Support failover in cluster (experimental)

aof/aof.go

@@ -259,6 +259,9 @@ func (persister *Persister) Fsync() {
 
 // Close gracefully stops aof persistence procedure
 func (persister *Persister) Close() {
+	if persister == nil {
+		return
+	}
 	if persister.aofFile != nil {
 		close(persister.aofChan)
 		<-persister.aofFinished // wait for aof finished

cluster/cluster.go

@@ -30,6 +30,7 @@ func MakeCluster() *Cluster {
 		},
 		StartAsSeed: config.Properties.ClusterAsSeed,
 		JoinAddress: config.Properties.ClusterSeed,
+		Master:      config.Properties.MasterInCluster,
 	})
 	if err != nil {
 		logger.Error(err.Error())

cluster/core/core.go

@@ -22,6 +22,9 @@ type Cluster struct {
 	slotsManager    *slotsManager
 	rebalanceManger *rebalanceManager
 	transactions    *TransactionManager
+	replicaManager  *replicaManager
+
+	closeChan chan struct{}
 
 	// allow inject route implementation
 	getSlotImpl  func(key string) uint32
@@ -33,7 +36,9 @@ type Config struct {
 	raft.RaftConfig
 	StartAsSeed    bool
 	JoinAddress    string
+	Master         string
 	connectionStub ConnectionFactory // for test
+	noCron         bool // for test
 }
 
 func (c *Cluster) SelfID() string {
@@ -123,7 +128,11 @@ func NewCluster(cfg *Config) (*Cluster, error) {
 			if err != nil {
 				return nil, err
 			}
-			result := conn.Send(utils.ToCmdLine(joinClusterCommand, cfg.RedisAdvertiseAddr, cfg.RaftAdvertiseAddr))
+			joinCmdLine := utils.ToCmdLine(joinClusterCommand, cfg.RedisAdvertiseAddr, cfg.RaftAdvertiseAddr)
+			if cfg.Master != "" {
+				joinCmdLine = append(joinCmdLine, []byte(cfg.Master))
+			}
+			result := conn.Send(joinCmdLine)
 			if err := protocol.Try2ErrorReply(result); err != nil {
 				return nil, err
 			}
@@ -137,6 +146,8 @@ func NewCluster(cfg *Config) (*Cluster, error) {
 		rebalanceManger: newRebalanceManager(),
 		slotsManager:    newSlotsManager(),
 		transactions:    newTransactionManager(),
+		replicaManager:  newReplicaManager(),
+		closeChan:       make(chan struct{}),
 	}
 	cluster.pickNodeImpl = func(slotID uint32) string {
 		return defaultPickNodeImpl(cluster, slotID)
@@ -146,6 +157,8 @@ func NewCluster(cfg *Config) (*Cluster, error) {
 	}
 	cluster.injectInsertCallback()
 	cluster.injectDeleteCallback()
+	cluster.registerOnFailover()
+	go cluster.clusterCron()
 	return cluster, nil
 }
 
@@ -155,6 +168,7 @@ func (cluster *Cluster) AfterClientClose(c redis.Connection) {
 }
 
 func (cluster *Cluster) Close() {
+	close(cluster.closeChan)
 	cluster.db.Close()
 	err := cluster.raftNode.Close()
 	if err != nil {

cluster/core/core_test.go

@@ -6,6 +6,7 @@ import (
 	"time"
 
 	"github.com/hdt3213/godis/cluster/raft"
+	"github.com/hdt3213/godis/interface/redis"
 	"github.com/hdt3213/godis/lib/utils"
 	"github.com/hdt3213/godis/redis/connection"
 	"github.com/hdt3213/godis/redis/protocol"
@@ -33,6 +34,7 @@ func TestClusterBootstrap(t *testing.T) {
 		},
 		StartAsSeed:    true,
 		connectionStub: connections,
+		noCron:         true,
 	}
 	leader, err := NewCluster(leaderCfg)
 	if err != nil {
@@ -72,6 +74,7 @@ func TestClusterBootstrap(t *testing.T) {
 		StartAsSeed:    false,
 		JoinAddress:    leaderCfg.RedisAdvertiseAddr,
 		connectionStub: connections,
+		noCron:         true,
 	}
 	follower, err := NewCluster(followerCfg)
 	if err != nil {
@@ -132,3 +135,104 @@ func TestClusterBootstrap(t *testing.T) {
 		}
 	}
 }
+
+func TestFailover(t *testing.T) {
+	// start leader
+	leaderDir := "test/0"
+	os.RemoveAll(leaderDir)
+	os.MkdirAll(leaderDir, 0777)
+	defer func() {
+		os.RemoveAll(leaderDir)
+	}()
+	RegisterCmd("slaveof", func(cluster *Cluster, c redis.Connection, cmdLine CmdLine) redis.Reply {
+		return protocol.MakeOkReply()
+	})
+
+	// connection stub
+	connections := NewInMemConnectionFactory()
+	leaderCfg := &Config{
+		RaftConfig: raft.RaftConfig{
+			RedisAdvertiseAddr: "127.0.0.1:6399",
+			RaftListenAddr:     "127.0.0.1:26666",
+			RaftAdvertiseAddr:  "127.0.0.1:26666",
+			Dir:                leaderDir,
+		},
+		StartAsSeed:    true,
+		connectionStub: connections,
+		noCron:         true,
+	}
+	leader, err := NewCluster(leaderCfg)
+	if err != nil {
+		t.Error(err)
+		return
+	}
+	connections.nodes[leaderCfg.RedisAdvertiseAddr] = leader
+
+	// start follower
+	followerDir := "test/1"
+	os.RemoveAll(followerDir)
+	os.MkdirAll(followerDir, 0777)
+	defer func() {
+		os.RemoveAll(followerDir)
+	}()
+	followerCfg := &Config{
+		RaftConfig: raft.RaftConfig{
+			RedisAdvertiseAddr: "127.0.0.1:6499",
+			RaftListenAddr:     "127.0.0.1:26667",
+			RaftAdvertiseAddr:  "127.0.0.1:26667",
+			Dir:                followerDir,
+		},
+		StartAsSeed:    false,
+		JoinAddress:    leaderCfg.RedisAdvertiseAddr,
+		connectionStub: connections,
+		noCron:         true,
+		Master:         leader.SelfID(),
+	}
+	follower, err := NewCluster(followerCfg)
+	if err != nil {
+		t.Error(err)
+		return
+	}
+	connections.nodes[followerCfg.RedisAdvertiseAddr] = follower
+
+	_ = follower.SelfID()
+	// check nodes
+	joined := false
+	for i := 0; i < 10; i++ {
+		nodes, err := leader.raftNode.GetNodes()
+		if err != nil {
+			t.Log(err)
+			continue
+		}
+		if len(nodes) == 2 {
+			t.Log("join success")
+			joined = true
+			break
+		}
+		time.Sleep(time.Second)
+	}
+	if !joined {
+		t.Error("join failed")
+		return
+	}
+
+	// rebalance
+	leader.replicaManager.masterHeartbeats[leader.SelfID()] = time.Now().Add(-time.Hour)
+	leader.doFailoverCheck()
+	time.Sleep(2 * time.Second)
+	for i := 0; i < 1000; i++ {
+		success := false
+		leader.raftNode.FSM.WithReadLock(func(fsm *raft.FSM) {
+			ms := fsm.MasterSlaves[follower.SelfID()]	
+			if ms != nil && len(ms.Slaves) > 0 {
+				success = true
+			}
+		})
+		if success {
+			t.Log("rebalance success")
+			break
+		} else {
+			time.Sleep(time.Second)
+		}
+	}
+}

