研发效能 zookeeper 的 Leader 选举源码解析

京东云开发者 · 2023年03月29日 · 3153 次阅读

作者:京东物流 梁吉超

zookeeper 是一个分布式服务框架,主要解决分布式应用中常见的多种数据问题,例如集群管理,状态同步等。为解决这些问题 zookeeper 需要 Leader 选举进行保障数据的强一致性机制和稳定性。本文通过集群的配置,对 leader 选举源进行解析,让读者们了解如何利用 BIO 通信机制,多线程多层队列实现高性能架构。****

01Leader 选举机制

Leader 选举机制采用半数选举算法。

每一个 zookeeper 服务端称之为一个节点,每


个节点都有投票权,把其选票投向每一个有选举权的节点,当其中一个节点选举出票数过半,这个节点就会成为 Leader,其它节点成为 Follower。

02Leader 选举集群配置

  1. 重命名 zoo_sample.cfg 文件为 zoo1.cfg ,zoo2.cfg,zoo3.cfg,zoo4.cfg

  2. 修改 zoo.cfg 文件,修改值如下:

【plain】
zoo1.cfg文件内容:
dataDir=/export/data/zookeeper-1
clientPort=2181
server.1=127.0.0.1:2001:3001
server.2=127.0.0.1:2002:3002:participant
server.3=127.0.0.1:2003:3003:participant
server.4=127.0.0.1:2004:3004:observer


zoo2.cfg文件内容:
dataDir=/export/data/zookeeper-2
clientPort=2182
server.1=127.0.0.1:2001:3001
server.2=127.0.0.1:2002:3002:participant
server.3=127.0.0.1:2003:3003:participant
server.4=127.0.0.1:2004:3004:observer


zoo3.cfg文件内容:
dataDir=/export/data/zookeeper-3
clientPort=2183
server.1=127.0.0.1:2001:3001
server.2=127.0.0.1:2002:3002:participant
server.3=127.0.0.1:2003:3003:participant
server.4=127.0.0.1:2004:3004:observer


zoo4.cfg文件内容:
dataDir=/export/data/zookeeper-4
clientPort=2184
server.1=127.0.0.1:2001:3001
server.2=127.0.0.1:2002:3002:participant
server.3=127.0.0.1:2003:3003:participant
server.4=127.0.0.1:2004:3004:observer





  1. server.第几号服务器 (对应 myid 文件内容)=ip:数据同步端口:选举端口:选举标识
  • participant 默认参与选举标识,可不写. observer 不参与选举

4.在/export/data/zookeeper-1,/export/data/zookeeper-2,/export/data/zookeeper-3,/export/data/zookeeper-4 目录下创建 myid 文件,文件内容分别写 1 ,2,3,4,用于标识 sid(全称:Server ID) 赋值。

  1. 启动三个 zookeeper 实例:
  • bin/zkServer.sh start conf/zoo1.cfg
  • bin/zkServer.sh start conf/zoo2.cfg
  • bin/zkServer.sh start conf/zoo3.cfg
  1. 每启动一个实例,都会读取启动参数配置 zoo.cfg 文件,这样实例就可以知道其作为服务端身份信息 sid 以及集群中有多少个实例参与选举。

03Leader 选举流程

图 1 第一轮到第二轮投票流程

前提:

设定票据数据格式 vote(sid,zxid,epoch)

  • sid 是 Server ID 每台服务的唯一标识,是 myid 文件内容;
  • zxid 是数据事务 id 号;
  • epoch 为选举周期,为方便理解下面讲解内容暂定为 1 初次选举,不写入下面内容里。

按照顺序启动 sid=1,sid=2 节点

第一轮投票:

  1. sid=1 节点:初始选票为自己,将选票 vote(1,0) 发送给 sid=2 节点;

  2. sid=2 节点:初始选票为自己,将选票 vote(2,0) 发送给 sid=1 节点;

  3. sid=1 节点:收到 sid=2 节点选票 vote(2,0) 和当前自己的选票 vote(1,0),首先比对 zxid 值,zxid 越大代表数据最新,优先选择 zxid 最大的选票,如果 zxid 相同,选举最大 sid。当前投票选举结果为 vote(2,0),sid=1 节点的选票变为 vote(2,0);

  4. sid=2 节点:收到 sid=1 节点选票 vote(1,0) 和当前自己的选票 vote(2,0),参照上述选举方式,选举结果为 vote(2,0),sid=2 节点的选票不变;

  5. 第一轮投票选举结束。

第二轮投票:

  1. sid=1 节点:当前自己的选票为 vote(2,0),将选票 vote(2,0) 发送给 sid=2 节点;

  2. sid=2 节点:当前自己的选票为 vote(2,0),将选票 vote(2,0) 发送给 sid=1 节点;

  3. sid=1 节点:收到 sid=2 节点选票 vote(2,0) 和自己的选票 vote(2,0), 按照半数选举算法,总共 3 个节点参与选举,已有 2 个节点选举出相同选票,推举 sid=2 节点为 Leader,自己角色变为 Follower;

