Compare commits

..

4 Commits

Author SHA1 Message Date
riccardom
dd2073aec9 Dont log twice 2026-06-25 09:44:23 +02:00
riccardom
c2a046ac8c Distinguish timeout from cancellation in logs 2026-06-22 19:08:07 +02:00
riccardom
ea08b3d2dc Propagate context to OS calls that can leverage its cancellation / timeout 2026-06-22 18:16:37 +02:00
riccardom
c660306cfb Wraps syestem info / posture checks into a goroutine with timeout
e.checks = checks is set before doing the SyncMeta,
so if it fails next time isCheckEquals compares true and bypasses
the update. This is to avoid another repeating the 15 seconds hang.
The checks will be synced on reconnect or posture checks changes
push from mgmt.
2026-06-22 18:16:37 +02:00
29 changed files with 346 additions and 670 deletions

View File

@@ -130,7 +130,7 @@ func debugConfigDump(cmd *cobra.Command, _ []string) error {
client := proto.NewDaemonServiceClient(conn)
resp, err := client.GetConfig(cmd.Context(), &proto.GetConfigRequest{
ProfileName: string(activeProf.ID),
ProfileName: activeProf.Name,
Username: currUser.Username,
})
if err != nil {

View File

@@ -51,20 +51,13 @@ type cachedRecord struct {
}
// Resolver caches critical NetBird infrastructure domains.
// records, refreshing, failedResolves, mgmtDomain and serverDomains are all
// guarded by mutex.
// records, refreshing, mgmtDomain and serverDomains are all guarded by mutex.
type Resolver struct {
records map[dns.Question]*cachedRecord
mgmtDomain *domain.Domain
serverDomains *dnsconfig.ServerDomains
mutex sync.RWMutex
// failedResolves records the last failed initial resolve per domain so a
// domain that never resolves isn't retried on every server-domains update
// until refreshBackoff elapses. Entries are cleared on success and pruned
// to the current server-domains set.
failedResolves map[domain.Domain]time.Time
chain ChainResolver
chainMaxPriority int
refreshGroup singleflight.Group
@@ -83,10 +76,9 @@ type Resolver struct {
// NewResolver creates a new management domains cache resolver.
func NewResolver() *Resolver {
return &Resolver{
records: make(map[dns.Question]*cachedRecord),
refreshing: make(map[dns.Question]*atomic.Bool),
failedResolves: make(map[domain.Domain]time.Time),
cacheTTL: resolveCacheTTL(),
records: make(map[dns.Question]*cachedRecord),
refreshing: make(map[dns.Question]*atomic.Bool),
cacheTTL: resolveCacheTTL(),
}
}
@@ -181,9 +173,7 @@ func (m *Resolver) continueToNext(w dns.ResponseWriter, r *dns.Msg) {
// AddDomain resolves a domain and stores its A/AAAA records in the cache.
// A family that resolves NODATA (nil err, zero records) evicts any stale
// entry for that qtype. When one family hard-errors while the other succeeds,
// the resolved family is still cached but AddDomain returns an error so the
// caller retries the incomplete resolve rather than treating it as complete.
// entry for that qtype.
func (m *Resolver) AddDomain(ctx context.Context, d domain.Domain) error {
dnsName := strings.ToLower(dns.Fqdn(d.PunycodeString()))
@@ -213,10 +203,6 @@ func (m *Resolver) AddDomain(ctx context.Context, d domain.Domain) error {
log.Debugf("added/updated domain=%s with %d A records and %d AAAA records",
d.SafeString(), len(aRecords), len(aaaaRecords))
if errA != nil || errAAAA != nil {
return fmt.Errorf("resolve %s: incomplete, a family failed: %w", d.SafeString(), errors.Join(errA, errAAAA))
}
return nil
}
@@ -476,7 +462,6 @@ func (m *Resolver) RemoveDomain(d domain.Domain) error {
delete(m.records, qAAAA)
delete(m.refreshing, qA)
delete(m.refreshing, qAAAA)
delete(m.failedResolves, d)
log.Debugf("removed domain=%s from cache", d.SafeString())
return nil
@@ -520,7 +505,6 @@ func (m *Resolver) UpdateFromServerDomains(ctx context.Context, serverDomains dn
allDomains := m.extractDomainsFromServerDomains(updatedServerDomains)
currentDomains := m.GetCachedDomains()
removedDomains = m.removeStaleDomains(currentDomains, allDomains)
m.pruneFailedResolves(allDomains)
}
m.addNewDomains(ctx, newDomains)
@@ -593,85 +577,13 @@ func (m *Resolver) isManagementDomain(domain domain.Domain) bool {
return m.mgmtDomain != nil && domain == *m.mgmtDomain
}
// addNewDomains resolves and caches domains that are not yet in the cache,
// running the lookups concurrently. Domains already cached are skipped and left
// to the stale-while-revalidate refresh path, so a sync never re-resolves them
// synchronously: once NetBird owns the OS resolver the resolve runs through the
// handler chain and would otherwise dial the managed upstreams under the engine
// sync lock on every update.
// addNewDomains resolves and caches all domains from the update
func (m *Resolver) addNewDomains(ctx context.Context, newDomains domain.List) {
var wg sync.WaitGroup
seen := make(map[domain.Domain]struct{}, len(newDomains))
for _, newDomain := range newDomains {
if _, dup := seen[newDomain]; dup {
continue
}
seen[newDomain] = struct{}{}
if !m.needsResolve(newDomain) {
continue
}
wg.Add(1)
go func(d domain.Domain) {
defer wg.Done()
if err := m.AddDomain(ctx, d); err != nil {
m.markResolveFailed(d)
log.Warnf("failed to add/update domain=%s: %v", d.SafeString(), err)
return
}
m.clearResolveFailed(d)
log.Debugf("added/updated management cache domain=%s", d.SafeString())
}(newDomain)
}
wg.Wait()
}
// needsResolve reports whether d should be resolved now. A recent failed or
// incomplete resolve gates retries on the backoff even when one family is
// already cached, so a transiently-failed family is retried instead of being
// treated as fully resolved. Otherwise a domain with any cached record is left
// to the stale-while-revalidate refresh path.
func (m *Resolver) needsResolve(d domain.Domain) bool {
dnsName := strings.ToLower(dns.Fqdn(d.PunycodeString()))
m.mutex.RLock()
defer m.mutex.RUnlock()
if failedAt, ok := m.failedResolves[d]; ok {
return time.Since(failedAt) >= refreshBackoff
}
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 false
}
}
return true
}
func (m *Resolver) markResolveFailed(d domain.Domain) {
m.mutex.Lock()
m.failedResolves[d] = time.Now()
m.mutex.Unlock()
}
func (m *Resolver) clearResolveFailed(d domain.Domain) {
m.mutex.Lock()
delete(m.failedResolves, d)
m.mutex.Unlock()
}
// pruneFailedResolves drops failure markers for domains no longer present in
// the server-domains set, keeping the map bounded to the current set (a
// failed-only domain has no cached record, so RemoveDomain never sees it).
func (m *Resolver) pruneFailedResolves(domains domain.List) {
m.mutex.Lock()
defer m.mutex.Unlock()
for d := range m.failedResolves {
if !slices.Contains(domains, d) {
delete(m.failedResolves, d)
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())
}
}
}

