// Copyright 2015 The etcd Authors // // Licensed under the Apache License, Version 2.0 (the "License"); // you may not use this file except in compliance with the License. // You may obtain a copy of the License at // // http://www.apache.org/licenses/LICENSE-2.0 // // Unless required by applicable law or agreed to in writing, software // distributed under the License is distributed on an "AS IS" BASIS, // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. // See the License for the specific language governing permissions and // limitations under the License. package etcdserver import ( "context" "encoding/json" "expvar" "fmt" "math" "math/rand" "net/http" "path" "regexp" "strconv" "sync" "sync/atomic" "time" "github.com/coreos/go-semver/semver" humanize "github.com/dustin/go-humanize" "github.com/prometheus/client_golang/prometheus" "go.etcd.io/etcd/pkg/v3/notify" "go.etcd.io/etcd/pkg/v3/runtime" "go.etcd.io/etcd/server/v3/config" "go.etcd.io/etcd/server/v3/etcdserver/apply" "go.etcd.io/etcd/server/v3/etcdserver/errors" "go.uber.org/zap" pb "go.etcd.io/etcd/api/v3/etcdserverpb" "go.etcd.io/etcd/api/v3/membershippb" "go.etcd.io/etcd/api/v3/version" "go.etcd.io/etcd/client/pkg/v3/fileutil" "go.etcd.io/etcd/client/pkg/v3/types" "go.etcd.io/etcd/client/pkg/v3/verify" "go.etcd.io/etcd/pkg/v3/idutil" "go.etcd.io/etcd/pkg/v3/pbutil" "go.etcd.io/etcd/pkg/v3/schedule" "go.etcd.io/etcd/pkg/v3/traceutil" "go.etcd.io/etcd/pkg/v3/wait" "go.etcd.io/etcd/raft/v3" "go.etcd.io/etcd/raft/v3/raftpb" "go.etcd.io/etcd/server/v3/auth" "go.etcd.io/etcd/server/v3/etcdserver/api" "go.etcd.io/etcd/server/v3/etcdserver/api/etcdhttp/types" "go.etcd.io/etcd/server/v3/etcdserver/api/membership" "go.etcd.io/etcd/server/v3/etcdserver/api/rafthttp" "go.etcd.io/etcd/server/v3/etcdserver/api/snap" stats "go.etcd.io/etcd/server/v3/etcdserver/api/v2stats" "go.etcd.io/etcd/server/v3/etcdserver/api/v2store" "go.etcd.io/etcd/server/v3/etcdserver/api/v3alarm" "go.etcd.io/etcd/server/v3/etcdserver/api/v3compactor" "go.etcd.io/etcd/server/v3/etcdserver/cindex" serverversion "go.etcd.io/etcd/server/v3/etcdserver/version" "go.etcd.io/etcd/server/v3/lease" "go.etcd.io/etcd/server/v3/lease/leasehttp" serverstorage "go.etcd.io/etcd/server/v3/storage" "go.etcd.io/etcd/server/v3/storage/backend" "go.etcd.io/etcd/server/v3/storage/mvcc" "go.etcd.io/etcd/server/v3/storage/schema" ) const ( DefaultSnapshotCount = 100000 // DefaultSnapshotCatchUpEntries is the number of entries for a slow follower // to catch-up after compacting the raft storage entries. // We expect the follower has a millisecond level latency with the leader. // The max throughput is around 10K. Keep a 5K entries is enough for helping // follower to catch up. DefaultSnapshotCatchUpEntries uint64 = 5000 StoreClusterPrefix = "/0" StoreKeysPrefix = "/1" // HealthInterval is the minimum time the cluster should be healthy // before accepting add member requests. HealthInterval = 5 * time.Second purgeFileInterval = 30 * time.Second // max number of in-flight snapshot messages etcdserver allows to have // This number is more than enough for most clusters with 5 machines. maxInFlightMsgSnap = 16 releaseDelayAfterSnapshot = 30 * time.Second // maxPendingRevokes is the maximum number of outstanding expired lease revocations. maxPendingRevokes = 16 recommendedMaxRequestBytes = 10 * 1024 * 1024 readyPercent = 0.9 DowngradeEnabledPath = "/downgrade/enabled" ) var ( // monitorVersionInterval should be smaller than the timeout // on the connection. Or we will not be able to reuse the connection // (since it will timeout). monitorVersionInterval = rafthttp.ConnWriteTimeout - time.Second recommendedMaxRequestBytesString = humanize.Bytes(uint64(recommendedMaxRequestBytes)) storeMemberAttributeRegexp = regexp.MustCompile(path.Join(membership.StoreMembersPrefix, "[[:xdigit:]]{1,16}", "attributes")) ) func init() { rand.Seed(time.Now().UnixNano()) expvar.Publish( "file_descriptor_limit", expvar.Func( func() interface{} { n, _ := runtime.FDLimit() return n }, ), ) } type Response struct { Term uint64 Index uint64 Event *v2store.Event Watcher v2store.Watcher Err error } type ServerV2 interface { Server Leader() types.ID // Do takes a V2 request and attempts to fulfill it, returning a Response. Do(ctx context.Context, r pb.Request) (Response, error) ClientCertAuthEnabled() bool } type ServerV3 interface { Server apply.RaftStatusGetter } func (s *EtcdServer) ClientCertAuthEnabled() bool { return s.Cfg.ClientCertAuthEnabled } type Server interface { // AddMember attempts to add a member into the cluster. It will return // ErrIDRemoved if member ID is removed from the cluster, or return // ErrIDExists if member ID exists in the cluster. AddMember(ctx context.Context, memb membership.Member) ([]*membership.Member, error) // RemoveMember attempts to remove a member from the cluster. It will // return ErrIDRemoved if member ID is removed from the cluster, or return // ErrIDNotFound if member ID is not in the cluster. RemoveMember(ctx context.Context, id uint64) ([]*membership.Member, error) // UpdateMember attempts to update an existing member in the cluster. It will // return ErrIDNotFound if the member ID does not exist. UpdateMember(ctx context.Context, updateMemb membership.Member) ([]*membership.Member, error) // PromoteMember attempts to promote a non-voting node to a voting node. It will // return ErrIDNotFound if the member ID does not exist. // return ErrLearnerNotReady if the member are not ready. // return ErrMemberNotLearner if the member is not a learner. PromoteMember(ctx context.Context, id uint64) ([]*membership.Member, error) // ClusterVersion is the cluster-wide minimum major.minor version. // Cluster version is set to the min version that an etcd member is // compatible with when first bootstrap. // // ClusterVersion is nil until the cluster is bootstrapped (has a quorum). // // During a rolling upgrades, the ClusterVersion will be updated // automatically after a sync. (5 second by default) // // The API/raft component can utilize ClusterVersion to determine if // it can accept a client request or a raft RPC. // NOTE: ClusterVersion might be nil when etcd 2.1 works with etcd 2.0 and // the leader is etcd 2.0. etcd 2.0 leader will not update clusterVersion since // this feature is introduced post 2.0. ClusterVersion() *semver.Version // StorageVersion is the storage schema version. It's supported starting // from 3.6. StorageVersion() *semver.Version Cluster() api.Cluster Alarms() []*pb.AlarmMember // LeaderChangedNotify returns a channel for application level code to be notified // when etcd leader changes, this function is intend to be used only in application // which embed etcd. // Caution: // 1. the returned channel is being closed when the leadership changes. // 2. so the new channel needs to be obtained for each raft term. // 3. user can loose some consecutive channel changes using this API. LeaderChangedNotify() <-chan struct{} } // EtcdServer is the production implementation of the Server interface type EtcdServer struct { // inflightSnapshots holds count the number of snapshots currently inflight. inflightSnapshots int64 // must use atomic operations to access; keep 64-bit aligned. appliedIndex uint64 // must use atomic operations to access; keep 64-bit aligned. committedIndex uint64 // must use atomic operations to access; keep 64-bit aligned. term uint64 // must use atomic operations to access; keep 64-bit aligned. lead uint64 // must use atomic operations to access; keep 64-bit aligned. consistIndex cindex.ConsistentIndexer // consistIndex is used to get/set/save consistentIndex r raftNode // uses 64-bit atomics; keep 64-bit aligned. readych chan struct{} Cfg config.ServerConfig lgMu *sync.RWMutex lg *zap.Logger w wait.Wait readMu sync.RWMutex // read routine notifies etcd server that it waits for reading by sending an empty struct to // readwaitC readwaitc chan struct{} // readNotifier is used to notify the read routine that it can process the request // when there is no error readNotifier *notifier // stop signals the run goroutine should shutdown. stop chan struct{} // stopping is closed by run goroutine on shutdown. stopping chan struct{} // done is closed when all goroutines from start() complete. done chan struct{} // leaderChanged is used to notify the linearizable read loop to drop the old read requests. leaderChanged *notify.Notifier errorc chan error memberId types.ID attributes membership.Attributes