cluster/core/cron.go

@@ -0,0 +1,36 @@
+package core
+
+import (
+	"sync/atomic"
+	"time"
+
+	"github.com/hdt3213/godis/cluster/raft"
+)
+
+func (cluster *Cluster) clusterCron() {
+	if cluster.config.noCron {
+		return
+	}
+	ticker := time.NewTicker(time.Second)
+	var running int32
+	for {
+		select {
+		case <-ticker.C:
+			if cluster.raftNode.State() == raft.Leader {
+				if atomic.CompareAndSwapInt32(&running, 0, 1) {
+					// Disable parallelism
+					go func() {
+						cluster.doFailoverCheck()
+						cluster.doRebalance()
+						atomic.StoreInt32(&running, 0)
+					}()
+				}
+			} else {
+				cluster.sendHearbeat()
+			}
+		case <-cluster.closeChan:
+			ticker.Stop()
+			return
+		}
+	}
+}

cluster/core/node_manager.go

@@ -23,10 +23,10 @@ func init() {
 	RegisterCmd(migrationChangeRouteCommand, execMigrationChangeRoute)
 }
 
-// execJoin handles cluster-join command
-// format: cluster-join redisAddress (advertised)raftAddress
+// execJoin handles cluster-join command as raft leader
+// format: cluster-join redisAddress(advertised), raftAddress, masterId
 func execJoin(cluster *Cluster, c redis.Connection, cmdLine CmdLine) redis.Reply {
-	if len(cmdLine) != 3 {
+	if len(cmdLine) < 3 {
 		return protocol.MakeArgNumErrReply(joinClusterCommand)
 	}
 	state := cluster.raftNode.State()
@@ -42,10 +42,26 @@ func execJoin(cluster *Cluster, c redis.Connection, cmdLine CmdLine) redis.Reply
 	// self node is leader
 	redisAddr := string(cmdLine[1])
 	raftAddr := string(cmdLine[2])
-	err := cluster.raftNode.HandleJoin(redisAddr, raftAddr)
+	err := cluster.raftNode.AddToRaft(redisAddr, raftAddr)
 	if err != nil {
 		return protocol.MakeErrReply(err.Error())
 	}
+	master := ""
+	if len(cmdLine) == 4 {
+		master = string(cmdLine[3])
+	}
+	_, err = cluster.raftNode.Propose(&raft.LogEntry{
+		Event: raft.EventJoin,
+		JoinTask: &raft.JoinTask{
+			NodeId: redisAddr,
+			Master: master,
+		},
+	})
+	if err != nil {
+		// todo: remove the node from raft
+		return protocol.MakeErrReply(err.Error())
+	}
+
 	// join sucees, rebalance node
 	return protocol.MakeOkReply()
 }
@@ -114,17 +130,14 @@ func (cluster *Cluster) triggerMigrationTask(task *raft.MigratingTask) error {
 }
 
 func (cluster *Cluster) makeRebalancePlan() ([]*raft.MigratingTask, error) {
-	nodes, err := cluster.raftNode.GetNodes()
-	if err != nil {
-		return nil, err
-	}
-	avgSlot := int(math.Ceil(float64(SlotCount) / float64(len(nodes))))
+
 	var migratings []*raft.MigratingTask
 	cluster.raftNode.FSM.WithReadLock(func(fsm *raft.FSM) {
+		avgSlot := int(math.Ceil(float64(SlotCount) / float64(len(fsm.MasterSlaves))))
 		var exportingNodes []string
 		var importingNodes []string
-		for _, node := range nodes {
-			nodeId := string(node.ID)
+		for _, ms := range fsm.MasterSlaves {
+			nodeId := ms.MasterId
 			nodeSlots := fsm.Node2Slot[nodeId]
 			if len(nodeSlots) > avgSlot+1 {
 				exportingNodes = append(exportingNodes, nodeId)
@@ -200,7 +213,7 @@ func (cluster *Cluster) waitCommitted(peer string, logIndex uint64) error {
 // format: cluster.migration.changeroute taskid
 func execMigrationChangeRoute(cluster *Cluster, c redis.Connection, cmdLine CmdLine) redis.Reply {
 	if len(cmdLine) != 2 {
-		return protocol.MakeArgNumErrReply(joinClusterCommand)
+		return protocol.MakeArgNumErrReply(migrationChangeRouteCommand)
 	}
 	state := cluster.raftNode.State()
 	if state != raft.Leader {