View File

@@ -21,7 +21,6 @@ type fakeChain struct {
mu sync.Mutex
calls map[string]int
answers map[string][]dns.RR
qErr map[string]error
err error
hasRoot bool
onLookup func()
@@ -31,7 +30,6 @@ func newFakeChain() *fakeChain {
return &fakeChain{
calls: map[string]int{},
answers: map[string][]dns.RR{},
qErr: map[string]error{},
hasRoot: true,
}
}
@@ -49,9 +47,6 @@ func (f *fakeChain) ResolveInternal(ctx context.Context, msg *dns.Msg, maxPriori
f.calls[key]++
answers := f.answers[key]
err := f.err
if err == nil {
err = f.qErr[key]
}
onLookup := f.onLookup
f.mu.Unlock()
@@ -80,12 +75,6 @@ func (f *fakeChain) setAnswer(name string, qtype uint16, ip string) {
}
}
func (f *fakeChain) setErr(name string, qtype uint16, err error) {
f.mu.Lock()
defer f.mu.Unlock()
f.qErr[name+"|"+dns.TypeToString[qtype]] = err
}
func (f *fakeChain) callCount(name string, qtype uint16) int {
f.mu.Lock()
defer f.mu.Unlock()

View File

@@ -1,183 +0,0 @@
package mgmt
import (
"context"
"errors"
"sync/atomic"
"testing"
"time"
"github.com/miekg/dns"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
dnsconfig "github.com/netbirdio/netbird/client/internal/dns/config"
"github.com/netbirdio/netbird/shared/management/domain"
)
// A domain already in the cache must not be re-resolved on a subsequent server
// domains update; it is left to the stale-while-revalidate refresh path.
func TestResolver_UpdateFromServerDomains_SkipsCached(t *testing.T) {
r := NewResolver()
chain := newFakeChain()
chain.setAnswer("signal.example.com.", dns.TypeA, "10.0.0.2")
r.SetChainResolver(chain, 50)
sd := dnsconfig.ServerDomains{Signal: domain.Domain("signal.example.com")}
_, err := r.UpdateFromServerDomains(context.Background(), sd)
require.NoError(t, err)
require.Equal(t, 1, chain.callCount("signal.example.com.", dns.TypeA),
"first update must resolve the domain")
_, err = r.UpdateFromServerDomains(context.Background(), sd)
require.NoError(t, err)
assert.Equal(t, 1, chain.callCount("signal.example.com.", dns.TypeA),
"cached domain must not be re-resolved on a subsequent update")
}
// New domains in a single update must resolve concurrently rather than serially.
func TestResolver_AddNewDomains_ResolvesConcurrently(t *testing.T) {
r := NewResolver()
chain := newFakeChain()
var inflight, maxInflight atomic.Int32
chain.onLookup = func() {
n := inflight.Add(1)
for {
old := maxInflight.Load()
if n <= old || maxInflight.CompareAndSwap(old, n) {
break
}
}
time.Sleep(50 * time.Millisecond)
inflight.Add(-1)
}
relays := []domain.Domain{"a.example.com", "b.example.com", "c.example.com", "d.example.com"}
for _, d := range relays {
chain.setAnswer(dns.Fqdn(string(d)), dns.TypeA, "10.0.0.2")
}
r.SetChainResolver(chain, 50)
start := time.Now()
_, err := r.UpdateFromServerDomains(context.Background(), dnsconfig.ServerDomains{Relay: relays})
require.NoError(t, err)
elapsed := time.Since(start)
assert.GreaterOrEqual(t, int(maxInflight.Load()), 2, "domains must resolve concurrently")
// Serial resolution of 4 domains would take at least 4*50ms; concurrent is far less.
assert.Less(t, elapsed, 300*time.Millisecond, "resolution should not be serial")
}
// A domain that fails to resolve must not be retried on every update; the
// failure backoff suppresses re-resolution until it expires.
func TestResolver_UpdateFromServerDomains_BacksOffFailures(t *testing.T) {
r := NewResolver()
chain := newFakeChain()
chain.err = errors.New("resolve boom")
r.SetChainResolver(chain, 50)
sd := dnsconfig.ServerDomains{Signal: domain.Domain("signal.example.com")}
_, err := r.UpdateFromServerDomains(context.Background(), sd)
require.NoError(t, err)
require.Equal(t, 1, chain.callCount("signal.example.com.", dns.TypeA),
"first update must attempt the resolve")
_, err = r.UpdateFromServerDomains(context.Background(), sd)
require.NoError(t, err)
assert.Equal(t, 1, chain.callCount("signal.example.com.", dns.TypeA),