cluster *membership.RaftCluster v2store v2store.Store snapshotter *snap.Snapshotter applyV2 ApplierV2 uberApply apply.UberApplier applyWait wait.WaitTime kv mvcc.WatchableKV lessor lease.Lessor bemu sync.RWMutex be backend.Backend beHooks *serverstorage.BackendHooks authStore auth.AuthStore alarmStore *v3alarm.AlarmStore stats *stats.ServerStats lstats *stats.LeaderStats SyncTicker *time.Ticker // compactor is used to auto-compact the KV. compactor v3compactor.Compactor // peerRt used to send requests (version, lease) to peers. peerRt http.RoundTripper reqIDGen *idutil.Generator // wgMu blocks concurrent waitgroup mutation while server stopping wgMu sync.RWMutex // wg is used to wait for the goroutines that depends on the server state // to exit when stopping the server. wg sync.WaitGroup // ctx is used for etcd-initiated requests that may need to be canceled // on etcd server shutdown. ctx context.Context cancel context.CancelFunc leadTimeMu sync.RWMutex leadElectedTime time.Time firstCommitInTerm *notify.Notifier clusterVersionChanged *notify.Notifier *AccessController // forceSnapshot can force snapshot be triggered after apply, independent of the snapshotCount. // Should only be set within apply code path. Used to force snapshot after cluster version downgrade. forceSnapshot bool } // NewServer creates a new EtcdServer from the supplied configuration. The // configuration is considered static for the lifetime of the EtcdServer. func NewServer(cfg config.ServerConfig) (srv *EtcdServer, err error) { b, err := bootstrap(cfg) if err != nil { return nil, err } defer func() { if err != nil { b.Close() } }() sstats := stats.NewServerStats(cfg.Name, b.cluster.cl.String()) lstats := stats.NewLeaderStats(cfg.Logger, b.cluster.nodeID.String()) heartbeat := time.Duration(cfg.TickMs) * time.Millisecond srv = &EtcdServer{ readych: make(chan struct{}), Cfg: cfg, lgMu: new(sync.RWMutex), lg: cfg.Logger, errorc: make(chan error, 1), v2store: b.storage.st, snapshotter: b.ss, r: *b.raft.newRaftNode(b.ss, b.storage.wal.w, b.cluster.cl), memberId: b.cluster.nodeID, attributes: membership.Attributes{Name: cfg.Name, ClientURLs: cfg.ClientURLs.StringSlice()}, cluster: b.cluster.cl, stats: sstats, lstats: lstats, SyncTicker: time.NewTicker(500 * time.Millisecond), peerRt: b.prt, reqIDGen: idutil.NewGenerator(uint16(b.cluster.nodeID), time.Now()), AccessController: &AccessController{CORS: cfg.CORS, HostWhitelist: cfg.HostWhitelist}, consistIndex: b.storage.backend.ci, firstCommitInTerm: notify.NewNotifier(), clusterVersionChanged: notify.NewNotifier(), } serverID.With(prometheus.Labels{"server_id": b.cluster.nodeID.String()}).Set(1) srv.cluster.SetVersionChangedNotifier(srv.clusterVersionChanged) srv.applyV2 = NewApplierV2(cfg.Logger, srv.v2store, srv.cluster) srv.be = b.storage.backend.be srv.beHooks = b.storage.backend.beHooks minTTL := time.Duration((3*cfg.ElectionTicks)/2) * heartbeat // always recover lessor before kv. When we recover the mvcc.KV it will reattach keys to its leases. // If we recover mvcc.KV first, it will attach the keys to the wrong lessor before it recovers. srv.lessor = lease.NewLessor(srv.Logger(), srv.be, srv.cluster, lease.LessorConfig{ MinLeaseTTL: int64(math.Ceil(minTTL.Seconds())), CheckpointInterval: cfg.LeaseCheckpointInterval, CheckpointPersist: cfg.LeaseCheckpointPersist, ExpiredLeasesRetryInterval: srv.Cfg.ReqTimeout(), }) tp, err := auth.NewTokenProvider(cfg.Logger, cfg.AuthToken, func(index uint64) <-chan struct{} { return srv.applyWait.Wait(index) }, time.Duration(cfg.TokenTTL)*time.Second, ) if err != nil { cfg.Logger.Warn("failed to create token provider", zap.Error(err)) return nil, err } mvccStoreConfig := mvcc.StoreConfig{ CompactionBatchLimit: cfg.CompactionBatchLimit, CompactionSleepInterval: cfg.CompactionSleepInterval, } srv.kv = mvcc.New(srv.Logger(), srv.be, srv.lessor, mvccStoreConfig) srv.authStore = auth.NewAuthStore(srv.Logger(), schema.NewAuthBackend(srv.Logger(), srv.be), tp, int(cfg.BcryptCost)) newSrv := srv // since srv == nil in defer if srv is returned as nil defer func() { // closing backend without first closing kv can cause // resumed compactions to fail with closed tx errors if err != nil { newSrv.kv.Close() } }() if num := cfg.AutoCompactionRetention; num != 0 { srv.compactor, err = v3compactor.New(cfg.Logger, cfg.AutoCompactionMode, num, srv.kv, srv) if err != nil { return nil, err } srv.compactor.Run() } if err = srv.restoreAlarms(); err != nil { return nil, err } srv.uberApply = srv.NewUberApplier() if srv.Cfg.EnableLeaseCheckpoint { // setting checkpointer enables lease checkpoint feature. srv.lessor.SetCheckpointer(func(ctx context.Context, cp *pb.LeaseCheckpointRequest) { srv.raftRequestOnce(ctx, pb.InternalRaftRequest{LeaseCheckpoint: cp}) }) } // Set the hook after EtcdServer finishes the initialization to avoid // the hook being called during the initialization process. srv.be.SetTxPostLockInsideApplyHook(srv.getTxPostLockInsideApplyHook()) // TODO: move transport initialization near the definition of remote tr := &rafthttp.Transport{ Logger: cfg.Logger, TLSInfo: cfg.PeerTLSInfo, DialTimeout: cfg.PeerDialTimeout(), ID: b.cluster.nodeID, URLs: cfg.PeerURLs, ClusterID: b.cluster.cl.ID(), Raft: srv, Snapshotter: b.ss, ServerStats: sstats, LeaderStats: lstats, ErrorC: srv.errorc, } if err = tr.Start(); err != nil { return nil, err } // add all remotes into transport for _, m := range b.cluster.remotes { if m.ID != b.cluster.nodeID { tr.AddRemote(m.ID, m.PeerURLs) } } for _, m := range b.cluster.cl.Members() { if m.ID != b.cluster.nodeID { tr.AddPeer(m.ID, m.PeerURLs) } } srv.r.transport = tr return srv, nil } func (s *EtcdServer) Logger() *zap.Logger { s.lgMu.RLock() l := s.lg s.lgMu.RUnlock() return l } func (s *EtcdServer) Config() config.ServerConfig { return s.Cfg } func tickToDur(ticks int, tickMs uint) string { return fmt.Sprintf("%v", time.Duration(ticks)*time.Duration(tickMs)*time.Millisecond) } func (s *EtcdServer) adjustTicks() { lg := s.Logger() clusterN := len(s.cluster.Members()) // single-node fresh start, or single-node recovers from snapshot if clusterN == 1 { ticks := s.Cfg.ElectionTicks - 1 lg.Info( "started as single-node; fast-forwarding election ticks", zap.String("local-member-id", s.MemberId().String()), zap.Int("forward-ticks", ticks), zap.String("forward-duration", tickToDur(ticks, s.Cfg.TickMs)), zap.Int("election-ticks", s.Cfg.ElectionTicks), zap.String("election-timeout", tickToDur(s.Cfg.ElectionTicks, s.Cfg.TickMs)), ) s.r.advanceTicks(ticks) return } if !s.Cfg.InitialElectionTickAdvance { lg.Info("skipping initial election tick advance", zap.Int("election-ticks", s.Cfg.ElectionTicks)) return } lg.Info("starting initial election tick advance", zap.Int("election-ticks", s.Cfg.ElectionTicks)) // retry up to "rafthttp.ConnReadTimeout", which is 5-sec // until peer connection reports; otherwise: // 1. all connections failed, or // 2. no active peers, or // 3. restarted single-node with no snapshot // then, do nothing, because advancing ticks would have no effect waitTime := rafthttp.ConnReadTimeout itv := 50 * time.Millisecond for i := int64(0); i < int64(waitTime/itv); i++ { select { case <-time.After(itv): case <-s.stopping: return } peerN := s.r.transport.ActivePeers() if peerN > 1 { // multi-node received peer connection reports // adjust ticks, in case slow leader message receive ticks := s.Cfg.ElectionTicks - 2 lg.Info( "initialized peer connections; fast-forwarding election ticks", zap.String("local-member-id", s.MemberId().String()), zap.Int("forward-ticks", ticks), zap.String("forward-duration", tickToDur(ticks, s.Cfg.TickMs)), zap.Int("election-ticks", s.Cfg.ElectionTicks), zap.String("election-timeout", tickToDur(s.Cfg.ElectionTicks, s.Cfg.TickMs)), zap.Int("active-remote-members", peerN), ) s.r.advanceTicks(ticks) return } } } // Start performs any initialization of the Server necessary for it to // begin serving requests. It must be called before Do or Process. // Start must be non-blocking; any long-running server functionality // should be implemented in goroutines. func (s *EtcdServer) Start() { s.start() s.GoAttach(func() { s.adjustTicks() }) s.GoAttach(func() { s.publishV3(s.Cfg.ReqTimeout()) }) s.GoAttach(s.purgeFile) s.GoAttach(func() { monitorFileDescriptor(s.Logger(), s.stopping) }) s.GoAttach(s.monitorClusterVersions) s.GoAttach(s.monitorStorageVersion) s.GoAttach(s.linearizableReadLoop) s.GoAttach(s.monitorKVHash) s.GoAttach(s.monitorDowngrade) } // start