cluster/core/replica_manager.go

@@ -0,0 +1,155 @@
+package core
+
+import (
+	"net"
+	"sync"
+	"time"
+
+	"github.com/hdt3213/godis/cluster/raft"
+	"github.com/hdt3213/godis/interface/redis"
+	"github.com/hdt3213/godis/lib/logger"
+	"github.com/hdt3213/godis/lib/utils"
+	"github.com/hdt3213/godis/redis/connection"
+	"github.com/hdt3213/godis/redis/protocol"
+)
+
+const heartbeatCommand = "cluster.heartbeat"
+
+func init() {
+	RegisterCmd(heartbeatCommand, execHeartbeat)
+}
+
+const (
+	statusNormal  = iota
+	statusFailing // failover in progress
+)
+
+type replicaManager struct {
+	mu               sync.RWMutex
+	masterHeartbeats map[string]time.Time // id -> lastHeartbeatTime
+}
+
+func newReplicaManager() *replicaManager {
+	return &replicaManager{
+		masterHeartbeats: make(map[string]time.Time),
+	}
+}
+
+// execHeartbeat receives heartbeat from follower as raft leader
+// cmdLine: cluster.heartbeat nodeId
+func execHeartbeat(cluster *Cluster, c redis.Connection, cmdLine CmdLine) redis.Reply {
+	if len(cmdLine) != 2 {
+		return protocol.MakeArgNumErrReply(heartbeatCommand)
+	}
+	id := string(cmdLine[1])
+	cluster.replicaManager.mu.Lock()
+	cluster.replicaManager.masterHeartbeats[id] = time.Now()
+	cluster.replicaManager.mu.Unlock()
+
+	return protocol.MakeOkReply()
+}
+
+func (cluster *Cluster) sendHearbeat() {
+	leaderConn, err := cluster.BorrowLeaderClient()
+	if err != nil {
+		logger.Error(err)
+	}
+	defer cluster.connections.ReturnPeerClient(leaderConn)
+	reply := leaderConn.Send(utils.ToCmdLine(heartbeatCommand, cluster.SelfID()))
+	if err := protocol.Try2ErrorReply(reply); err != nil {
+		logger.Error(err)
+	}
+}
+
+const followerTimeout = 10 * time.Second
+
+func (cluster *Cluster) doFailoverCheck() {
+	// find timeout masters
+	var timeoutMasters []*raft.MasterSlave
+	ddl := time.Now().Add(-followerTimeout)
+	cluster.replicaManager.mu.RLock()
+	for masterId, lastTime := range cluster.replicaManager.masterHeartbeats {
+		if lastTime.Second() == 0 {
+			// do not set new joined node as timeout
+			cluster.replicaManager.masterHeartbeats[masterId] = time.Now()
+		}
+		if lastTime.Before(ddl) {
+			slaves := cluster.raftNode.GetSlaves(masterId)
+			if slaves != nil && len(slaves.Slaves) > 0 {
+				timeoutMasters = append(timeoutMasters, slaves)
+			}
+
+		}
+	}
+	cluster.replicaManager.mu.RUnlock()
+
+	// trigger failover
+	for _, failed := range timeoutMasters {
+		cluster.triggerFailover(failed)
+	}
+}
+
+func (cluster *Cluster) triggerFailover(failed *raft.MasterSlave) error {
+	newMaster := failed.Slaves[0]
+	id := utils.RandString(20)
+	// propose change
+	_, err := cluster.raftNode.Propose(&raft.LogEntry{
+		Event: raft.EventStartFailover,
+		FailoverTask: &raft.FailoverTask{
+			ID:          id,
+			OldMasterId: failed.MasterId,
+			NewMasterId: newMaster,
+		},
+	})
+	if err != nil {
+		return err
+	}
+	logger.Infof("proposed start failover id=%s, oldMaster=%s, newMaster=%s", id, failed.MasterId, newMaster)
+	// send slave of to new master
+	conn, err := cluster.connections.BorrowPeerClient(newMaster)
+	if err != nil {
+		return err
+	}
+	defer cluster.connections.ReturnPeerClient(conn)
+
+	reply := conn.Send(utils.ToCmdLine("slaveof", "no", "one"))
+	if err := protocol.Try2ErrorReply(reply); err != nil {
+		return err
+	}
+
+	// new master is ready to receive commands, change route
+	_, err = cluster.raftNode.Propose(&raft.LogEntry{
+		Event: raft.EventFinishFailover,
+		FailoverTask: &raft.FailoverTask{
+			ID:          id,
+			OldMasterId: failed.MasterId,
+			NewMasterId: newMaster,
+		},
+	})
+	if err != nil {
+		return err
+	}
+	logger.Infof("proposed finish failover id=%s, oldMaster=%s, newMaster=%s", id, failed.MasterId, newMaster)
+	// other slaves will listen to raft to change master
+	return nil
+}
+
+func (cluster *Cluster) registerOnFailover() {
+	cluster.raftNode.SetOnFailover(func(newMaster string) {
+		if newMaster != "" && newMaster != cluster.SelfID() {
+			// old master failed and other node become the new master
+			// this node may be the old master
+			ip, port, err := net.SplitHostPort(newMaster)
+			if err != nil {
+				logger.Errorf("illegal new master: %s", newMaster)
+				return
+			}
+			c := connection.NewFakeConn()
+			ret := cluster.db.Exec(c, utils.ToCmdLine("slaveof", ip, port))
+			if err := protocol.Try2ErrorReply(ret); err != nil {
+				logger.Errorf("slave of failed: %v", err)
+				return
+			}
+		}
+	})
+}

cluster/raft/fsm.go

@@ -3,7 +3,6 @@ package raft
 import (
 	"encoding/json"
 	"io"
-	"sort"
 	"sync"
 