"failed resolve must back off and not retry on the next update")
}
// A domain listed under more than one server-domain type (e.g. STUN and TURN on
// the same host) must be resolved once per update, not once per occurrence.
func TestResolver_AddNewDomains_DedupesDuplicateDomains(t *testing.T) {
r := NewResolver()
chain := newFakeChain()
chain.setAnswer("dup.example.com.", dns.TypeA, "10.0.0.9")
r.SetChainResolver(chain, 50)
sd := dnsconfig.ServerDomains{
Stuns: []domain.Domain{"dup.example.com"},
Turns: []domain.Domain{"dup.example.com"},
}
_, err := r.UpdateFromServerDomains(context.Background(), sd)
require.NoError(t, err)
assert.Equal(t, 1, chain.callCount("dup.example.com.", dns.TypeA),
"a domain appearing under multiple server-domain types must resolve once")
}
// A failure marker must be dropped once its domain leaves the server-domains set
// so the map stays bounded to the current set.
func TestResolver_UpdateFromServerDomains_PrunesFailedResolves(t *testing.T) {
r := NewResolver()
chain := newFakeChain()
chain.err = errors.New("resolve boom")
r.SetChainResolver(chain, 50)
_, err := r.UpdateFromServerDomains(context.Background(), dnsconfig.ServerDomains{Signal: domain.Domain("gone.example.com")})
require.NoError(t, err)
r.mutex.RLock()
_, marked := r.failedResolves[domain.Domain("gone.example.com")]
r.mutex.RUnlock()
require.True(t, marked, "failed resolve must be recorded")
_, err = r.UpdateFromServerDomains(context.Background(), dnsconfig.ServerDomains{Signal: domain.Domain("other.example.com")})
require.NoError(t, err)
r.mutex.RLock()
_, stillMarked := r.failedResolves[domain.Domain("gone.example.com")]
r.mutex.RUnlock()
assert.False(t, stillMarked, "failure marker for a domain no longer in the set must be pruned")
}
// When one family hard-errors while the other resolves, the domain is cached
// for the working family but recorded as incomplete so the failed family is
// retried under backoff instead of being treated as fully resolved forever.
func TestResolver_AddNewDomains_RetriesPartialFamilyFailure(t *testing.T) {
d := domain.Domain("relay.example.com")
r := NewResolver()
chain := newFakeChain()
chain.setAnswer("relay.example.com.", dns.TypeA, "10.0.0.2")
chain.setErr("relay.example.com.", dns.TypeAAAA, errors.New("servfail"))
r.SetChainResolver(chain, 50)
_, err := r.UpdateFromServerDomains(context.Background(), dnsconfig.ServerDomains{Relay: []domain.Domain{d}})
require.NoError(t, err)
r.mutex.RLock()
_, aCached := r.records[dns.Question{Name: "relay.example.com.", Qtype: dns.TypeA, Qclass: dns.ClassINET}]
_, marked := r.failedResolves[d]
r.mutex.RUnlock()
require.True(t, aCached, "the working family must still be cached")
require.True(t, marked, "a partial failure must be recorded so the failed family is retried")
assert.False(t, r.needsResolve(d), "within the backoff window the domain is not retried")
r.mutex.Lock()
r.failedResolves[d] = time.Now().Add(-2 * refreshBackoff)
r.mutex.Unlock()
assert.True(t, r.needsResolve(d), "after the backoff elapses the domain is retried to pick up the missing family")
}
// A family that returns NODATA (legitimately absent, e.g. an IPv4-only host) is
// not a failure: the domain must not be marked for retry, otherwise it would be
// re-resolved on every sync.
func TestResolver_AddNewDomains_NodataIsNotFailure(t *testing.T) {
d := domain.Domain("v4only.example.com")
r := NewResolver()
chain := newFakeChain()
chain.setAnswer("v4only.example.com.", dns.TypeA, "10.0.0.2")
r.SetChainResolver(chain, 50)
_, err := r.UpdateFromServerDomains(context.Background(), dnsconfig.ServerDomains{Relay: []domain.Domain{d}})
require.NoError(t, err)
r.mutex.RLock()
_, marked := r.failedResolves[d]
r.mutex.RUnlock()
assert.False(t, marked, "a NODATA family must not be recorded as a failure")
assert.False(t, r.needsResolve(d), "an IPv4-only host must not be re-resolved on later syncs")
}