prepares and starts server in a new goroutine. It is no longer safe to // modify a server's fields after it has been sent to Start. // This function is just used for testing. func (s *EtcdServer) start() { lg := s.Logger() if s.Cfg.SnapshotCount == 0 { lg.Info( "updating snapshot-count to default", zap.Uint64("given-snapshot-count", s.Cfg.SnapshotCount), zap.Uint64("updated-snapshot-count", DefaultSnapshotCount), ) s.Cfg.SnapshotCount = DefaultSnapshotCount } if s.Cfg.SnapshotCatchUpEntries == 0 { lg.Info( "updating snapshot catch-up entries to default", zap.Uint64("given-snapshot-catchup-entries", s.Cfg.SnapshotCatchUpEntries), zap.Uint64("updated-snapshot-catchup-entries", DefaultSnapshotCatchUpEntries), ) s.Cfg.SnapshotCatchUpEntries = DefaultSnapshotCatchUpEntries } s.w = wait.New() s.applyWait = wait.NewTimeList() s.done = make(chan struct{}) s.stop = make(chan struct{}) s.stopping = make(chan struct{}, 1) s.ctx, s.cancel = context.WithCancel(context.Background()) s.readwaitc = make(chan struct{}, 1) s.readNotifier = newNotifier() s.leaderChanged = notify.NewNotifier() if s.ClusterVersion() != nil { lg.Info( "starting etcd server", zap.String("local-member-id", s.MemberId().String()), zap.String("local-server-version", version.Version), zap.String("cluster-id", s.Cluster().ID().String()), zap.String("cluster-version", version.Cluster(s.ClusterVersion().String())), ) membership.ClusterVersionMetrics.With(prometheus.Labels{"cluster_version": version.Cluster(s.ClusterVersion().String())}).Set(1) } else { lg.Info( "starting etcd server", zap.String("local-member-id", s.MemberId().String()), zap.String("local-server-version", version.Version), zap.String("cluster-version", "to_be_decided"), ) } // TODO: if this is an empty log, writes all peer infos // into the first entry go s.run() } func (s *EtcdServer) purgeFile() { lg := s.Logger() var dberrc, serrc, werrc <-chan error var dbdonec, sdonec, wdonec <-chan struct{} if s.Cfg.MaxSnapFiles > 0 { dbdonec, dberrc = fileutil.PurgeFileWithDoneNotify(lg, s.Cfg.SnapDir(), "snap.db", s.Cfg.MaxSnapFiles, purgeFileInterval, s.stopping) sdonec, serrc = fileutil.PurgeFileWithDoneNotify(lg, s.Cfg.SnapDir(), "snap", s.Cfg.MaxSnapFiles, purgeFileInterval, s.stopping) } if s.Cfg.MaxWALFiles > 0 { wdonec, werrc = fileutil.PurgeFileWithDoneNotify(lg, s.Cfg.WALDir(), "wal", s.Cfg.MaxWALFiles, purgeFileInterval, s.stopping) } select { case e := <-dberrc: lg.Fatal("failed to purge snap db file", zap.Error(e)) case e := <-serrc: lg.Fatal("failed to purge snap file", zap.Error(e)) case e := <-werrc: lg.Fatal("failed to purge wal file", zap.Error(e)) case <-s.stopping: if dbdonec != nil { <-dbdonec } if sdonec != nil { <-sdonec } if wdonec != nil { <-wdonec } return } } func (s *EtcdServer) Cluster() api.Cluster { return s.cluster } func (s *EtcdServer) ApplyWait() <-chan struct{} { return s.applyWait.Wait(s.getCommittedIndex()) } type ServerPeer interface { ServerV2 RaftHandler() http.Handler LeaseHandler() http.Handler } func (s *EtcdServer) LeaseHandler() http.Handler { if s.lessor == nil { return nil } return leasehttp.NewHandler(s.lessor, s.ApplyWait) } func (s *EtcdServer) RaftHandler() http.Handler { return s.r.transport.Handler() } type ServerPeerV2 interface { ServerPeer HashKVHandler() http.Handler DowngradeEnabledHandler() http.Handler } func (s *EtcdServer) DowngradeInfo() *serverversion.DowngradeInfo { return s.cluster.DowngradeInfo() } type downgradeEnabledHandler struct { lg *zap.Logger cluster api.Cluster server *EtcdServer } func (s *EtcdServer) DowngradeEnabledHandler() http.Handler { return &downgradeEnabledHandler{ lg: s.Logger(), cluster: s.cluster, server: s, } } func (h *downgradeEnabledHandler) ServeHTTP(w http.ResponseWriter, r *http.Request) { if r.Method != http.MethodGet { w.Header().Set("Allow", http.MethodGet) http.Error(w, "Method Not Allowed", http.StatusMethodNotAllowed) return } w.Header().Set("X-Etcd-Cluster-ID", h.cluster.ID().String()) if r.URL.Path != DowngradeEnabledPath { http.Error(w, "bad path", http.StatusBadRequest) return } ctx, cancel := context.WithTimeout(context.Background(), h.server.Cfg.ReqTimeout()) defer cancel() // serve with linearized downgrade info if err := h.server.linearizableReadNotify(ctx); err != nil { http.Error(w, fmt.Sprintf("failed linearized read: %v", err), http.StatusInternalServerError) return } enabled := h.server.DowngradeInfo().Enabled w.Header().Set("Content-Type", "text/plain") w.Write([]byte(strconv.FormatBool(enabled))) } // Process takes a raft message and applies it to the server's raft state // machine, respecting any timeout of the given context. func (s *EtcdServer) Process(ctx context.Context, m raftpb.Message) error { lg := s.Logger() if s.cluster.IsIDRemoved(types.ID(m.From)) { lg.Warn( "rejected Raft message from removed member", zap.String("local-member-id", s.MemberId().String()), zap.String("removed-member-id", types.ID(m.From).String()), ) return httptypes.NewHTTPError(http.StatusForbidden, "cannot process message from removed member") } if m.Type == raftpb.MsgApp { s.stats.RecvAppendReq(types.ID(m.From).String(), m.Size()) } return s.r.Step(ctx, m) } func (s *EtcdServer) IsIDRemoved(id uint64) bool { return s.cluster.IsIDRemoved(types.ID(id)) } func (s *EtcdServer) ReportUnreachable(id uint64) { s.r.ReportUnreachable(id) } // ReportSnapshot reports snapshot sent status to the raft state machine, // and clears the used snapshot from the snapshot store. func (s *EtcdServer) ReportSnapshot(id uint64, status raft.SnapshotStatus) { s.r.ReportSnapshot(id, status) } type etcdProgress struct { confState raftpb.ConfState snapi uint64 appliedt uint64 appliedi uint64 } // raftReadyHandler contains a set of EtcdServer operations to be called by raftNode, // and helps decouple state machine logic from Raft algorithms. // TODO: add a state machine interface to toApply the commit entries and do snapshot/recover type raftReadyHandler struct { getLead func() (lead uint64) updateLead func(lead uint64) updateLeadership func(newLeader bool) updateCommittedIndex func(uint64) } func (s *EtcdServer) run() { lg := s.Logger() sn, err := s.r.raftStorage.Snapshot() if err != nil { lg.Panic("failed to get snapshot from Raft storage", zap.Error(err)) } // asynchronously accept toApply packets, dispatch progress in-order sched := schedule.NewFIFOScheduler() var ( smu sync.RWMutex syncC <-chan time.Time ) setSyncC := func(ch <-chan time.Time) { smu.Lock() syncC = ch smu.Unlock() } getSyncC := func() (ch <-chan time.Time) { smu.RLock() ch = syncC smu.RUnlock() return } rh := &raftReadyHandler{ getLead: func() (lead uint64) { return s.getLead() }, updateLead: func(lead uint64) { s.setLead(lead) }, updateLeadership: func(newLeader bool) { if !s.isLeader() { if s.lessor != nil { s.lessor.Demote() } if s.compactor != nil { s.compactor.Pause() } setSyncC(nil) } else { if newLeader { t := time.Now() s.leadTimeMu.Lock() s.leadElectedTime = t s.leadTimeMu.Unlock() } setSyncC(s.SyncTicker.C) if s.compactor != nil { s.compactor.Resume() } } if newLeader { s.leaderChanged.Notify() } // TODO: remove the nil checking // current test utility does not provide the stats if s.stats != nil { s.stats.BecomeLeader() } }, updateCommittedIndex: func(ci uint64) { cci := s.getCommittedIndex() if ci > cci { s.setCommittedIndex(ci) } }, } s.r.start(rh) ep := etcdProgress{ confState: sn.Metadata.ConfState, snapi: sn.Metadata.Index, appliedt: sn.Metadata.Term, appliedi: sn.Metadata.Index, } defer func() { s.wgMu.Lock() // block concurrent waitgroup adds in GoAttach while stopping close(s.stopping) s.wgMu.Unlock() s.cancel() sched.Stop() // wait for gouroutines before closing raft so wal stays open s.wg.Wait() s.SyncTicker.Stop() // must stop raft after scheduler-- etcdserver can leak rafthttp pipelines // by adding a peer after raft stops the transport s.r.stop() s.Cleanup() close(s.done) }() var expiredLeaseC <-chan []*lease.Lease if s.lessor != nil { expiredLeaseC = s.lessor.ExpiredLeasesC() } for { select { case ap := <-s.r.apply(): f := func(context.Context) { s.applyAll(&ep, &ap) } sched.Schedule(f) case leases := <-expiredLeaseC: s.revokeExpiredLeases(leases) case err := <-s.errorc: lg.Warn("server error", zap.Error(err)) lg.Warn("data-dir used by this member must be removed") return case <-getSyncC(): if s.v2store.HasTTLKeys() { s.sync(s.Cfg.ReqTimeout()) } case <-s.stop: return } } } func (s *EtcdServer) revokeExpiredLeases(leases []*lease.Lease) { s.GoAttach(func() { lg := s.Logger() // Increases throughput of expired leases deletion process through parallelization c := make(chan struct{}, maxPendingRevokes) for _, curLease := range leases { select { case c <- struct{}{}: case <-s.stopping: return } f := func(lid int64) { s.GoAttach(func() { ctx := s.authStore.WithRoot(s.ctx) _, lerr := s.LeaseRevoke(ctx, &pb.LeaseRevokeRequest{ID: lid}) if lerr == nil { leaseExpired.Inc() } else { lg.Warn( "failed to revoke lease", zap.String("lease-id", fmt.Sprintf("%016x", lid)), zap.Error(lerr), ) } <-c }) } f(int64(curLease.ID)) } }) } // Cleanup removes allocated objects by EtcdServer.NewServer in // situation that EtcdServer::Start was not called (that takes care of cleanup). func (s *EtcdServer) Cleanup() { // kv, lessor and backend can be nil if running without v3 enabled // or running unit tests. if s.lessor != nil { s.lessor.Stop() } if s.kv != nil { s.kv.Close() } if s.authStore != nil { s.authStore.Close() } if s.be != nil { s.be.Close() } if s.compactor != nil { s.compactor.Stop() } } func (s *EtcdServer) applyAll(ep *etcdProgress, apply *toApply) { s.applySnapshot(ep, apply) s.applyEntries(ep, apply) proposalsApplied.Set(float64(ep.appliedi)) s.applyWait.Trigger(ep.appliedi) // wait for the raft routine to finish the disk writes before triggering a // snapshot. or applied index might be greater than the last index in raft // storage, since the raft routine might be slower than toApply routine. <-apply.notifyc s.triggerSnapshot(ep) select { // snapshot requested via send() case m := <-s.r.msgSnapC: merged := s.createMergedSnapshotMessage(m, ep.appliedt, ep.appliedi, ep.confState) s.sendMergedSnap(merged) default: } } func (s *EtcdServer) applySnapshot(ep *etcdProgress, toApply *toApply) { if raft.IsEmptySnap(toApply.snapshot) { return } applySnapshotInProgress.Inc() lg := s.Logger() lg.Info( "applying snapshot", zap.Uint64("current-snapshot-index", ep.snapi), zap.Uint64("current-applied-index", ep.appliedi), zap.Uint64("incoming-leader-snapshot-index", toApply.snapshot.Metadata.Index), zap.Uint64("incoming-leader-snapshot-term", toApply.snapshot.Metadata.Term), ) defer func() { lg.Info( "applied snapshot", zap.Uint64("current-snapshot-index", ep.snapi), zap.Uint64("current-applied-index", ep.appliedi), zap.Uint64("incoming-leader-snapshot-index", toApply.snapshot.Metadata.Index), zap.Uint64("incoming-leader-snapshot-term", toApply.snapshot.Metadata.Term), ) applySnapshotInProgress.Dec() }() if toApply.snapshot.Metadata.Index <= ep.appliedi { lg.Panic( "unexpected leader snapshot from outdated index", zap.Uint64("current-snapshot-index", ep.snapi), zap.Uint64("current-applied-index", ep.appliedi), zap.Uint64("incoming-leader-snapshot-index", toApply.snapshot.Metadata.Index), zap.Uint64("incoming-leader-snapshot-term", toApply.snapshot.Metadata.Term), ) } // wait for raftNode to persist snapshot onto the disk <-toApply.notifyc newbe, err := serverstorage.OpenSnapshotBackend(s.Cfg, s.snapshotter, toApply.snapshot, s.beHooks) if err != nil { lg.Panic("failed to open snapshot backend", zap.Error(err)) } // We need to set the backend to consistIndex before recovering the lessor, // because lessor.Recover will commit the boltDB transaction, accordingly it // will get the old consistent_index persisted into the db in OnPreCommitUnsafe. // Eventually the new consistent_index value coming from snapshot is overwritten // by the old value. s.consistIndex.SetBackend(newbe) verifySnapshotIndex(toApply.snapshot, s.consistIndex.ConsistentIndex()) // always recover lessor before kv. When we recover the mvcc.KV it will reattach keys to its leases. // If we recover mvcc.KV first, it will attach the keys to the wrong lessor before it recovers. if s.lessor != nil { lg.Info("restoring lease store") s.lessor.Recover(newbe, func() lease.TxnDelete { return s.kv.Write(traceutil.TODO()) }) lg.Info("restored lease store") } lg.Info("restoring mvcc store") if err := s.kv.Restore(newbe); err != nil { lg.Panic("failed to restore mvcc store", zap.Error(err)) } newbe.SetTxPostLockInsideApplyHook(s.getTxPostLockInsideApplyHook()) lg.Info("restored mvcc store", zap.Uint64("consistent-index", s.consistIndex.ConsistentIndex())) // Closing old backend might block until all the txns // on the backend are finished. // We do not want to wait on closing the old backend. s.bemu.Lock() oldbe := s.be go func() { lg.Info("closing old backend file") defer func() { lg.Info("closed old backend file") }() if err := oldbe.Close(); err != nil { lg.Panic("failed to close old backend", zap.Error(err)) } }() s.be = newbe s.bemu.Unlock() lg.Info("restoring alarm store") if err := s.restoreAlarms(); err != nil { lg.Panic("failed to restore alarm store", zap.Error(err)) } lg.Info("restored alarm store") if s.authStore != nil { lg.Info("restoring auth store") s.authStore.Recover(schema.NewAuthBackend(lg, newbe)) lg.Info("restored auth store") } lg.Info("restoring v2 store") if err := s.v2store.Recovery(toApply.snapshot.Data); err != nil { lg.Panic("failed to restore v2 store", zap.Error(err)) } if err := serverstorage.AssertNoV2StoreContent(lg, s.v2store, s.Cfg.V2Deprecation); err != nil { lg.Panic("illegal v2store content", zap.Error(err)) } lg.Info("restored v2 store") s.cluster.SetBackend(schema.NewMembershipBackend(lg, newbe)) lg.Info("restoring cluster configuration") s.cluster.Recover(api.UpdateCapability) lg.Info("restored cluster configuration") lg.Info("removing old peers from network") // recover raft transport s.r.transport.RemoveAllPeers() lg.Info("removed old peers from network") lg.Info("adding peers from new cluster configuration") for _, m := range s.cluster.Members() { if m.ID == s.MemberId() { continue } s.r.transport.AddPeer(m.ID, m.PeerURLs) } lg.Info("added peers from new cluster configuration") ep.appliedt = toApply.snapshot.Metadata.Term ep.appliedi = toApply.snapshot.Metadata.Index ep.snapi = ep.appliedi ep.confState = toApply.snapshot.Metadata.ConfState // As backends and implementations like alarmsStore changed, we need // to re-bootstrap Appliers. s.uberApply = s.NewUberApplier() } func (s *EtcdServer) NewUberApplier() apply.UberApplier { return apply.NewUberApplier(s.lg, s.be, s.KV(), s.alarmStore, s.authStore, s.lessor, s.cluster, s, s, s.consistIndex, s.Cfg.WarningApplyDuration, s.Cfg.ExperimentalTxnModeWriteWithSharedBuffer, s.Cfg.QuotaBackendBytes) } func verifySnapshotIndex(snapshot raftpb.Snapshot, cindex uint64) { verify.Verify(func() { if cindex != snapshot.Metadata.Index { panic(fmt.Sprintf("consistent_index(%d) isn't equal to snapshot index (%d)", cindex, snapshot.Metadata.Index)) } }) } func (s *EtcdServer) applyEntries(ep *etcdProgress, apply *toApply) { if len(apply.entries) == 0 { return } firsti := apply.entries[0].Index if firsti > ep.appliedi+1 { lg := s.Logger() lg.Panic( "unexpected committed entry index", zap.Uint64("current-applied-index", ep.appliedi), zap.Uint64("first-committed-entry-index", firsti), ) } var ents []raftpb.Entry if ep.appliedi+1-firsti < uint64(len(apply.entries)) { ents = apply.entries[ep.appliedi+1-firsti:] } if len(ents) == 0 { return } var shouldstop bool if ep.appliedt, ep.appliedi, shouldstop = s.apply(ents, &ep.confState); shouldstop { go s.stopWithDelay(10*100*time.Millisecond, fmt.Errorf("the member has been permanently removed from the cluster")) } } func (s *EtcdServer) ForceSnapshot() { s.forceSnapshot = true } func (s *EtcdServer) triggerSnapshot(ep *etcdProgress) { if !s.shouldSnapshot(ep) { return } lg := s.Logger() lg.Info( "triggering snapshot", zap.String("local-member-id", s.MemberId().String()), zap.Uint64("local-member-applied-index", ep.appliedi), zap.Uint64("local-member-snapshot-index", ep.snapi), zap.Uint64("local-member-snapshot-count", s.Cfg.SnapshotCount), zap.Bool("snapshot-forced", s.forceSnapshot), ) s.forceSnapshot = false s.snapshot(ep.appliedi, ep.confState) ep.snapi = ep.appliedi } func (s *EtcdServer) shouldSnapshot(ep *etcdProgress) bool { return (s.forceSnapshot && ep.appliedi != ep.snapi) || (ep.appliedi-ep.snapi > s.Cfg.SnapshotCount) } func (s *EtcdServer) hasMultipleVotingMembers() bool { return s.cluster != nil && len(s.cluster.VotingMemberIDs()) > 1 } func (s *EtcdServer) isLeader() bool { return uint64(s.MemberId()) == s.Lead() } // MoveLeader transfers the leader to the given transferee. func (s *EtcdServer) MoveLeader(ctx context.Context, lead, transferee uint64) error { if !s.cluster.IsMemberExist(types.ID(transferee)) || s.cluster.Member(types.ID(transferee)).IsLearner { return errors.ErrBadLeaderTransferee } now := time.Now() interval := time.Duration(s.Cfg.TickMs) * time.Millisecond lg := s.Logger() lg.Info( "leadership transfer starting", zap.String("local-member-id", s.MemberId().String()), zap.String("current-leader-member-id", types.ID(lead).String()), zap.String("transferee-member-id", types.ID(transferee).String()), ) s.r.TransferLeadership(ctx, lead, transferee) for s.Lead() != transferee { select { case <-ctx.Done(): // time out return errors.ErrTimeoutLeaderTransfer case <-time.After(interval): } } // TODO: drain all requests, or drop all messages to the old leader lg.Info( "leadership transfer finished", zap.String("local-member-id", s.MemberId().String()), zap.String("old-leader-member-id", types.ID(lead).String()), zap.String("new-leader-member-id", types.ID(transferee).String()), zap.Duration("took", time.Since(now)), ) return nil } // TransferLeadership transfers the leader to the chosen transferee. func (s *EtcdServer) TransferLeadership() error { lg := s.Logger() if !s.isLeader() { lg.Info( "skipped leadership transfer; local server is not leader", zap.String("local-member-id", s.MemberId().String()), zap.String("current-leader-member-id", types.ID(s.Lead()).String()), ) return nil } if !s.hasMultipleVotingMembers() { lg.Info( "skipped leadership transfer for single voting member cluster", zap.String("local-member-id", s.MemberId().String()), zap.String("current-leader-member-id", types.ID(s.Lead()).String()), ) return nil } transferee, ok := longestConnected(s.r.transport, s.cluster.VotingMemberIDs()) if !ok { return errors.ErrUnhealthy } tm := s.Cfg.ReqTimeout() ctx, cancel := context.WithTimeout(s.ctx, tm) err := s.MoveLeader(ctx, s.Lead(), uint64(transferee)) cancel() return err } // HardStop stops the server without coordination with other members in the cluster. func (s *EtcdServer) HardStop() { select { case s.stop <- struct{}{}: case <-s.done: return } <-s.done } // Stop stops the server gracefully, and shuts down the running goroutine. // Stop should be called after a Start(s), otherwise it will block forever. // When stopping leader, Stop transfers its leadership to one of its peers // before stopping the server. // Stop terminates the Server and performs any necessary finalization. // Do and Process cannot be called after Stop has been invoked. func (s *EtcdServer) Stop() { lg := s.Logger() if err := s.TransferLeadership(); err != nil { lg.Warn("leadership transfer failed", zap.String("local-member-id", s.MemberId().String()), zap.Error(err)) } s.HardStop() } // ReadyNotify returns a channel that will be closed when the server // is ready to serve client requests func (s *EtcdServer) ReadyNotify() <-chan struct{} { return s.readych } func (s *EtcdServer) stopWithDelay(d time.Duration, err error) { select { case <-time.After(d): case <-s.done: } select { case s.errorc <- err: default: } } // StopNotify returns a channel that receives a empty struct // when the server is stopped. func (s *EtcdServer) StopNotify() <-chan struct{} { return s.done } // StoppingNotify returns a channel that receives a empty struct // when the server is being stopped. func (s *EtcdServer) StoppingNotify() <-chan struct{} { return s.stopping } func (s *EtcdServer) checkMembershipOperationPermission(ctx context.Context) error { if s.authStore == nil { // In the context of ordinary etcd process, s.authStore will never be nil. // This branch is for handling cases in server_test.go return nil } // Note that this permission check is done in the API layer, // so TOCTOU problem can be caused potentially in a schedule like this: // update membership with user A -> revoke root role of A -> toApply membership change // in the state machine layer // However, both of membership change and role management requires the root privilege. // So careful operation by admins can prevent the problem. authInfo, err := s.AuthInfoFromCtx(ctx) if err != nil { return err } return s.AuthStore().IsAdminPermitted(authInfo) } func (s *EtcdServer) AddMember(ctx context.Context, memb membership.Member) ([]*membership.Member, error) { if err := s.checkMembershipOperationPermission(ctx); err != nil { return nil, err } // TODO: move Member to protobuf type b, err := json.Marshal(memb) if err != nil { return nil, err } // by default StrictReconfigCheck is enabled; reject new members if unhealthy. if err := s.mayAddMember(memb); err != nil { return nil, err } cc := raftpb.ConfChange{ Type: raftpb.ConfChangeAddNode, NodeID: uint64(memb.ID), Context: b, } if memb.IsLearner { cc.Type = raftpb.ConfChangeAddLearnerNode } return s.configure(ctx, cc) } func (s *EtcdServer) mayAddMember(memb membership.Member) error { lg := s.Logger() if !s.Cfg.StrictReconfigCheck { return nil } // protect quorum when adding voting member if !memb.IsLearner && !s.cluster.IsReadyToAddVotingMember() { lg.Warn( "rejecting member add request; not enough healthy members", zap.String("local-member-id", s.MemberId().String()), zap.String("requested-member-add", fmt.Sprintf("%+v", memb)), zap.Error(errors.ErrNotEnoughStartedMembers), ) return errors.ErrNotEnoughStartedMembers } if !isConnectedFullySince(s.r.transport, time.Now().Add(-HealthInterval), s.MemberId(), s.cluster.VotingMembers()) { lg.Warn( "rejecting member add request; local member has not been connected to all peers, reconfigure breaks active quorum", zap.String("local-member-id", s.MemberId().String()), zap.String("requested-member-add", fmt.Sprintf("%+v", memb)), zap.Error(errors.ErrUnhealthy), ) return errors.ErrUnhealthy } return nil } func (s *EtcdServer) RemoveMember(ctx context.Context, id uint64) ([]*membership.Member, error) { if err := s.checkMembershipOperationPermission(ctx); err != nil { return nil, err } // by default StrictReconfigCheck is enabled; reject removal if leads to quorum loss if err := s.mayRemoveMember(types.ID(id)); err != nil { return nil, err } cc := raftpb.ConfChange{ Type: raftpb.ConfChangeRemoveNode, NodeID: id, } return s.configure(ctx, cc) } // PromoteMember promotes a learner node to a voting node. func (s *EtcdServer) PromoteMember(ctx context.Context, id uint64) ([]*membership.Member, error) { // only raft leader has information on whether the to-be-promoted learner node is ready. If promoteMember call // fails with ErrNotLeader, forward the request to leader node via HTTP. If promoteMember call fails with error // other than ErrNotLeader, return the error. resp, err := s.promoteMember(ctx, id) if err == nil { learnerPromoteSucceed.Inc() return resp, nil } if err != errors.ErrNotLeader { learnerPromoteFailed.WithLabelValues(err.Error()).Inc() return resp, err } cctx, cancel := context.WithTimeout(ctx, s.Cfg.ReqTimeout()) defer cancel() // forward to leader for cctx.Err() == nil { leader, err := s.waitLeader(cctx) if err != nil { return nil, err } for _, url := range leader.PeerURLs { resp, err := promoteMemberHTTP(cctx, url, id, s.peerRt) if err == nil { return resp, nil } // If member promotion failed, return early. Otherwise keep retry. if err == errors.ErrLearnerNotReady || err == membership.ErrIDNotFound || err == membership.ErrMemberNotLearner { return nil, err } } } if cctx.Err() == context.DeadlineExceeded { return nil, errors.ErrTimeout } return nil, errors.ErrCanceled } // promoteMember checks whether the to-be-promoted learner node is ready before sending the promote // request to raft. // The function returns ErrNotLeader if the local node is not raft leader (therefore does not have // enough information to determine if the learner node is ready), returns ErrLearnerNotReady if the // local node is leader (therefore has enough information) but decided the learner node is not ready // to be promoted. func (s *EtcdServer) promoteMember(ctx context.Context, id uint64) ([]*membership.Member, error) { if err := s.checkMembershipOperationPermission(ctx); err != nil { return nil, err } // check if we can promote this learner. if err := s.mayPromoteMember(types.ID(id)); err != nil { return nil, err } // build the context for the promote confChange. mark IsLearner to false and IsPromote to true. promoteChangeContext := membership.ConfigChangeContext{ Member: membership.Member{ ID: types.ID(id), }, IsPromote: true, } b, err := json.Marshal(promoteChangeContext) if err != nil { return nil, err } cc := raftpb.ConfChange{ Type: raftpb.ConfChangeAddNode, NodeID: id, Context: b, } return s.configure(ctx, cc) } func (s *EtcdServer) mayPromoteMember(id types.ID) error { lg := s.Logger() err := s.isLearnerReady(uint64(id)) if err != nil { return err } if !s.Cfg.StrictReconfigCheck { return nil } if !s.cluster.IsReadyToPromoteMember(uint64(id)) { lg.Warn( "rejecting member promote request; not enough healthy members", zap.String("local-member-id", s.MemberId().String()), zap.String("requested-member-remove-id", id.String()), zap.Error(errors.ErrNotEnoughStartedMembers), ) return errors.ErrNotEnoughStartedMembers } return nil } // check whether the learner catches up with leader or not. // Note: it will return nil if member is not found in cluster or if member is not learner. // These two conditions will be checked before toApply phase later. func (s *EtcdServer) isLearnerReady(id uint64) error { rs := s.raftStatus() // leader's raftStatus.Progress is not nil if rs.Progress == nil { return errors.ErrNotLeader } var learnerMatch uint64 isFound := false leaderID := rs.ID for memberID, progress := range rs.Progress { if id == memberID { // check its status learnerMatch = progress.Match isFound = true break } } if isFound { leaderMatch := rs.Progress[leaderID].Match // the learner's Match not caught up with leader yet if float64(learnerMatch) < float64(leaderMatch)*readyPercent { return errors.ErrLearnerNotReady } } return nil } func (s *EtcdServer) mayRemoveMember(id types.ID) error { if !s.Cfg.StrictReconfigCheck { return nil } lg := s.Logger() isLearner := s.cluster.IsMemberExist(id) && s.cluster.Member(id).IsLearner // no need to check quorum when removing non-voting member if isLearner { return nil } if !s.cluster.IsReadyToRemoveVotingMember(uint64(id)) { lg.Warn( "rejecting member remove request; not enough healthy members", zap.String("local-member-id", s.MemberId().String()), zap.String("requested-member-remove-id", id.String()), zap.Error(errors.ErrNotEnoughStartedMembers), ) return errors.ErrNotEnoughStartedMembers } // downed member is safe to remove since it's not part of the active quorum if t := s.r.transport.ActiveSince(id); id != s.MemberId() && t.IsZero() { return nil } // protect quorum if some members are down m := s.cluster.VotingMembers() active := numConnectedSince(s.r.transport, time.Now().Add(-HealthInterval), s.MemberId(), m) if (active - 1) < 1+((len(m)-1)/2) { lg.Warn( "rejecting member remove request; local member has not been connected to all peers, reconfigure breaks active quorum", zap.String("local-member-id", s.MemberId().String()), zap.String("requested-member-remove", id.String()), zap.Int("active-peers", active), zap.Error(errors.ErrUnhealthy), ) return errors.ErrUnhealthy } return nil } func (s *EtcdServer) UpdateMember(ctx context.Context, memb membership.Member) ([]*membership.Member, error) { b, merr := json.Marshal(memb) if merr != nil { return nil, merr } if err := s.checkMembershipOperationPermission(ctx); err != nil { return nil, err } cc := raftpb.ConfChange{ Type: raftpb.ConfChangeUpdateNode, NodeID: uint64(memb.ID), Context: b, } return s.configure(ctx, cc) } func (s *EtcdServer) setCommittedIndex(v uint64) { atomic.StoreUint64(&s.committedIndex, v) } func (s *EtcdServer) getCommittedIndex() uint64 { return atomic.LoadUint64(&s.committedIndex) } func (s *EtcdServer) setAppliedIndex(v uint64) { atomic.StoreUint64(&s.appliedIndex, v) } func (s *EtcdServer) getAppliedIndex() uint64 { return atomic.LoadUint64(&s.appliedIndex) } func (s *EtcdServer) setTerm(v uint64) { atomic.StoreUint64(&s.term, v) } func (s *EtcdServer) getTerm() uint64 { return atomic.LoadUint64(&s.term) } func (s *EtcdServer) setLead(v uint64) { atomic.StoreUint64(&s.lead, v) } func (s *EtcdServer) getLead() uint64 { return atomic.LoadUint64(&s.lead) } func (s *EtcdServer) LeaderChangedNotify() <-chan struct{} { return s.leaderChanged.Receive() } // FirstCommitInTermNotify returns channel that will be unlocked on first // entry committed in new term, which is necessary for new leader to answer // read-only requests (leader is not able to respond any read-only requests // as long as linearizable semantic is required) func (s *EtcdServer) FirstCommitInTermNotify() <-chan struct{} { return s.firstCommitInTerm.Receive() } func (s *EtcdServer) MemberId() types.ID { return s.memberId } func (s *EtcdServer) Leader() types.ID { return types.ID(s.getLead()) } func (s *EtcdServer) Lead() uint64 { return s.getLead() } func (s *EtcdServer) CommittedIndex() uint64 { return s.getCommittedIndex() } func (s *EtcdServer) AppliedIndex() uint64 { return s.getAppliedIndex() } func (s *EtcdServer) Term() uint64 { return s.getTerm() } type confChangeResponse struct { membs []*membership.Member err error } // configure sends a configuration change through consensus and // then waits for it to be applied to the server. It // will block until the change is performed or there is an error. func (s *EtcdServer) configure(ctx context.Context, cc raftpb.ConfChange) ([]*membership.Member, error) { lg := s.Logger() cc.ID = s.reqIDGen.Next() ch := s.w.Register(cc.ID) start := time.Now() if err := s.r.ProposeConfChange(ctx, cc); err != nil { s.w.Trigger(cc.ID, nil) return nil, err } select { case x := <-ch: if x == nil { lg.Panic("failed to configure") } resp := x.(*confChangeResponse) lg.Info( "applied a configuration change through raft", zap.String("local-member-id", s.MemberId().String()), zap.String("raft-conf-change", cc.Type.String()), zap.String("raft-conf-change-node-id", types.ID(cc.NodeID).String()), ) return resp.membs, resp.err case <-ctx.Done(): s.w.Trigger(cc.ID, nil) // GC wait return nil, s.parseProposeCtxErr(ctx.Err(), start) case <-s.stopping: return nil, errors.ErrStopped } } // sync proposes a SYNC request and is non-blocking. // This makes no guarantee that the request will be proposed or performed. // The request will be canceled after the given timeout. func (s *EtcdServer) sync(timeout time.Duration) { req := pb.Request{ Method: "SYNC", ID: s.reqIDGen.Next(), Time: time.Now().UnixNano(), } data := pbutil.MustMarshal(&req) // There is no promise that node has leader when do SYNC request, // so it uses goroutine to propose. ctx, cancel := context.WithTimeout(s.ctx, timeout) s.GoAttach(func() { s.r.Propose(ctx, data) cancel() }) } // publishV3 registers server information into the cluster using v3 request. The // information is the JSON representation of this server's member struct, updated // with the static clientURLs of the server. // The function keeps attempting to register until it succeeds, // or its server is stopped. func (s *EtcdServer) publishV3(timeout time.Duration) { req := &membershippb.ClusterMemberAttrSetRequest{ Member_ID: uint64(s.MemberId()), MemberAttributes: &membershippb.Attributes{ Name: s.attributes.Name, ClientUrls: s.attributes.ClientURLs, }, } lg := s.Logger() for { select { case <-s.stopping: lg.Warn( "stopped publish because server is stopping", zap.String("local-member-id", s.MemberId().String()), zap.String("local-member-attributes", fmt.Sprintf("%+v", s.attributes)), zap.Duration("publish-timeout", timeout), ) return default: } ctx, cancel := context.WithTimeout(s.ctx, timeout) _, err := s.raftRequest(ctx, pb.InternalRaftRequest{ClusterMemberAttrSet: req}) cancel() switch err { case nil: close(s.readych) lg.Info( "published local member to cluster through raft", zap.String("local-member-id", s.MemberId().String()), zap.String("local-member-attributes", fmt.Sprintf("%+v", s.attributes)), zap.String("cluster-id", s.cluster.ID().String()), zap.Duration("publish-timeout", timeout), ) return default: lg.Warn( "failed to publish local member to cluster through raft", zap.String("local-member-id", s.MemberId().String()), zap.String("local-member-attributes", fmt.Sprintf("%+v", s.attributes)), zap.Duration("publish-timeout", timeout), zap.Error(err), ) } } } func (s *EtcdServer) sendMergedSnap(merged snap.Message) { atomic.AddInt64(&s.inflightSnapshots, 1) lg := s.Logger() fields := []zap.Field{ zap.String("from", s.MemberId().String()), zap.String("to", types.ID(merged.To).String()), zap.Int64("bytes", merged.TotalSize), zap.String("size", humanize.Bytes(uint64(merged.TotalSize))), } now := time.Now() s.r.transport.SendSnapshot(merged) lg.Info("sending merged snapshot", fields...) s.GoAttach(func() { select { case ok := <-merged.CloseNotify(): // delay releasing inflight snapshot for another 30 seconds to // block log compaction. // If the follower still fails to catch up, it is probably just too slow // to catch