 	"github.com/hashicorp/raft"
@@ -18,13 +17,22 @@ import (
 // If the target node crashes during migrating, the migration will be canceled.
 // All related commands will be routed to the source node
 type FSM struct {
-	mu         sync.RWMutex
-	Node2Slot  map[string][]uint32       // nodeID -> slotIDs, slotIDs is in ascending order and distinct
-	Slot2Node  map[uint32]string         // slotID -> nodeID
-	Migratings map[string]*MigratingTask // taskId -> task
+	mu           sync.RWMutex
+	Node2Slot    map[string][]uint32       // nodeID -> slotIDs, slotIDs is in ascending order and distinct
+	Slot2Node    map[uint32]string         // slotID -> nodeID
+	Migratings   map[string]*MigratingTask // taskId -> task
+	MasterSlaves map[string]*MasterSlave   // masterId -> MasterSlave
+	SlaveMasters map[string]string         // slaveId -> masterId
+	Failovers    map[string]*FailoverTask  // taskId -> task
+	changed      func(*FSM)                // called while fsm changed, within readlock
 }
 
-// MigratingTask
+type MasterSlave struct {
+	MasterId string
+	Slaves   []string
+}
+
+// MigratingTask is a running migrating task
 // It is immutable
 type MigratingTask struct {
 	ID         string
@@ -35,17 +43,33 @@ type MigratingTask struct {
 	Slots []uint32
 }
 
-// InitTask
+// InitTask assigns all slots to seed node while starting a new cluster
+// It is designed to init the FSM for a new cluster
 type InitTask struct {
 	Leader    string
 	SlotCount int
 }
 
+// FailoverTask represents a failover or joinery/quitting of slaves
+type FailoverTask struct {
+	ID          string
+	OldMasterId string
+	NewMasterId string
+}
+
+type JoinTask struct {
+	NodeId string
+	Master string
+}
+
 // implements FSM.Apply after you created a new raft event
 const (
 	EventStartMigrate = iota + 1
 	EventFinishMigrate
 	EventSeedStart
+	EventStartFailover
+	EventFinishFailover
+	EventJoin
 )
 
 // LogEntry is an entry in raft log, stores a change of cluster
@@ -53,6 +77,8 @@ type LogEntry struct {
 	Event         int
 	MigratingTask *MigratingTask `json:"MigratingTask,omitempty"`
 	InitTask      *InitTask
+	FailoverTask  *FailoverTask
+	JoinTask      *JoinTask
 }
 
 // Apply is called once a log entry is committed by a majority of the cluster.
@@ -82,52 +108,33 @@ func (fsm *FSM) Apply(log *raft.Log) interface{} {
 			slots[i] = uint32(i)
 		}
 		fsm.Node2Slot[entry.InitTask.Leader] = slots
+		fsm.addNode(entry.InitTask.Leader, "")
+	} else if entry.Event == EventStartFailover {
+		task := entry.FailoverTask
+		fsm.Failovers[task.ID] = task
+	} else if entry.Event == EventFinishFailover {
+		task := entry.FailoverTask
+		// change route
+		fsm.failover(task.OldMasterId, task.NewMasterId)
+		slots := fsm.Node2Slot[task.OldMasterId]
+		fsm.addSlots(task.NewMasterId, slots)
+		fsm.removeSlots(task.OldMasterId, slots)
+		delete(fsm.Failovers, task.ID)
+	} else if entry.Event == EventJoin {
+		task := entry.JoinTask
+		fsm.addNode(task.NodeId, task.Master)
+	}
+	if fsm.changed != nil {
+		fsm.changed(fsm)
 	}
-
 	return nil
 }
 
-func (fsm *FSM) addSlots(nodeID string, slots []uint32) {
-	for _, slotId := range slots {
-		/// update node2Slot
-		index := sort.Search(len(fsm.Node2Slot[nodeID]), func(i int) bool {
-			return fsm.Node2Slot[nodeID][i] >= slotId
-		})
-		if !(index < len(fsm.Node2Slot[nodeID]) && fsm.Node2Slot[nodeID][index] == slotId) {
-			// not found in node's slots, insert
-			fsm.Node2Slot[nodeID] = append(fsm.Node2Slot[nodeID][:index],
-				append([]uint32{slotId}, fsm.Node2Slot[nodeID][index:]...)...)
-		}
-		/// update slot2Node
-		fsm.Slot2Node[slotId] = nodeID
-	}
-}
-
-func (fsm *FSM) removeSlots(nodeID string, slots []uint32) {
-	for _, slotId := range slots {
-		/// update node2slot
-		index := sort.Search(len(fsm.Node2Slot[nodeID]), func(i int) bool { return fsm.Node2Slot[nodeID][i] >= slotId })
-		// found slot remove
-		for index < len(fsm.Node2Slot[nodeID]) && fsm.Node2Slot[nodeID][index] == slotId {
-			fsm.Node2Slot[nodeID] = append(fsm.Node2Slot[nodeID][:index], fsm.Node2Slot[nodeID][index+1:]...)
-		}
-		// update slot2node
-		if fsm.Slot2Node[slotId] == nodeID {
-			delete(fsm.Slot2Node, slotId)
-		}
-	}
-}
-
-func (fsm *FSM) GetMigratingTask(taskId string) *MigratingTask {
-	fsm.mu.RLock()
-	defer fsm.mu.RUnlock()
-	return fsm.Migratings[taskId]
-}
-
 // FSMSnapshot stores necessary data to restore FSM
 type FSMSnapshot struct {
-	Slot2Node  map[uint32]string // slotID -> nodeID
-	Migratings map[string]*MigratingTask
+	Slot2Node    map[uint32]string // slotID -> nodeID
+	Migratings   map[string]*MigratingTask
+	MasterSlaves map[string]*MasterSlave
 }
 
 func (snapshot *FSMSnapshot) Persist(sink raft.SnapshotSink) error {
@@ -161,9 +168,14 @@ func (fsm *FSM) Snapshot() (raft.FSMSnapshot, error) {
 	for k, v := range fsm.Migratings {
 		migratings[k] = v
 	}
+	masterSlaves := make(map[string]*MasterSlave)
+	for k, v := range fsm.MasterSlaves {
+		masterSlaves[k] = v
+	}
 	return &FSMSnapshot{
-		Slot2Node:  slot2Node,
-		Migratings: migratings,
+		Slot2Node:    slot2Node,
+		Migratings:   migratings,
+		MasterSlaves: masterSlaves,
 	}, nil
 }
 