View File

@@ -82,6 +82,12 @@ const (
PeerConnectionTimeoutMax = 45000 // ms
PeerConnectionTimeoutMin = 30000 // ms
disableAutoUpdate = "disabled"
// systemInfoTimeout bounds how long the sync loop waits for system info / posture
// check gathering. The gathering runs uncancellable system calls (process scan,
// exec, os.Stat); without this bound a single stuck call freezes handleSync, and
// thus syncMsgMux, for as long as the call hangs (observed multi-minute freezes).
systemInfoTimeout = 15 * time.Second
)
var ErrResetConnection = fmt.Errorf("reset connection")
@@ -1066,11 +1072,22 @@ func (e *Engine) updateChecksIfNew(checks []*mgmProto.Checks) error {
}
e.checks = checks
info, err := system.GetInfoWithChecks(e.ctx, checks)
if err != nil {
log.Warnf("failed to get system info with checks: %v", err)
info = system.GetInfo(e.ctx)
info, ok := system.GetInfoWithChecksTimeout(e.ctx, systemInfoTimeout, checks)
if !ok {
// Gathering timed out; skip the meta sync this cycle rather than blocking the
// sync loop (and syncMsgMux) on a stuck system call. A later sync will retry.
return nil
}
e.applyInfoFlags(info)
if err := e.mgmClient.SyncMeta(info); err != nil {
return fmt.Errorf("could not sync meta: error %s", err)
}
return nil
}
// applyInfoFlags sets the engine's config-derived feature flags on the gathered system info.
func (e *Engine) applyInfoFlags(info *system.Info) {
info.SetFlags(
e.config.RosenpassEnabled,
e.config.RosenpassPermissive,
@@ -1089,12 +1106,6 @@ func (e *Engine) updateChecksIfNew(checks []*mgmProto.Checks) error {
e.config.EnableSSHRemotePortForwarding,
e.config.DisableSSHAuth,
)
if err := e.mgmClient.SyncMeta(info); err != nil {
log.Errorf("could not sync meta: error %s", err)
return err
}
return nil
}
func (e *Engine) updateConfig(conf *mgmProto.PeerConfig) error {
@@ -1240,31 +1251,15 @@ func (e *Engine) receiveManagementEvents() {
e.shutdownWg.Add(1)
go func() {
defer e.shutdownWg.Done()
info, err := system.GetInfoWithChecks(e.ctx, e.checks)
if err != nil {
log.Warnf("failed to get system info with checks: %v", err)
info, ok := system.GetInfoWithChecksTimeout(e.ctx, systemInfoTimeout, e.checks)
if !ok {
// Gathering timed out; connect the stream with base info so management
// connectivity still comes up rather than blocking here.
info = system.GetInfo(e.ctx)
}
info.SetFlags(
e.config.RosenpassEnabled,
e.config.RosenpassPermissive,
&e.config.ServerSSHAllowed,
e.config.DisableClientRoutes,
e.config.DisableServerRoutes,
e.config.DisableDNS,
e.config.DisableFirewall,
e.config.BlockLANAccess,
e.config.BlockInbound,
e.config.DisableIPv6,
e.config.LazyConnectionEnabled,
e.config.EnableSSHRoot,
e.config.EnableSSHSFTP,
e.config.EnableSSHLocalPortForwarding,
e.config.EnableSSHRemotePortForwarding,
e.config.DisableSSHAuth,
)
e.applyInfoFlags(info)
err = e.mgmClient.Sync(e.ctx, info, e.handleSync)
err := e.mgmClient.Sync(e.ctx, info, e.handleSync)
if err != nil {
// happens if management is unavailable for a long time.
// We want to cancel the operation of the whole client