up. We cannot avoid the snapshot cycle anyway. if ok { select { case <-time.After(releaseDelayAfterSnapshot): case <-s.stopping: } } atomic.AddInt64(&s.inflightSnapshots, -1) lg.Info("sent merged snapshot", append(fields, zap.Duration("took", time.Since(now)))...) case <-s.stopping: lg.Warn("canceled sending merged snapshot; server stopping", fields...) return } }) } // toApply takes entries received from Raft (after it has been committed) and // applies them to the current state of the EtcdServer. // The given entries should not be empty. func (s *EtcdServer) apply( es []raftpb.Entry, confState *raftpb.ConfState, ) (appliedt uint64, appliedi uint64, shouldStop bool) { s.lg.Debug("Applying entries", zap.Int("num-entries", len(es))) for i := range es { e := es[i] s.lg.Debug("Applying entry", zap.Uint64("index", e.Index), zap.Uint64("term", e.Term), zap.Stringer("type", e.Type)) switch e.Type { case raftpb.EntryNormal: s.applyEntryNormal(&e) s.setAppliedIndex(e.Index) s.setTerm(e.Term) case raftpb.EntryConfChange: // We need to toApply all WAL entries on top of v2store // and only 'unapplied' (e.Index>backend.ConsistentIndex) on the backend. shouldApplyV3 := membership.ApplyV2storeOnly // set the consistent index of current executing entry if e.Index > s.consistIndex.ConsistentIndex() { s.consistIndex.SetConsistentApplyingIndex(e.Index, e.Term) shouldApplyV3 = membership.ApplyBoth } var cc raftpb.ConfChange pbutil.MustUnmarshal(&cc, e.Data) removedSelf, err := s.applyConfChange(cc, confState, shouldApplyV3) s.setAppliedIndex(e.Index) s.setTerm(e.Term) shouldStop = shouldStop || removedSelf s.w.Trigger(cc.ID, &confChangeResponse{s.cluster.Members(), err}) default: lg := s.Logger() lg.Panic( "unknown entry type; must be either EntryNormal or EntryConfChange", zap.String("type", e.Type.String()), ) } appliedi, appliedt = e.Index, e.Term } return appliedt, appliedi, shouldStop } // applyEntryNormal applies an EntryNormal type raftpb request to the EtcdServer func (s *EtcdServer) applyEntryNormal(e *raftpb.Entry) { shouldApplyV3 := membership.ApplyV2storeOnly applyV3Performed := false var ar *apply.Result index := s.consistIndex.ConsistentIndex() if e.Index > index { // set the consistent index of current executing entry s.consistIndex.SetConsistentApplyingIndex(e.Index, e.Term) shouldApplyV3 = membership.ApplyBoth defer func() { // The txPostLockInsideApplyHook will not get called in some cases, // in which we should move the consistent index forward directly. if !applyV3Performed || (ar != nil && ar.Err != nil) { s.consistIndex.SetConsistentIndex(e.Index, e.Term) } }() } s.lg.Debug("toApply entry normal", zap.Uint64("consistent-index", index), zap.Uint64("entry-index", e.Index), zap.Bool("should-applyV3", bool(shouldApplyV3))) // raft state machine may generate noop entry when leader confirmation. // skip it in advance to avoid some potential bug in the future if len(e.Data) == 0 { s.firstCommitInTerm.Notify() // promote lessor when the local member is leader and finished // applying all entries from the last term. if s.isLeader() { s.lessor.Promote(s.Cfg.ElectionTimeout()) } return } var raftReq pb.InternalRaftRequest if !pbutil.MaybeUnmarshal(&raftReq, e.Data) { // backward compatible var r pb.Request rp := &r pbutil.MustUnmarshal(rp, e.Data) s.lg.Debug("applyEntryNormal", zap.Stringer("V2request", rp)) s.w.Trigger(r.ID, s.applyV2Request((*RequestV2)(rp), shouldApplyV3)) return } s.lg.Debug("applyEntryNormal", zap.Stringer("raftReq", &raftReq)) if raftReq.V2 != nil { req := (*RequestV2)(raftReq.V2) s.w.Trigger(req.ID, s.applyV2Request(req, shouldApplyV3)) return } id := raftReq.ID if id == 0 { if raftReq.Header == nil { s.lg.Panic("applyEntryNormal, could not find a header") } id = raftReq.Header.ID } needResult := s.w.IsRegistered(id) if needResult || !noSideEffect(&raftReq) { if !needResult && raftReq.Txn != nil { removeNeedlessRangeReqs(raftReq.Txn) } applyV3Performed = true ar = s.uberApply.Apply(&raftReq, shouldApplyV3) } // do not re-toApply applied entries. if !shouldApplyV3 { return } if ar == nil { return } if ar.Err != errors.ErrNoSpace || len(s.alarmStore.Get(pb.AlarmType_NOSPACE)) > 0 { s.w.Trigger(id, ar) return } lg := s.Logger() lg.Warn( "message exceeded backend quota; raising alarm", zap.Int64("quota-size-bytes", s.Cfg.QuotaBackendBytes), zap.String("quota-size", humanize.Bytes(uint64(s.Cfg.QuotaBackendBytes))), zap.Error(ar.Err), ) s.GoAttach(func() { a := &pb.AlarmRequest{ MemberID: uint64(s.MemberId()), Action: pb.AlarmRequest_ACTIVATE, Alarm: pb.AlarmType_NOSPACE, } s.raftRequest(s.ctx, pb.InternalRaftRequest{Alarm: a}) s.w.Trigger(id, ar) }) } func noSideEffect(r *pb.InternalRaftRequest) bool { return r.Range != nil || r.AuthUserGet != nil || r.AuthRoleGet != nil || r.AuthStatus != nil } func removeNeedlessRangeReqs(txn *pb.TxnRequest) { f := func(ops []*pb.RequestOp) []*pb.RequestOp { j := 0 for i := 0; i < len(ops); i++ { if _, ok := ops[i].Request.(*pb.RequestOp_RequestRange); ok { continue } ops[j] = ops[i] j++ } return ops[:j] } txn.Success = f(txn.Success) txn.Failure = f(txn.Failure) } // applyConfChange applies a ConfChange to the server. It is only // invoked with a ConfChange that has already passed through Raft func (s *EtcdServer) applyConfChange(cc raftpb.ConfChange, confState *raftpb.ConfState, shouldApplyV3 membership.ShouldApplyV3) (bool, error) { if err := s.cluster.ValidateConfigurationChange(cc); err != nil { cc.NodeID = raft.None s.r.ApplyConfChange(cc) // The txPostLock callback will not get called in this case, // so we should set the consistent index directly. if s.consistIndex != nil && membership.ApplyBoth == shouldApplyV3 { applyingIndex, applyingTerm := s.consistIndex.ConsistentApplyingIndex() s.consistIndex.SetConsistentIndex(applyingIndex, applyingTerm) } return false, err } lg := s.Logger() *confState = *s.r.ApplyConfChange(cc) s.beHooks.SetConfState(confState) switch cc.Type { case raftpb.ConfChangeAddNode, raftpb.ConfChangeAddLearnerNode: confChangeContext := new(membership.ConfigChangeContext) if err := json.Unmarshal(cc.Context, confChangeContext); err != nil { lg.Panic("failed to unmarshal member", zap.Error(err)) } if cc.NodeID != uint64(confChangeContext.Member.ID) { lg.Panic( "got different member ID", zap.String("member-id-from-config-change-entry", types.ID(cc.NodeID).String()), zap.String("member-id-from-message", confChangeContext.Member.ID.String()), ) } if confChangeContext.IsPromote { s.cluster.PromoteMember(confChangeContext.Member.ID, shouldApplyV3) } else { s.cluster.AddMember(&confChangeContext.Member, shouldApplyV3) if confChangeContext.Member.ID != s.MemberId() { s.r.transport.AddPeer(confChangeContext.Member.ID, confChangeContext.PeerURLs) } } // update the isLearner metric when this server id is equal to the id in raft member confChange if confChangeContext.Member.ID == s.MemberId() { if cc.Type == raftpb.ConfChangeAddLearnerNode { isLearner.Set(1) } else { isLearner.Set(0) } } case raftpb.ConfChangeRemoveNode: id := types.ID(cc.NodeID) s.cluster.RemoveMember(id, shouldApplyV3) if id == s.MemberId() { return true, nil } s.r.transport.RemovePeer(id) case raftpb.ConfChangeUpdateNode: m := new(membership.Member) if err := json.Unmarshal(cc.Context, m); err != nil { lg.Panic("failed to unmarshal member", zap.Error(err)) } if cc.NodeID != uint64(m.ID) { lg.Panic( "got different member ID", zap.String("member-id-from-config-change-entry", types.ID(cc.NodeID).String()), zap.String("member-id-from-message", m.ID.String()), ) } s.cluster.UpdateRaftAttributes(m.ID, m.RaftAttributes, shouldApplyV3) if m.ID != s.MemberId() { s.r.transport.UpdatePeer(m.ID, m.PeerURLs) } } return false, nil } // TODO: non-blocking snapshot func (s *EtcdServer) snapshot(snapi uint64, confState raftpb.ConfState) { clone := s.v2store.Clone() // commit kv to write metadata (for example: consistent index) to disk. // // This guarantees that Backend's consistent_index is >= index of last snapshot. // // KV().commit() updates the consistent index in backend. // All operations that update consistent index must be called sequentially // from applyAll function. // So KV().Commit() cannot run in parallel with toApply. It has to be called outside // the go routine created below. s.KV().Commit() s.GoAttach(func() { lg := s.Logger() d, err := clone.SaveNoCopy() // TODO: current store will never fail to do a snapshot // what should we do if the store might fail? if err != nil { lg.Panic("failed to save