@@ -181,23 +193,18 @@ func (fsm *FSM) Restore(src io.ReadCloser) error {
 	}
 	fsm.Slot2Node = snapshot.Slot2Node
 	fsm.Migratings = snapshot.Migratings
+	fsm.MasterSlaves = snapshot.MasterSlaves
 	fsm.Node2Slot = make(map[string][]uint32)
 	for slot, node := range snapshot.Slot2Node {
 		fsm.Node2Slot[node] = append(fsm.Node2Slot[node], slot)
 	}
+	for master, slaves := range snapshot.MasterSlaves {
+		for _, slave := range slaves.Slaves {
+			fsm.SlaveMasters[slave] = master
+		}
+	}
+	if fsm.changed != nil {
+		fsm.changed(fsm)
+	}
 	return nil
 }
-
-// PickNode returns node hosting slot, ignore migrating
-func (fsm *FSM) PickNode(slot uint32) string {
-	fsm.mu.RLock()
-	defer fsm.mu.RUnlock()
-	return fsm.Slot2Node[slot]
-}
-
-// WithReadLock allow invoker do something complicated with read lock
-func (fsm *FSM) WithReadLock(fn func(fsm *FSM)) {
-	fsm.mu.RLock()
-	defer fsm.mu.RUnlock()
-	fn(fsm)
-}

cluster/raft/fsm_utils.go

@@ -0,0 +1,114 @@
+package raft
+
+import (
+	"errors"
+	"sort"
+)
+
+// PickNode returns node hosting slot, ignore migrating
+func (fsm *FSM) PickNode(slot uint32) string {
+	fsm.mu.RLock()
+	defer fsm.mu.RUnlock()
+	return fsm.Slot2Node[slot]
+}
+
+// WithReadLock allow invoker do something complicated with read lock
+func (fsm *FSM) WithReadLock(fn func(fsm *FSM)) {
+	fsm.mu.RLock()
+	defer fsm.mu.RUnlock()
+	fn(fsm)
+}
+
+func (fsm *FSM) GetMigratingTask(taskId string) *MigratingTask {
+	fsm.mu.RLock()
+	defer fsm.mu.RUnlock()
+	return fsm.Migratings[taskId]
+}
+
+func (fsm *FSM) addSlots(nodeID string, slots []uint32) {
+	for _, slotId := range slots {
+		/// update node2Slot
+		index := sort.Search(len(fsm.Node2Slot[nodeID]), func(i int) bool {
+			return fsm.Node2Slot[nodeID][i] >= slotId
+		})
+		if !(index < len(fsm.Node2Slot[nodeID]) && fsm.Node2Slot[nodeID][index] == slotId) {
+			// not found in node's slots, insert
+			fsm.Node2Slot[nodeID] = append(fsm.Node2Slot[nodeID][:index],
+				append([]uint32{slotId}, fsm.Node2Slot[nodeID][index:]...)...)
+		}
+		/// update slot2Node
+		fsm.Slot2Node[slotId] = nodeID
+	}
+}
+
+func (fsm *FSM) removeSlots(nodeID string, slots []uint32) {
+	for _, slotId := range slots {
+		/// update node2slot
+		index := sort.Search(len(fsm.Node2Slot[nodeID]), func(i int) bool { return fsm.Node2Slot[nodeID][i] >= slotId })
+		// found slot remove
+		for index < len(fsm.Node2Slot[nodeID]) && fsm.Node2Slot[nodeID][index] == slotId {
+			fsm.Node2Slot[nodeID] = append(fsm.Node2Slot[nodeID][:index], fsm.Node2Slot[nodeID][index+1:]...)
+		}
+		// update slot2node
+		if fsm.Slot2Node[slotId] == nodeID {
+			delete(fsm.Slot2Node, slotId)
+		}
+	}
+}
+
+func (fsm *FSM) failover(oldMasterId, newMasterId string) {
+	oldSlaves := fsm.MasterSlaves[oldMasterId].Slaves
+	newSlaves := make([]string, 0, len(oldSlaves))
+	// change other slaves
+	for _, slave := range oldSlaves {
+		if slave != newMasterId {
+			fsm.SlaveMasters[slave] = newMasterId
+			newSlaves = append(newSlaves, slave)
+		}
+	}
+	// change old master
+	delete(fsm.MasterSlaves, oldMasterId)
+	fsm.SlaveMasters[oldMasterId] = newMasterId
+	newSlaves = append(newSlaves, oldMasterId)
+
+	// change new master
+	delete(fsm.SlaveMasters, newMasterId)
+	fsm.MasterSlaves[newMasterId] = &MasterSlave{
+		MasterId: newMasterId,
+		Slaves:   newSlaves,
+	}
+}
+
+// getMaster returns "" if id points to a master node
+func (fsm *FSM) getMaster(id string) string {
+	master := ""
+	fsm.WithReadLock(func(fsm *FSM) {
+		master = fsm.SlaveMasters[id]
+	})
+	return master
+}
+
+func (fsm *FSM) addNode(id, masterId string) error {
+	if masterId == "" {
+		fsm.MasterSlaves[id] = &MasterSlave{
+			MasterId: id,
+		}
+	} else {
+		master := fsm.MasterSlaves[masterId]
+		if master == nil {
+			return errors.New("master not found")
+		}
+		exists := false
+		for _, slave := range master.Slaves {
+			if slave == id {
+				exists = true
+				break
+			}
+		}
+		if !exists {
+			master.Slaves = append(master.Slaves, id)
+		}
+		fsm.SlaveMasters[id] = masterId
+	}
+	return nil
+}

cluster/raft/raft.go

@@ -7,7 +7,6 @@ import (
 	"net"
 	"os"
 	"path/filepath"
-	"strconv"
 	"time"
 