View File

@@ -2,8 +2,10 @@ package system
import (
"context"
"errors"
"net/netip"
"strings"
"time"
log "github.com/sirupsen/logrus"
"google.golang.org/grpc/metadata"
@@ -154,7 +156,7 @@ func GetInfoWithChecks(ctx context.Context, checks []*proto.Checks) (*Info, erro
processCheckPaths = append(processCheckPaths, check.GetFiles()...)
}
files, err := checkFileAndProcess(processCheckPaths)
files, err := checkFileAndProcess(ctx, processCheckPaths)
if err != nil {
return nil, err
}
@@ -166,3 +168,39 @@ func GetInfoWithChecks(ctx context.Context, checks []*proto.Checks) (*Info, erro
log.Debugf("all system information gathered successfully")
return info, nil
}
// GetInfoWithChecksTimeout is GetInfoWithChecks bounded by timeout. Posture-check gathering
// runs uncancellable system calls (process enumeration, os.Stat), so calling it inline can
// block the caller for as long as such a call hangs. It runs in a goroutine instead: if it
// does not return within timeout the caller gets (nil, false) and should proceed with
// degraded behavior rather than block. On a gathering error it falls back to base GetInfo.
//
// The buffered channel lets the abandoned goroutine finish and exit once its blocking call
// returns, so it does not leak beyond the duration of that call.
func GetInfoWithChecksTimeout(ctx context.Context, timeout time.Duration, checks []*proto.Checks) (*Info, bool) {
ctx, cancel := context.WithTimeout(ctx, timeout)
defer cancel()
infoCh := make(chan *Info, 1)
go func() {
info, err := GetInfoWithChecks(ctx, checks)
if err != nil {
log.Warnf("failed to get system info with checks: %v", err)
info = GetInfo(ctx)
}
infoCh <- info
}()
select {
case info := <-infoCh:
return info, true
case <-ctx.Done():
if errors.Is(ctx.Err(), context.DeadlineExceeded) {
log.Warnf("gathering system info with checks timed out after %s", timeout)
} else {
// Parent context canceled (e.g. shutdown), not a timeout.
log.Warnf("gathering system info with checks canceled: %v", ctx.Err())
}
return nil, false
}
}

View File

@@ -50,7 +50,7 @@ func GetInfo(ctx context.Context) *Info {
}
// checkFileAndProcess checks if the file path exists and if a process is running at that path.
func checkFileAndProcess(paths []string) ([]File, error) {
func checkFileAndProcess(_ context.Context, _ []string) ([]File, error) {
return []File{}, nil
}

View File

@@ -32,7 +32,7 @@ func GetInfo(ctx context.Context) *Info {
sysName := string(bytes.Split(utsname.Sysname[:], []byte{0})[0])
machine := string(bytes.Split(utsname.Machine[:], []byte{0})[0])
release := string(bytes.Split(utsname.Release[:], []byte{0})[0])
swVersion, err := exec.Command("sw_vers", "-productVersion").Output()
swVersion, err := exec.CommandContext(ctx, "sw_vers", "-productVersion").Output()
if err != nil {
log.Warnf("got an error while retrieving macOS version with sw_vers, error: %s. Using darwin version instead.\n", err)
swVersion = []byte(release)

View File

@@ -105,7 +105,7 @@ func isDuplicated(addresses []NetworkAddress, addr NetworkAddress) bool {
}
// checkFileAndProcess checks if the file path exists and if a process is running at that path.
func checkFileAndProcess(paths []string) ([]File, error) {
func checkFileAndProcess(_ context.Context, _ []string) ([]File, error) {
return []File{}, nil
}

View File

@@ -103,7 +103,7 @@ func collectLocationInfo(info *Info) {
}
}
func checkFileAndProcess(_ []string) ([]File, error) {
func checkFileAndProcess(_ context.Context, _ []string) ([]File, error) {
return []File{}, nil
}

View File

@@ -3,6 +3,7 @@ package system
import (
"context"
"testing"
"time"
"github.com/stretchr/testify/assert"
"google.golang.org/grpc/metadata"
@@ -34,6 +35,20 @@ func Test_CustomHostname(t *testing.T) {
assert.Equal(t, want, got.Hostname)
}
func TestGetInfoWithChecksTimeout_Success(t *testing.T) {
info, ok := GetInfoWithChecksTimeout(context.Background(), 30*time.Second, nil)
assert.True(t, ok, "expected gathering to complete within the timeout")
assert.NotNil(t, info)
}
func TestGetInfoWithChecksTimeout_Timeout(t *testing.T) {
// A 1ns budget expires before the (real) system-info gathering can finish, so the
// caller must get (nil, false) instead of blocking on the in-flight goroutine.
info, ok := GetInfoWithChecksTimeout(context.Background(), time.Nanosecond, nil)
assert.False(t, ok, "expected timeout to be reported")
assert.Nil(t, info)
}
func Test_NetAddresses(t *testing.T) {
addr, err := networkAddresses()
if err != nil {

View File

@@ -3,24 +3,30 @@
package system
import (
"context"
"os"
"slices"
"github.com/shirou/gopsutil/v3/process"
)
// getRunningProcesses returns a list of running process paths.
func getRunningProcesses() ([]string, error) {
processIDs, err := process.Pids()
// getRunningProcesses returns a list of running process paths. The context bounds the work:
// the per-PID loop bails as soon as ctx is done, and the gopsutil calls honor it where they
// can, so a stuck enumeration cannot run unbounded.
func getRunningProcesses(ctx context.Context) ([]string, error) {
processIDs, err := process.PidsWithContext(ctx)
if err != nil {
return nil, err
}
processMap := make(map[string]bool)
for _, pID := range processIDs {
if err := ctx.Err(); err != nil {
return nil, err
}
p := &process.Process{Pid: pID}
path, _ := p.Exe()
path, _ := p.ExeWithContext(ctx)
if path != "" {
processMap[path] = false
}
@@ -35,18 +41,21 @@ func getRunningProcesses() ([]string, error) {
}
// checkFileAndProcess checks if the file path exists and if a process is running at that path.
func checkFileAndProcess(paths []string) ([]File, error) {
func checkFileAndProcess(ctx context.Context, paths []string) ([]File, error) {
files := make([]File, len(paths))
if len(paths) == 0 {
return files, nil
}
runningProcesses, err := getRunningProcesses()
runningProcesses, err := getRunningProcesses(ctx)
if err != nil {
return nil, err
}
for i, path := range paths {
if err := ctx.Err(); err != nil {
return nil, err
}
file := File{Path: path}
_, err := os.Stat(path)