v2 store", zap.Error(err)) } snap, err := s.r.raftStorage.CreateSnapshot(snapi, &confState, d) if err != nil { // the snapshot was done asynchronously with the progress of raft. // raft might have already got a newer snapshot. if err == raft.ErrSnapOutOfDate { return } lg.Panic("failed to create snapshot", zap.Error(err)) } // SaveSnap saves the snapshot to file and appends the corresponding WAL entry. if err = s.r.storage.SaveSnap(snap); err != nil { lg.Panic("failed to save snapshot", zap.Error(err)) } if err = s.r.storage.Release(snap); err != nil { lg.Panic("failed to release wal", zap.Error(err)) } lg.Info( "saved snapshot", zap.Uint64("snapshot-index", snap.Metadata.Index), ) // When sending a snapshot, etcd will pause compaction. // After receives a snapshot, the slow follower needs to get all the entries right after // the snapshot sent to catch up. If we do not pause compaction, the log entries right after // the snapshot sent might already be compacted. It happens when the snapshot takes long time // to send and save. Pausing compaction avoids triggering a snapshot sending cycle. if atomic.LoadInt64(&s.inflightSnapshots) != 0 { lg.Info("skip compaction since there is an inflight snapshot") return } // keep some in memory log entries for slow followers. compacti := uint64(1) if snapi > s.Cfg.SnapshotCatchUpEntries { compacti = snapi - s.Cfg.SnapshotCatchUpEntries } err = s.r.raftStorage.Compact(compacti) if err != nil { // the compaction was done asynchronously with the progress of raft. // raft log might already been compact. if err == raft.ErrCompacted { return } lg.Panic("failed to compact", zap.Error(err)) } lg.Info( "compacted Raft logs", zap.Uint64("compact-index", compacti), ) }) } // CutPeer drops messages to the specified peer. func (s *EtcdServer) CutPeer(id types.ID) { tr, ok := s.r.transport.(*rafthttp.Transport) if ok { tr.CutPeer(id) } } // MendPeer recovers the message dropping behavior of the given peer. func (s *EtcdServer) MendPeer(id types.ID) { tr, ok := s.r.transport.(*rafthttp.Transport) if ok { tr.MendPeer(id) } } func (s *EtcdServer) PauseSending() { s.r.pauseSending() } func (s *EtcdServer) ResumeSending() { s.r.resumeSending() } func (s *EtcdServer) ClusterVersion() *semver.Version { if s.cluster == nil { return nil } return s.cluster.Version() } func (s *EtcdServer) StorageVersion() *semver.Version { // `applySnapshot` sets a new backend instance, so we need to acquire the bemu lock. s.bemu.RLock() defer s.bemu.RUnlock() v, err := schema.DetectSchemaVersion(s.lg, s.be.ReadTx()) if err != nil { s.lg.Warn("Failed to detect schema version", zap.Error(err)) return nil } return &v } // monitorClusterVersions every monitorVersionInterval checks if it's the leader and updates cluster version if needed. func (s *EtcdServer) monitorClusterVersions() { monitor := serverversion.NewMonitor(s.Logger(), NewServerVersionAdapter(s)) for { select { case <-s.firstCommitInTerm.Receive(): case <-time.After(monitorVersionInterval): case <-s.stopping: return } if s.Leader() != s.MemberId() { continue } monitor.UpdateClusterVersionIfNeeded() } } // monitorStorageVersion every monitorVersionInterval updates storage version if needed. func (s *EtcdServer) monitorStorageVersion() { monitor := serverversion.NewMonitor(s.Logger(), NewServerVersionAdapter(s)) for { select { case <-time.After(monitorVersionInterval): case <-s.clusterVersionChanged.Receive(): case <-s.stopping: return } monitor.UpdateStorageVersionIfNeeded() } } func (s *EtcdServer) updateClusterVersionV2(ver string) { lg := s.Logger() if s.cluster.Version() == nil { lg.Info( "setting up initial cluster version using v2 API", zap.String("cluster-version", version.Cluster(ver)), ) } else { lg.Info( "updating cluster version using v2 API", zap.String("from", version.Cluster(s.cluster.Version().String())), zap.String("to", version.Cluster(ver)), ) } req := pb.Request{ Method: "PUT", Path: membership.StoreClusterVersionKey(), Val: ver, } ctx, cancel := context.WithTimeout(s.ctx, s.Cfg.ReqTimeout()) _, err := s.Do(ctx, req) cancel() switch err { case nil: lg.Info("cluster version is updated", zap.String("cluster-version", version.Cluster(ver))) return case errors.ErrStopped: lg.Warn("aborting cluster version update; server is stopped", zap.Error(err)) return default: lg.Warn("failed to update cluster version", zap.Error(err)) } } func (s *EtcdServer) updateClusterVersionV3(ver string) { lg := s.Logger() if s.cluster.Version() == nil { lg.Info( "setting up initial cluster version using v3 API", zap.String("cluster-version", version.Cluster(ver)), ) } else { lg.Info( "updating cluster version using v3 API", zap.String("from", version.Cluster(s.cluster.Version().String())), zap.String("to", version.Cluster(ver)), ) } req := membershippb.ClusterVersionSetRequest{Ver: ver} ctx, cancel := context.WithTimeout(s.ctx, s.Cfg.ReqTimeout()) _, err := s.raftRequest(ctx, pb.InternalRaftRequest{ClusterVersionSet: &req}) cancel() switch err { case nil: lg.Info("cluster version is updated", zap.String("cluster-version", version.Cluster(ver))) return case errors.ErrStopped: lg.Warn("aborting cluster version update; server is stopped", zap.Error(err)) return default: lg.Warn("failed to update cluster version", zap.Error(err)) } } // monitorDowngrade every DowngradeCheckTime checks if it's the leader and cancels downgrade if needed. func (s *EtcdServer) monitorDowngrade() { monitor := serverversion.NewMonitor(s.Logger(), NewServerVersionAdapter(s)) t := s.Cfg.DowngradeCheckTime if t == 0 { return } for { select { case <-time.After(t): case <-s.stopping: return } if !s.isLeader() { continue } monitor.CancelDowngradeIfNeeded() } } func (s *EtcdServer) parseProposeCtxErr(err error, start time.Time) error { switch err { case context.Canceled: return errors.ErrCanceled case context.DeadlineExceeded: s.leadTimeMu.RLock() curLeadElected := s.leadElectedTime s.leadTimeMu.RUnlock() prevLeadLost := curLeadElected.Add(-2 * time.Duration(s.Cfg.ElectionTicks) * time.Duration(s.Cfg.TickMs) * time.Millisecond) if start.After(prevLeadLost) && start.Before(curLeadElected) { return errors.ErrTimeoutDueToLeaderFail } lead := types.ID(s.getLead()) switch lead { case types.ID(raft.None): // TODO: return error to specify it happens because the cluster does not have leader now case s.MemberId(): if !isConnectedToQuorumSince(s.r.transport, start, s.MemberId(), s.cluster.Members()) { return errors.ErrTimeoutDueToConnectionLost } default: if !isConnectedSince(s.r.transport, start, lead) { return errors.ErrTimeoutDueToConnectionLost } } return errors.ErrTimeout default: return err } } func (s *EtcdServer) KV() mvcc.WatchableKV { return s.kv } func (s *EtcdServer) Backend() backend.Backend { s.bemu.RLock() defer s.bemu.RUnlock() return s.be } func (s *EtcdServer) AuthStore() auth.AuthStore { return s.authStore } func (s *EtcdServer) restoreAlarms() error { as, err := v3alarm.NewAlarmStore(s.lg, schema.NewAlarmBackend(s.lg, s.be)) if err != nil { return err } s.alarmStore = as return nil } // GoAttach creates a goroutine on a given function and tracks it using // the etcdserver waitgroup. // The passed function should interrupt on s.StoppingNotify(). func (s *EtcdServer) GoAttach(f func()) { s.wgMu.RLock() // this blocks with ongoing close(s.stopping) defer s.wgMu.RUnlock() select { case <-s.stopping: lg := s.Logger() lg.Warn("server has stopped; skipping GoAttach") return default: } // now safe to add since waitgroup wait has not started yet s.wg.Add(1) go func() { defer s.wg.Done() f() }() } func (s *EtcdServer) Alarms() []*pb.AlarmMember { return s.alarmStore.Get(pb.AlarmType_NONE) } // IsLearner returns if the local member is raft learner func (s *EtcdServer) IsLearner() bool { return s.cluster.IsLocalMemberLearner() } // IsMemberExist returns if the member with the given id exists in cluster. func (s *EtcdServer) IsMemberExist(id types.ID) bool { return s.cluster.IsMemberExist(id) } // raftStatus returns the raft status of this etcd node. func (s *EtcdServer) raftStatus() raft.Status { return s.r.Node.Status() } func (s *EtcdServer) Version() *serverversion.Manager { return serverversion.NewManager(s.Logger(), NewServerVersionAdapter(s)) } func (s *EtcdServer) getTxPostLockInsideApplyHook() func() { return func() { applyingIdx, applyingTerm := s.consistIndex.ConsistentApplyingIndex() if applyingIdx > s.consistIndex.UnsafeConsistentIndex() { s.consistIndex.SetConsistentIndex(applyingIdx, applyingTerm) } } }