 	"github.com/hashicorp/raft"
@@ -22,6 +21,13 @@ type Node struct {
 	stableStore   raft.StableStore
 	snapshotStore raft.SnapshotStore
 	transport     raft.Transport
+	watcher       watcher
+}
+
+type watcher struct {
+	watch         func(*FSM)
+	currentMaster string
+	onFailover    func(newMaster string)
 }
 
 type RaftConfig struct {
@@ -74,9 +80,12 @@ func StartNode(cfg *RaftConfig) (*Node, error) {
 	}
 
 	storage := &FSM{
-		Node2Slot:  make(map[string][]uint32),
-		Slot2Node:  make(map[uint32]string),
-		Migratings: make(map[string]*MigratingTask),
+		Node2Slot:    make(map[string][]uint32),
+		Slot2Node:    make(map[uint32]string),
+		Migratings:   make(map[string]*MigratingTask),
+		MasterSlaves: make(map[string]*MasterSlave),
+		SlaveMasters: make(map[string]string),
+		Failovers:    make(map[string]*FailoverTask),
 	}
 
 	logStore := boltDB
@@ -85,8 +94,7 @@ func StartNode(cfg *RaftConfig) (*Node, error) {
 	if err != nil {
 		return nil, err
 	}
-
-	return &Node{
+	node := &Node{
 		Cfg:           cfg,
 		inner:         inner,
 		FSM:           storage,
@@ -94,7 +102,9 @@ func StartNode(cfg *RaftConfig) (*Node, error) {
 		stableStore:   stableStore,
 		snapshotStore: snapshotStore,
 		transport:     transport,
-	}, nil
+	}
+	node.setupWatch()
+	return node, nil
 }
 
 func (node *Node) HasExistingState() (bool, error) {
@@ -130,37 +140,13 @@ func (node *Node) BootstrapCluster(slotCount int) error {
 	return err
 }
 
-func (node *Node) Shutdown() error {
+func (node *Node) Close() error {
 	future := node.inner.Shutdown()
-	return future.Error()
+	return fmt.Errorf("raft shutdown %v", future.Error())
 }
 
-func (node *Node) State() raft.RaftState {
-	return node.inner.State()
-}
-
-func (node *Node) CommittedIndex() (uint64, error) {
-	stats := node.inner.Stats()
-	committedIndex0 := stats["commit_index"]
-	return strconv.ParseUint(committedIndex0, 10, 64)
-}
-
-func (node *Node) GetLeaderRedisAddress() string {
-	// redis advertise address used as leader id
-	_, id := node.inner.LeaderWithID()
-	return string(id)
-}
-
-func (node *Node) GetNodes() ([]raft.Server, error) {
-	configFuture := node.inner.GetConfiguration()
-	if err := configFuture.Error(); err != nil {
-		return nil, fmt.Errorf("failed to get raft configuration: %v", err)
-	}
-	return configFuture.Configuration().Servers, nil
-}
-
-// HandleJoin handles join request, node must be leader
-func (node *Node) HandleJoin(redisAddr, raftAddr string) error {
+// AddToRaft handles join request, node must be leader
+func (node *Node) AddToRaft(redisAddr, raftAddr string) error {
 	configFuture := node.inner.GetConfiguration()
 	if err := configFuture.Error(); err != nil {
 		return fmt.Errorf("failed to get raft configuration: %v", err)
@@ -175,6 +161,23 @@ func (node *Node) HandleJoin(redisAddr, raftAddr string) error {
 	return future.Error()
 }
 
+func (node *Node) HandleEvict(redisAddr string) error {
+	configFuture := node.inner.GetConfiguration()
+	if err := configFuture.Error(); err != nil {
+		return fmt.Errorf("failed to get raft configuration: %v", err)
+	}
+	id := raft.ServerID(redisAddr)
+	for _, srv := range configFuture.Configuration().Servers {
+		if srv.ID == id {
+			err := node.inner.RemoveServer(srv.ID, 0, 0).Error()
+			if err != nil {
+				return err
+			}
+		}
+	}
+	return nil
+}
+
 func (node *Node) Propose(event *LogEntry) (uint64, error) {
 	bin, err := json.Marshal(event)
 	if err != nil {
@@ -188,7 +191,21 @@ func (node *Node) Propose(event *LogEntry) (uint64, error) {
 	return future.Index(), nil
 }
 
-func (node *Node) Close() error {
-	future := node.inner.Shutdown()
-	return fmt.Errorf("raft shutdown %v", future.Error())
-}
+func (node *Node) setupWatch() {
+	node.watcher.watch = func(f *FSM) {
+		newMaster := f.SlaveMasters[node.Self()]
+		if newMaster != node.watcher.currentMaster {
+			node.watcher.currentMaster = newMaster
+			if node.watcher.onFailover != nil {
+				node.watcher.onFailover(newMaster)
+			}
+		}
+	}
+}
+
+// SetOnFailover sets onFailover callback
+// After a failover, onFailover will receive the new master
+func (node *Node) SetOnFailover(fn func(newMaster string)) {
+	node.watcher.currentMaster = node.FSM.getMaster(node.Self())
+	node.watcher.onFailover = fn
+}