View File

@@ -1,6 +1,7 @@
package system
import (
"context"
"testing"
"github.com/shirou/gopsutil/v3/process"
@@ -9,7 +10,7 @@ import (
func Benchmark_getRunningProcesses(b *testing.B) {
b.Run("getRunningProcesses new", func(b *testing.B) {
for i := 0; i < b.N; i++ {
ps, err := getRunningProcesses()
ps, err := getRunningProcesses(context.Background())
if err != nil {
b.Fatalf("unexpected error: %v", err)
}
@@ -29,12 +30,38 @@ func Benchmark_getRunningProcesses(b *testing.B) {
}
}
})
s, _ := getRunningProcesses()
s, _ := getRunningProcesses(context.Background())
b.Logf("getRunningProcesses returned %d processes", len(s))
s, _ = getRunningProcessesOld()
b.Logf("getRunningProcessesOld returned %d processes", len(s))
}
func TestCheckFileAndProcess_ContextCanceled(t *testing.T) {
ctx, cancel := context.WithCancel(context.Background())
cancel()
// With a canceled context and non-empty paths the gathering must bail with an error
// instead of running the (potentially blocking) process scan / stat loop.
if _, err := checkFileAndProcess(ctx, []string{"/does/not/exist"}); err == nil {
t.Fatal("expected error on canceled context, got nil")
}
}
func TestCheckFileAndProcess_EmptyPaths(t *testing.T) {
// No check paths means no work to do: it must return immediately with no error,
// even on a canceled context (nothing to scan or stat).
ctx, cancel := context.WithCancel(context.Background())
cancel()
files, err := checkFileAndProcess(ctx, nil)
if err != nil {
t.Fatalf("unexpected error for empty paths: %v", err)
}
if len(files) != 0 {
t.Fatalf("expected no files, got %d", len(files))
}
}
func getRunningProcessesOld() ([]string, error) {
processes, err := process.Processes()
if err != nil {

View File

@@ -11,7 +11,7 @@ import (
const (
reconnThreshold = 5 * time.Minute
baseBlockDuration = 30 * time.Minute // Duration for which a peer is banned after exceeding the reconnection limit
baseBlockDuration = 10 * time.Minute // Duration for which a peer is banned after exceeding the reconnection limit
reconnLimitForBan = 30 // Number of reconnections within the reconnTreshold that triggers a ban
metaChangeLimit = 3 // Number of reconnections with different metadata that triggers a ban of one peer
)
@@ -139,13 +139,22 @@ func (l *loginFilter) addLogin(wgPubKey string, metaHash uint64) {
state.lastSeen = now
}
func metaHash(meta nbpeer.PeerSystemMeta) uint64 {
func metaHash(meta nbpeer.PeerSystemMeta, pubip string) uint64 {
h := fnv.New64a()
h.Write([]byte(meta.WtVersion))
h.Write([]byte(meta.OSVersion))
h.Write([]byte(meta.KernelVersion))
h.Write([]byte(meta.Hostname))
h.Write([]byte(meta.SystemSerialNumber))
h.Write([]byte(pubip))
return h.Sum64()
macs := uint64(0)
for _, na := range meta.NetworkAddresses {
for _, r := range na.Mac {
macs += uint64(r)
}
}
return h.Sum64() + macs
}

