConnMgr) SetExcludeList(ctx context.Context, peerIDs map[string]bool) {
if e.lazyConnMgr == nil {
return
}
excludedPeers := make([]lazyconn.PeerConfig, 0, len(peerIDs))
for peerID := range peerIDs {
var peerConn *peer.Conn
var exists bool
if peerConn, exists = e.peerStore.PeerConn(peerID); !exists {
log.Warnf("failed to find peer conn for peerID: %s", peerID)
continue
}
lazyPeerCfg := lazyconn.PeerConfig{
PublicKey: peerID,
AllowedIPs: peerConn.WgConfig().AllowedIps,
PeerConnID: peerConn.ConnID(),
Log: peerConn.Log,
}
excludedPeers = append(excludedPeers, lazyPeerCfg)
}
added := e.lazyConnMgr.ExcludePeer(excludedPeers)
for _, peerID := range added {
var peerConn *peer.Conn
var exists bool
if peerConn, exists = e.peerStore.PeerConn(peerID); !exists {
// if the peer not exist in the store, it means that the engine will call the AddPeerConn in next step
continue
}
peerConn.Log.Infof("peer has been added to lazy connection exclude list, opening permanent connection")
if err := peerConn.Open(ctx); err != nil {
peerConn.Log.Errorf("failed to open connection: %v", err)
}
}
}
func (e *ConnMgr) AddPeerConn(ctx context.Context, peerKey string, conn *peer.Conn) (exists bool) {
if success := e.peerStore.AddPeerConn(peerKey, conn); !success {
return true
}
if !e.isStartedWithLazyMgr() {
if err := conn.Open(ctx); err != nil {
conn.Log.Errorf("failed to open connection: %v", err)
}
return
}
if !lazyconn.IsSupported(conn.AgentVersionString()) {
conn.Log.Warnf("peer does not support lazy connection (%s), open permanent connection", conn.AgentVersionString())
if err := conn.Open(ctx); err != nil {
conn.Log.Errorf("failed to open connection: %v", err)
}
return
}
lazyPeerCfg := lazyconn.PeerConfig{
PublicKey: peerKey,
AllowedIPs: conn.WgConfig().AllowedIps,
PeerConnID: conn.ConnID(),
Log: conn.Log,
}
excluded, err := e.lazyConnMgr.AddPeer(lazyPeerCfg)
if err != nil {
conn.Log.Errorf("failed to add peer to lazyconn manager: %v", err)
if err := conn.Open(ctx); err != nil {
conn.Log.Errorf("failed to open connection: %v", err)
}
return
}
if excluded {
conn.Log.Infof("peer is on lazy conn manager exclude list, opening connection")
if err := conn.Open(ctx); err != nil {
conn.Log.Errorf("failed to open connection: %v", err)
}
return
}
conn.Log.Infof("peer added to lazy conn manager")
return
}
func (e *ConnMgr) RemovePeerConn(peerKey string) {
conn, ok := e.peerStore.Remove(peerKey)
if !ok {
return
}
defer conn.Close(false)
if !e.isStartedWithLazyMgr() {
return
}
e.lazyConnMgr.RemovePeer(peerKey)
conn.Log.Infof("removed peer from lazy conn manager")
}
// ActivatePeer wakes an idle lazy connection. Unlike the rest of ConnMgr it is
// safe for concurrent use: the lazy manager pointer is read under lazyConnMgrMu
// and the manager itself is internally synchronized, so callers outside the
// engine loop (DNS warm-up) do not need engine.syncMsgMux.
func (e *ConnMgr) ActivatePeer(ctx context.Context, conn *peer.Conn) {
e.lazyConnMgrMu.RLock()
lazyConnMgr := e.lazyConnMgr
started := lazyConnMgr != nil && e.lazyCtxCancel != nil
e.lazyConnMgrMu.RUnlock()
if !started {
return
}
if found := lazyConnMgr.ActivatePeer(conn.GetKey()); found {
if err := conn.Open(ctx); err != nil {
conn.Log.Errorf("failed to open connection: %v", err)
}
}
}
// DeactivatePeer deactivates a peer connection in the lazy connection manager.