cluster/raft/utils.go

@@ -0,0 +1,42 @@
+package raft
+
+import (
+	"fmt"
+	"strconv"
+
+	"github.com/hashicorp/raft"
+)
+
+func (node *Node) Self() string {
+	return node.Cfg.ID()
+}
+
+func (node *Node) State() raft.RaftState {
+	return node.inner.State()
+}
+
+func (node *Node) CommittedIndex() (uint64, error) {
+	stats := node.inner.Stats()
+	committedIndex0 := stats["commit_index"]
+	return strconv.ParseUint(committedIndex0, 10, 64)
+}
+
+func (node *Node) GetLeaderRedisAddress() string {
+	// redis advertise address used as leader id
+	_, id := node.inner.LeaderWithID()
+	return string(id)
+}
+
+func (node *Node) GetNodes() ([]raft.Server, error) {
+	configFuture := node.inner.GetConfiguration()
+	if err := configFuture.Error(); err != nil {
+		return nil, fmt.Errorf("failed to get raft configuration: %v", err)
+	}
+	return configFuture.Configuration().Servers, nil
+}
+
+func (node *Node) GetSlaves(id string) *MasterSlave {
+	node.FSM.mu.RLock()
+	defer node.FSM.mu.RUnlock()
+	return node.FSM.MasterSlaves[id]
+}

config/config.go

@@ -46,6 +46,7 @@ type ServerProperties struct {
 	ClusterSeed       string `cfg:"cluster-seed"`
 	RaftListenAddr    string `cfg:"raft-listen-address"`
 	RaftAdvertiseAddr string `cfg:"raft-advertise-address"`
+	MasterInCluster   string `cfg:"master-in-cluster"`
 	// config file path
 	CfPath string `cfg:"cf,omitempty"`
 }

database/persistence.go

@@ -105,8 +105,8 @@ func (server *Server) AddAof(dbIndex int, cmdLine CmdLine) {
 	}
 }
 
-func (server *Server) bindPersister(aofHandler *aof.Persister) {
-	server.persister = aofHandler
+func (server *Server) bindPersister(persister *aof.Persister) {
+	server.persister = persister
 	// bind SaveCmdLine
 	for _, db := range server.dbSet {
 		singleDB := db.Load().(*DB)

database/replication_master.go

@@ -311,7 +311,7 @@ func (server *Server) execPSync(c redis.Connection, args [][]byte) redis.Reply {
 			if err == nil {
 				return
 			}
-			if err != nil && err != cannotPartialSync {
+			if err != cannotPartialSync {
 				server.removeSlave(slave)
 				logger.Errorf("masterTryPartialSyncWithSlave error: %v", err)
 				return
@@ -422,7 +422,7 @@ func (listener *replAofListener) Callback(cmdLines []CmdLine) {
 	}
 }
 
-func (server *Server) initMaster() {
+func (server *Server) initMasterStatus() {
 	server.masterStatus = &masterStatus{
 		mu:           sync.RWMutex{},
 		replId:       utils.RandHexString(40),

database/replication_master_test.go

@@ -11,13 +11,14 @@ import (
 	"testing"
 	"time"
 
+	"github.com/hdt3213/godis/aof"
 	"github.com/hdt3213/godis/config"
-	rdb "github.com/hdt3213/rdb/parser"
 	"github.com/hdt3213/godis/lib/utils"
 	"github.com/hdt3213/godis/redis/connection"
 	"github.com/hdt3213/godis/redis/parser"
 	"github.com/hdt3213/godis/redis/protocol"
 	"github.com/hdt3213/godis/redis/protocol/asserts"
+	rdb "github.com/hdt3213/rdb/parser"
 )
 
 func mockServer() *Server {
@@ -31,7 +32,7 @@ func mockServer() *Server {
 		server.dbSet[i] = holder
 	}
 	server.slaveStatus = initReplSlaveStatus()
-	server.initMaster()
+	server.initMasterStatus()
 	return server
 }
 
@@ -212,6 +213,7 @@ func TestReplicationMasterRewriteRDB(t *testing.T) {
 		Databases:      16,
 		AppendOnly:     true,
 		AppendFilename: aofFilename,
+		AppendFsync:    aof.FsyncAlways,
 	}
 	master := mockServer()
 	aofHandler, err := NewPersister(master, config.Properties.AppendFilename, true, config.Properties.AppendFsync)

database/replication_slave.go

@@ -272,6 +272,11 @@ func (server *Server) psyncHandshake() (bool, error) {
 	if err != nil {
 		return false, errors.New("send failed " + err.Error())
 	}
+	return server.parsePsyncHandshake()
+}
+
+func (server *Server) parsePsyncHandshake() (bool, error) {
+	var err error
 	psyncPayload := <-server.slaveStatus.masterChan
 	if psyncPayload.Err != nil {
 		return false, errors.New("read response failed: " + psyncPayload.Err.Error())

database/replication_slave_test.go

@@ -2,18 +2,21 @@ package database
 
 import (
 	"bytes"
-	"github.com/hdt3213/godis/aof"
-	"github.com/hdt3213/godis/config"
-	"github.com/hdt3213/godis/lib/utils"
-	"github.com/hdt3213/godis/redis/client"
-	"github.com/hdt3213/godis/redis/connection"
-	"github.com/hdt3213/godis/redis/protocol"
-	"github.com/hdt3213/godis/redis/protocol/asserts"
+	"context"
 	"io/ioutil"
 	"os"
 	"path"
 	"testing"
 	"time"
+
+	"github.com/hdt3213/godis/aof"
+	"github.com/hdt3213/godis/config"
+	"github.com/hdt3213/godis/lib/utils"
+	"github.com/hdt3213/godis/redis/client"
+	"github.com/hdt3213/godis/redis/connection"
+	"github.com/hdt3213/godis/redis/parser"
+	"github.com/hdt3213/godis/redis/protocol"
+	"github.com/hdt3213/godis/redis/protocol/asserts"
 )
 
 func TestReplicationSlaveSide(t *testing.T) {
@@ -151,12 +154,15 @@ func TestReplicationSlaveSide(t *testing.T) {
 	}
 