View File

@@ -164,7 +164,9 @@ func BenchmarkHashingMethods(b *testing.B) {
KernelVersion: "5.15.0-76-generic",
Hostname: "prod-server-database-01",
SystemSerialNumber: "PC-1234567890",
NetworkAddresses: []nbpeer.NetworkAddress{{Mac: "00:1B:44:11:3A:B7"}, {Mac: "00:1B:44:11:3A:B8"}},
}
pubip := "8.8.8.8"
var resultString string
var resultUint uint64
@@ -173,7 +175,7 @@ func BenchmarkHashingMethods(b *testing.B) {
b.ReportAllocs()
b.ResetTimer()
for i := 0; i < b.N; i++ {
resultString = builderString(meta)
resultString = builderString(meta, pubip)
}
})
@@ -181,7 +183,7 @@ func BenchmarkHashingMethods(b *testing.B) {
b.ReportAllocs()
b.ResetTimer()
for i := 0; i < b.N; i++ {
resultString = fnvHashToString(meta)
resultString = fnvHashToString(meta, pubip)
}
})
@@ -189,7 +191,7 @@ func BenchmarkHashingMethods(b *testing.B) {
b.ReportAllocs()
b.ResetTimer()
for i := 0; i < b.N; i++ {
resultUint = metaHash(meta)
resultUint = metaHash(meta, pubip)
}
})
@@ -197,20 +199,29 @@ func BenchmarkHashingMethods(b *testing.B) {
_ = resultUint
}
func fnvHashToString(meta nbpeer.PeerSystemMeta) string {
func fnvHashToString(meta nbpeer.PeerSystemMeta, pubip string) string {
h := fnv.New64a()
if len(meta.NetworkAddresses) != 0 {
for _, na := range meta.NetworkAddresses {
h.Write([]byte(na.Mac))
}
}
h.Write([]byte(meta.WtVersion))
h.Write([]byte(meta.OSVersion))
h.Write([]byte(meta.KernelVersion))
h.Write([]byte(meta.Hostname))
h.Write([]byte(meta.SystemSerialNumber))
h.Write([]byte(pubip))
return strconv.FormatUint(h.Sum64(), 16)
}
func builderString(meta nbpeer.PeerSystemMeta) string {
estimatedSize := len(meta.WtVersion) + len(meta.OSVersion) + len(meta.KernelVersion) + len(meta.Hostname) + len(meta.SystemSerialNumber) + 4
func builderString(meta nbpeer.PeerSystemMeta, pubip string) string {
mac := getMacAddress(meta.NetworkAddresses)
estimatedSize := len(meta.WtVersion) + len(meta.OSVersion) + len(meta.KernelVersion) + len(meta.Hostname) + len(meta.SystemSerialNumber) +
len(pubip) + len(mac) + 6
var b strings.Builder
b.Grow(estimatedSize)
@@ -224,10 +235,23 @@ func builderString(meta nbpeer.PeerSystemMeta) string {
b.WriteString(meta.Hostname)
b.WriteByte('|')
b.WriteString(meta.SystemSerialNumber)
b.WriteByte('|')
b.WriteString(pubip)
return b.String()
}
func getMacAddress(nas []nbpeer.NetworkAddress) string {
if len(nas) == 0 {
return ""
}
macs := make([]string, 0, len(nas))
for _, na := range nas {
macs = append(macs, na.Mac)
}
return strings.Join(macs, "/")
}
func BenchmarkLoginFilter_ParallelLoad(b *testing.B) {
filter := newLoginFilterWithCfg(testAdvancedCfg())
numKeys := 100000

View File

@@ -254,7 +254,7 @@ func (s *Server) Sync(req *proto.EncryptedMessage, srv proto.ManagementService_S
return mapError(ctx, err)
}
metahashed := metaHash(peerMeta)
metahashed := metaHash(peerMeta, sRealIP)
if userID == "" && !s.loginFilter.allowLogin(peerKey.String(), metahashed) {
if s.appMetrics != nil {
s.appMetrics.GRPCMetrics().CountSyncRequestBlocked()
@@ -306,7 +306,7 @@ func (s *Server) Sync(req *proto.EncryptedMessage, srv proto.ManagementService_S
log.WithContext(ctx).Tracef("peer system meta has to be provided on sync. Peer %s, remote addr %s", peerKey.String(), realIP)
}
metahash := metaHash(peerMeta)
metahash := metaHash(peerMeta, realIP.String())
s.loginFilter.addLogin(peerKey.String(), metahash)
peer, netMap, postureChecks, dnsFwdPort, err := s.accountManager.SyncAndMarkPeer(ctx, accountID, peerKey.String(), peerMeta, realIP, syncStart)
@@ -732,7 +732,7 @@ func (s *Server) Login(ctx context.Context, req *proto.EncryptedMessage) (*proto
}
peerMeta := extractPeerMeta(ctx, loginReq.GetMeta())
metahashed := metaHash(peerMeta)
metahashed := metaHash(peerMeta, sRealIP)
if !s.loginFilter.allowLogin(peerKey.String(), metahashed) {
if s.logBlockedPeers {
log.WithContext(ctx).Tracef("peer %s with meta hash %d is blocked from login", peerKey.String(), metahashed)
@@ -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
}

View File

@@ -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
}

View File

@@ -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)

View File

@@ -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.

