raft cluster

wip: raft does not care about migrating

wip: optimize code

wip: raft election

wip

wip: fix raft leader missing log entries

wip

fix a dead lock

batch set slot route

wip: raft persist

wip

refactor cluster suite

remove relay

rename relay2

refactor: allow customizing client factory

test raft

refactor re-balance

avoid errors caused by inconsistent status on follower nodes during raft commits

test raft election

cluster/topo_gcluster.go

@@ -0,0 +1,146 @@
+package cluster
+
+import (
+	"fmt"
+	"github.com/hdt3213/godis/aof"
+	"github.com/hdt3213/godis/interface/redis"
+	"github.com/hdt3213/godis/lib/logger"
+	"github.com/hdt3213/godis/redis/protocol"
+	"strconv"
+	"strings"
+)
+
+func init() {
+	registerCmd("gcluster", execGCluster)
+}
+
+func execGCluster(cluster *Cluster, c redis.Connection, args [][]byte) redis.Reply {
+	if len(args) < 2 {
+		return protocol.MakeArgNumErrReply("gcluster")
+	}
+	subCmd := strings.ToLower(string(args[1]))
+	switch subCmd {
+	case "set-slot":
+		// Command line: gcluster set-slot <slotID> <targetNodeID>
+		// Other node request current node to migrate a slot to it.
+		// Current node will set the slot as migrating state.
+		// After this function return, all requests of target slot will be routed to target node
+		return execGClusterSetSlot(cluster, c, args[2:])
+	case "migrate":
+		// Command line: gcluster migrate <slotId>
+		// Current node will  dump the given slot to the node sending this request
+		// The given slot must in migrating state
+		return execGClusterMigrate(cluster, c, args[2:])
+	case "migrate-done":
+		// command line: gcluster migrate-done <slotId>
+		// The new node hosting given slot tells current node that migration has finished, remains data can be deleted
+		return execGClusterMigrateDone(cluster, c, args[2:])
+	case "request-donate":
+		// command line: gcluster donate <nodeID>
+		// picks some slots and gives them to the calling node for load balance
+		return execGClusterDonateSlot(cluster, c, args[2:])
+	}
+	return protocol.MakeErrReply(" ERR unknown gcluster sub command '" + subCmd + "'")
+}
+
+// execGClusterSetSlot set a current node hosted slot as migrating
+// args is [slotID, newNodeId]
+func execGClusterSetSlot(cluster *Cluster, c redis.Connection, args [][]byte) redis.Reply {
+	if len(args) != 2 {
+		return protocol.MakeArgNumErrReply("gcluster")
+	}
+	slotId0, err := strconv.Atoi(string(args[0]))
+	if err != nil || slotId0 >= slotCount {
+		return protocol.MakeErrReply("ERR value is not a valid slot id")
+	}
+	slotId := uint32(slotId0)
+	targetNodeID := string(args[1])
+	targetNode := cluster.topology.GetNode(targetNodeID)
+	if targetNode == nil {
+		return protocol.MakeErrReply("ERR node not found")
+	}
+	cluster.setSlotMovingOut(slotId, targetNodeID)
+	logger.Info(fmt.Sprintf("set slot %d to node %s", slotId, targetNodeID))
+	return protocol.MakeOkReply()
+}
+
+// execGClusterDonateSlot picks some slots and gives them to the calling node for load balance
+// args is [callingNodeId]
+func execGClusterDonateSlot(cluster *Cluster, c redis.Connection, args [][]byte) redis.Reply {
+	targetNodeID := string(args[0])
+	nodes := cluster.topology.GetNodes() // including the new node
+	avgSlot := slotCount / len(nodes)
+	cluster.slotMu.Lock()
+	defer cluster.slotMu.Unlock()
+	limit := len(cluster.slots) - avgSlot
+	if limit <= 0 {
+		return protocol.MakeEmptyMultiBulkReply()
+	}
+	result := make([][]byte, 0, limit)
+	// use the randomness of the for-each-in-map to randomly select slots
+	for slotID, slot := range cluster.slots {
+		if slot.state == slotStateHost {
+			slot.state = slotStateMovingOut
+			slot.newNodeID = targetNodeID
+			slotIDBin := []byte(strconv.FormatUint(uint64(slotID), 10))
+			result = append(result, slotIDBin)
+			if len(result) == limit {
+				break
+			}
+		}
+	}
+	return protocol.MakeMultiBulkReply(result)
+}
+
+// execGClusterMigrate Command line: gcluster migrate slotId
+// Current node will  dump data in the given slot to the node sending this request
+// The given slot must in migrating state
+func execGClusterMigrate(cluster *Cluster, c redis.Connection, args [][]byte) redis.Reply {
+	slotId0, err := strconv.Atoi(string(args[0]))
+	if err != nil || slotId0 >= slotCount {
+		return protocol.MakeErrReply("ERR value is not a valid slot id")
+	}
+	slotId := uint32(slotId0)
+	slot := cluster.getHostSlot(slotId)
+	if slot == nil || slot.state != slotStateMovingOut {
+		return protocol.MakeErrReply("ERR only dump migrating slot")
+	}
+	// migrating slot is immutable
+	logger.Info("start dump slot", slotId)
+	slot.keys.ForEach(func(key string) bool {
+		entity, ok := cluster.db.GetEntity(0, key)
+		if ok {
+			ret := aof.EntityToCmd(key, entity)
+			// todo: handle error and close connection
+			_, _ = c.Write(ret.ToBytes())
+			expire := cluster.db.GetExpiration(0, key)
+			if expire != nil {
+				ret = aof.MakeExpireCmd(key, *expire)
+				_, _ = c.Write(ret.ToBytes())
+			}
+
+		}
+		return true
+	})
+	logger.Info("finish dump slot ", slotId)
+	// send a ok reply to tell requesting node dump finished
+	return protocol.MakeOkReply()
+}
+
+// execGClusterMigrateDone command line: gcluster migrate-done <slotId>
+func execGClusterMigrateDone(cluster *Cluster, c redis.Connection, args [][]byte) redis.Reply {
+	slotId0, err := strconv.Atoi(string(args[0]))
+	if err != nil || slotId0 >= slotCount {
+		return protocol.MakeErrReply("ERR value is not a valid slot id")
+	}
+	slotId := uint32(slotId0)
+	slot := cluster.getHostSlot(slotId)
+	if slot == nil || slot.state != slotStateMovingOut {
+		return protocol.MakeErrReply("ERR slot is not moving out")
+	}
+	cluster.cleanDroppedSlot(slotId)
+	cluster.slotMu.Lock()
+	delete(cluster.slots, slotId)
+	cluster.slotMu.Unlock()
+	return protocol.MakeOkReply()
+}