 	// check slave aof file
-	aofLoader := MakeAuxiliaryServer()
-	aofHandler2, err := NewPersister(aofLoader, config.Properties.AppendFilename, true, aof.FsyncNo)
-	aofLoader.bindPersister(aofHandler2)
-	ret = aofLoader.Exec(conn, utils.ToCmdLine("get", "zz"))
+	aofCheckServer := MakeAuxiliaryServer()
+	aofHandler2, err := NewPersister(aofCheckServer, config.Properties.AppendFilename, true, aof.FsyncNo)
+	if err != nil {
+		t.Error("create persister failed")
+	}
+	aofCheckServer.bindPersister(aofHandler2)
+	ret = aofCheckServer.Exec(conn, utils.ToCmdLine("get", "zz"))
 	asserts.AssertNullBulk(t, ret)
-	ret = aofLoader.Exec(conn, utils.ToCmdLine("get", "1"))
+	ret = aofCheckServer.Exec(conn, utils.ToCmdLine("get", "1"))
 	asserts.AssertBulkReply(t, ret, "4")
 
 	err = server.slaveStatus.close()
@@ -164,3 +170,82 @@ func TestReplicationSlaveSide(t *testing.T) {
 		t.Error("cannot close")
 	}
 }
+
+func TestReplicationFailover(t *testing.T) {
+	tmpDir, err := os.MkdirTemp("", "godis")
+	if err != nil {
+		t.Error(err)
+		return
+	}
+	aofFilename := path.Join(tmpDir, "a.aof")
+	defer func() {
+		_ = os.Remove(aofFilename)
+	}()
+	config.Properties = &config.ServerProperties{
+		Databases:      16,
+		AppendOnly:     true,
+		AppendFilename: aofFilename,
+	}
+	conn := connection.NewFakeConn()
+	server := mockServer()
+	aofHandler, err := NewPersister(server, aofFilename, true, aof.FsyncAlways)
+	if err != nil {
+		t.Error(err)
+		return
+	}
+	server.bindPersister(aofHandler)
+	
+	masterCli, err := client.MakeClient("127.0.0.1:6379")
+	if err != nil {
+		t.Error(err)
+		return
+	}
+	masterCli.Start()
+
+	// sync with master
+	ret := masterCli.Send(utils.ToCmdLine("set", "1", "1"))
+	asserts.AssertStatusReply(t, ret, "OK")
+	ret = server.Exec(conn, utils.ToCmdLine("SLAVEOF", "127.0.0.1", "6379"))
+	asserts.AssertStatusReply(t, ret, "OK")
+	success := false
+	for i := 0; i < 30; i++ {
+		// wait for sync
+		time.Sleep(time.Second)
+		ret = server.Exec(conn, utils.ToCmdLine("GET", "1"))
+		bulkRet, ok := ret.(*protocol.BulkReply)
+		if ok {
+			if bytes.Equal(bulkRet.Arg, []byte("1")) {
+				success = true
+				break
+			}
+		}
+	}
+	if !success {
+		t.Error("sync failed")
+		return
+	}
+
+	t.Log("slave of no one")
+	ret = server.Exec(conn, utils.ToCmdLine("SLAVEOF", "no", "one"))
+	asserts.AssertStatusReply(t, ret, "OK")
+	server.Exec(conn, utils.ToCmdLine("set", "2", "2"))
+
+	replConn := connection.NewFakeConn()
+	server.Exec(replConn, utils.ToCmdLine("psync", "?", "-1"))
+	masterChan := parser.ParseStream(replConn)
+	serverB := mockServer()
+	serverB.slaveStatus.masterChan = masterChan
+	serverB.slaveStatus.configVersion = 0
+	serverB.parsePsyncHandshake()
+	serverB.loadMasterRDB(0)
+	server.masterCron()
+	ctx, cancel := context.WithTimeout(context.Background(), 2*time.Second)
+    defer cancel()
+    go serverB.receiveAOF(ctx, 0)
+
+    time.Sleep(3 * time.Second)
+	ret = serverB.Exec(conn, utils.ToCmdLine("get", "1"))
+	asserts.AssertBulkReply(t, ret, "1")
+	ret = serverB.Exec(conn, utils.ToCmdLine("get", "2"))
+	asserts.AssertBulkReply(t, ret, "2")
+}

database/server.go

@@ -85,7 +85,7 @@ func NewStandaloneServer() *Server {
 		}
 	}
 	server.slaveStatus = initReplSlaveStatus()
-	server.initMaster()
+	server.initMasterStatus()
 	server.startReplCron()
 	server.role = masterRole // The initialization process does not require atomicity
 	return server

lib/pool/pool.go

@@ -42,6 +42,7 @@ func New(factory func() (interface{}, error), finalizer func(x interface{}), cfg
 // getOnNoIdle try to create a new connection or waiting for connection being returned
 // invoker should have pool.mu
 func (pool *Pool) getOnNoIdle() (interface{}, error) {
+	pool.mu.Lock()
 	if pool.activeCount >= pool.MaxActive {
 		// waiting for connection being returned
 		req := make(chan interface{}, 1)
@@ -74,10 +75,10 @@ func (pool *Pool) Get() (interface{}, error) {
 		pool.mu.Unlock()
 		return nil, ErrClosed
 	}
+	pool.mu.Unlock()
 
 	select {
 	case item := <-pool.idles:
-		pool.mu.Unlock()
 		return item, nil
 	default:
 		// no pooled item, create one

redis/connection/conn.go

@@ -57,7 +57,9 @@ func (c *Connection) RemoteAddr() string {
 // Close disconnect with the client
 func (c *Connection) Close() error {
 	c.sendingData.WaitWithTimeout(10 * time.Second)
-	_ = c.conn.Close()
+	if c.conn != nil { // may be a fake conn for tests
+		_ = c.conn.Close()
+	}
 	c.subs = nil
 	c.password = ""
 	c.queue = nil
@@ -219,6 +221,7 @@ func (c *Connection) IsSlave() bool {
 	return c.flags&flagSlave > 0
 }
 
+// SetMaster marks c as a connection with master
 func (c *Connection) SetMaster() {
 	c.flags |= flagMaster
 }