// If locally the lazy connection is disabled, we force the peer connection open.
func (e *ConnMgr) DeactivatePeer(conn *peer.Conn) {
if !e.isStartedWithLazyMgr() {
return
}
conn.Log.Infof("closing peer connection: remote peer initiated inactive, idle lazy state and sent GOAWAY")
e.lazyConnMgr.DeactivatePeer(conn.ConnID())
}
func (e *ConnMgr) Close() {
if !e.isStartedWithLazyMgr() {
return
}
e.lazyCtxCancel()
e.wg.Wait()
e.lazyConnMgrMu.Lock()
e.lazyConnMgr = nil
e.lazyConnMgrMu.Unlock()
}
func (e *ConnMgr) initLazyManager(engineCtx context.Context) {
cfg := manager.Config{
InactivityThreshold: inactivityThresholdEnv(),
ReconcileAllowedIPs: e.reconcileRoutedIPs,
}
e.lazyConnMgrMu.Lock()
e.lazyConnMgr = manager.NewManager(cfg, engineCtx, e.peerStore, e.iface)
e.lazyCtx, e.lazyCtxCancel = context.WithCancel(engineCtx)
e.lazyConnMgrMu.Unlock()
e.wg.Add(1)
go func() {
defer e.wg.Done()
e.lazyConnMgr.Start(e.lazyCtx)
}()
}
func (e *ConnMgr) addPeersToLazyConnManager() error {
peers := e.peerStore.PeersPubKey()
lazyPeerCfgs := make([]lazyconn.PeerConfig, 0, len(peers))
for _, peerID := range peers {
var peerConn *peer.Conn
var exists bool
if peerConn, exists = e.peerStore.PeerConn(peerID); !exists {
log.Warnf("failed to find peer conn for peerID: %s", peerID)
continue
}
lazyPeerCfg := lazyconn.PeerConfig{
PublicKey: peerID,
AllowedIPs: peerConn.WgConfig().AllowedIps,
PeerConnID: peerConn.ConnID(),
Log: peerConn.Log,
}
lazyPeerCfgs = append(lazyPeerCfgs, lazyPeerCfg)
}
return e.lazyConnMgr.AddActivePeers(lazyPeerCfgs)
}
func (e *ConnMgr) closeManager(ctx context.Context) {
if e.lazyConnMgr == nil {
return
}
e.lazyCtxCancel()
e.wg.Wait()
e.lazyConnMgrMu.Lock()
e.lazyConnMgr = nil
e.lazyConnMgrMu.Unlock()
for _, peerID := range e.peerStore.PeersPubKey() {
e.peerStore.PeerConnOpen(ctx, peerID)
}
}
func (e *ConnMgr) isStartedWithLazyMgr() bool {
return e.lazyConnMgr != nil && e.lazyCtxCancel != nil
}
// resolveLazyForce determines the local override. NB_LAZY_CONN takes precedence; when it
// is unset the MDM policy override (mdmState) applies. Either wins in both directions over
// the management feature flag; StateUnset for both defers to management.
func resolveLazyForce(mdmState lazyconn.State) lazyForce {
state := lazyconn.EnvState()
if state == lazyconn.StateUnset {
state = mdmState
}
switch state {
case lazyconn.StateOn:
return lazyForceOn
case lazyconn.StateOff:
return lazyForceOff
default:
return lazyForceNone
}
}
func inactivityThresholdEnv() *time.Duration {
envValue := os.Getenv(lazyconn.EnvInactivityThreshold)
if envValue == "" {
return nil
}
parsedMinutes, err := strconv.Atoi(envValue)
if err != nil || parsedMinutes <= 0 {
return nil
}
d := time.Duration(parsedMinutes) * time.Minute
return &d
}