View File

@@ -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)
@@ -1916,117 +1916,6 @@ func TestDefaultAccountManager_MarkPeerConnected_PeerLoginExpiration(t *testing.
}
}
func TestDefaultAccountManager_MarkPeerDisconnected_SchedulesInactivityExpiration(t *testing.T) {
manager, _, err := createManager(t)
require.NoError(t, err, "unable to create account manager")
accountID, err := manager.GetAccountIDByUserID(context.Background(), auth.UserAuth{UserId: userID})
require.NoError(t, err, "unable to create an account")
key, err := wgtypes.GenerateKey()
require.NoError(t, err, "unable to generate WireGuard key")
peerPubKey := key.PublicKey().String()
_, _, _, _, err = manager.AddPeer(context.Background(), "", "", userID, &nbpeer.Peer{
Key: peerPubKey,
Meta: nbpeer.PeerSystemMeta{Hostname: "test-peer"},
InactivityExpirationEnabled: true,
}, false)
require.NoError(t, err, "unable to add peer")
_, err = manager.UpdateAccountSettings(context.Background(), accountID, userID, &types.Settings{
PeerLoginExpiration: time.Hour,
PeerLoginExpirationEnabled: true,
PeerInactivityExpiration: time.Hour,
PeerInactivityExpirationEnabled: true,
Extra: &types.ExtraSettings{},
})
require.NoError(t, err, "expecting to update account settings successfully but got error")
// Establish a session so the matching-token disconnect is actually applied.
streamStartTime := time.Now().UTC()
err = manager.MarkPeerConnected(context.Background(), peerPubKey, accountID, streamStartTime.UnixNano(), nil)
require.NoError(t, err, "unable to mark peer connected")
// Install the mock only now, so the assertion observes the disconnect, not
// the earlier connect.
scheduled := make(chan struct{}, 1)
manager.peerInactivityExpiry = &MockScheduler{
CancelFunc: func(ctx context.Context, IDs []string) {},
ScheduleFunc: func(ctx context.Context, in time.Duration, ID string, job func() (nextRunIn time.Duration, reschedule bool)) {
select {
case scheduled <- struct{}{}:
default:
}
},
}
err = manager.MarkPeerDisconnected(context.Background(), peerPubKey, accountID, streamStartTime.UnixNano())
require.NoError(t, err, "unable to mark peer disconnected")
select {
case <-scheduled:
// expected: disconnect re-armed the inactivity expiry timer
case <-time.After(time.Second):
t.Fatal("expected inactivity expiration to be rescheduled when an eligible peer disconnects")
}
}
func TestDefaultAccountManager_MarkPeerDisconnected_SkipsInactivityExpirationWhenDisabled(t *testing.T) {
manager, _, err := createManager(t)
require.NoError(t, err, "unable to create account manager")
accountID, err := manager.GetAccountIDByUserID(context.Background(), auth.UserAuth{UserId: userID})
require.NoError(t, err, "unable to create an account")
key, err := wgtypes.GenerateKey()
require.NoError(t, err, "unable to generate WireGuard key")
peerPubKey := key.PublicKey().String()
_, _, _, _, err = manager.AddPeer(context.Background(), "", "", userID, &nbpeer.Peer{
Key: peerPubKey,
Meta: nbpeer.PeerSystemMeta{Hostname: "test-peer"},
InactivityExpirationEnabled: true,
}, false)
require.NoError(t, err, "unable to add peer")
// Peer is eligible (SSO + inactivity enabled) but the account-level setting
// stays disabled, so disconnect must not schedule anything.
_, err = manager.UpdateAccountSettings(context.Background(), accountID, userID, &types.Settings{
PeerLoginExpiration: time.Hour,
PeerLoginExpirationEnabled: true,
PeerInactivityExpiration: time.Hour,
PeerInactivityExpirationEnabled: false,
Extra: &types.ExtraSettings{},
})
require.NoError(t, err, "expecting to update account settings successfully but got error")
streamStartTime := time.Now().UTC()
err = manager.MarkPeerConnected(context.Background(), peerPubKey, accountID, streamStartTime.UnixNano(), nil)
require.NoError(t, err, "unable to mark peer connected")
scheduled := make(chan struct{}, 1)
manager.peerInactivityExpiry = &MockScheduler{
CancelFunc: func(ctx context.Context, IDs []string) {},
ScheduleFunc: func(ctx context.Context, in time.Duration, ID string, job func() (nextRunIn time.Duration, reschedule bool)) {
select {
case scheduled <- struct{}{}:
default:
}
},
}
err = manager.MarkPeerDisconnected(context.Background(), peerPubKey, accountID, streamStartTime.UnixNano())
require.NoError(t, err, "unable to mark peer disconnected")
select {
case <-scheduled:
t.Fatal("inactivity expiration must not be scheduled while the account-level setting is disabled")
case <-time.After(200 * time.Millisecond):
// expected: nothing scheduled
}
}
func TestDefaultAccountManager_OnPeerDisconnected_LastSeenCheck(t *testing.T) {
manager, _, err := createManager(t)
require.NoError(t, err, "unable to create account manager")
@@ -2046,7 +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)
@@ -2067,7 +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)
@@ -2091,7 +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)
@@ -2101,7 +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)
@@ -2163,7 +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)
}()
}
@@ -2204,7 +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{}

View File

@@ -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")
}

View File

@@ -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
}
@@ -188,40 +192,27 @@ func (am *DefaultAccountManager) MarkPeerDisconnected(ctx context.Context, peerP
}
}
if peer.AddedWithSSOLogin() && peer.InactivityExpirationEnabled {
settings, err := am.Store.GetAccountSettings(ctx, store.LockingStrengthNone, accountID)
if err != nil {
log.WithContext(ctx).Warnf("failed getting account settings to schedule inactivity expiration for peer %s: %v", peer.ID, err)
} else if settings.PeerInactivityExpirationEnabled {
am.checkAndSchedulePeerInactivityExpiration(ctx, accountID)
}
}
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)
}
}
@@ -989,8 +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)
@@ -1020,10 +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 {
@@ -1051,10 +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)
}
@@ -1071,8 +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)

View File

@@ -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

View File

@@ -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"`

View File

@@ -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{}).

View File

@@ -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")

View File

@@ -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

View File

@@ -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()

View File

@@ -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