  4. sid=2 节点:收到 sid=1 节点选票 vote(2,0) 和自己的选票 vote(2,0),按照半数选举算法推举 sid=2 节点为 Leader,自己角色变为 Leader。

这时启动 sid=3 节点后,集群里已经选举出 leader,sid=1 和 sid=2 节点会将自己的 leader 选票发回给 sid=3 节点,通过半数选举结果还是 sid=2 节点为 leader。

3.1 Leader 选举采用多层队列架构

zookeeper 选举底层主要分为选举应用层和消息传输队列层,第一层应用层队列统一接收和发送选票,而第二层传输层队列,是按照服务端 sid 分成了多个队列,是为了避免给每台服务端发送消息互相影响。比如对某台机器发送不成功不会影响正常服务端的发送。

图 2 多层队列上下关系交互流程图

04 解析代码入口类

通过查看 zkServer.sh 文件内容找到服务启动类:

org.apache.zookeeper.server.quorum.QuorumPeerMain

05 选举流程代码解析

图 3 选举代码实现流程图

  1. 加载配置文件 QuorumPeerConfig.parse(path);

针对 Leader 选举关键配置信息如下:

  • 读取 dataDir 目录找到 myid 文件内容,设置当前应用 sid 标识,做为投票人身份信息。下面遇到 myid 变量为当前节点自己 sid 标识。
  • - 设置 peerType 当前应用是否参与选举
  • new QuorumMaj() 解析 server.前缀加载集群成员信息,加载 allMembers 所有成员,votingMembers参与选举成员,observingMembers 观察者成员,设置half值 votingMembers.size()/2.
【Java】
public QuorumMaj(Properties props) throws ConfigException {
        for (Entry<Object, Object> entry : props.entrySet()) {
            String key = entry.getKey().toString();
            String value = entry.getValue().toString();
            //读取集群配置文件中的server.开头的应用实例配置信息
            if (key.startsWith("server.")) {
                int dot = key.indexOf('.');
                long sid = Long.parseLong(key.substring(dot + 1));
                QuorumServer qs = new QuorumServer(sid, value);
                allMembers.put(Long.valueOf(sid), qs);
                if (qs.type == LearnerType.PARTICIPANT)
//应用实例绑定的角色为PARTICIPANT意为参与选举
                    votingMembers.put(Long.valueOf(sid), qs);
                else {
                    //观察者成员
                    observingMembers.put(Long.valueOf(sid), qs);
                }
            } else if (key.equals("version")) {
                version = Long.parseLong(value, 16);
            }
        }
        //过半基数
        half = votingMembers.size() / 2;
    }





  1. QuorumPeerMain.runFromConfig(config) 启动服务;

  2. QuorumPeer.startLeaderElection() 开启选举服务;

  • 设置当前选票 new Vote(sid,zxid,epoch)
【plain】
synchronized public void startLeaderElection(){
try {
           if (getPeerState() == ServerState.LOOKING) {
               //首轮:当前节点默认投票对象为自己
               currentVote = new Vote(myid, getLastLoggedZxid(), getCurrentEpoch());
           }
       } catch(IOException e) {
           RuntimeException re = new RuntimeException(e.getMessage());
           re.setStackTrace(e.getStackTrace());
           throw re;
       }
//........
}





  • 创建选举管理类:QuorumCnxnManager;
  • 初始化 recvQueue接收投票队列 (第二层传输队列);
  • 初始化 queueSendMap按 sid 发送投票队列 (第二层传输队列);
  • 初始化 senderWorkerMap发送投票工作线程容器,表示着与 sid 投票节点已连接;
  • 初始化选举监听线程类 QuorumCnxnManager.Listener。
【Java】
//QuorumPeer.createCnxnManager()
public QuorumCnxManager(QuorumPeer self,
                        final long mySid,
                        Map<Long,QuorumPeer.QuorumServer> view,
                        QuorumAuthServer authServer,
                        QuorumAuthLearner authLearner,
                        int socketTimeout,
                        boolean listenOnAllIPs,
                        int quorumCnxnThreadsSize,
                        boolean quorumSaslAuthEnabled) {
    //接收投票队列(第二层传输队列)
    this.recvQueue = new ArrayBlockingQueue<Message>(RECV_CAPACITY);
    //按sid发送投票队列(第二层传输队列)
    this.queueSendMap = new ConcurrentHashMap<Long, ArrayBlockingQueue<ByteBuffer>>();
    //发送投票工作线程容器,表示着与sid投票节点已连接 
    this.senderWorkerMap = new ConcurrentHashMap<Long, SendWorker>();
    this.lastMessageSent = new ConcurrentHashMap<Long, ByteBuffer>();


    String cnxToValue = System.getProperty("zookeeper.cnxTimeout");
    if(cnxToValue != null){
        this.cnxTO = Integer.parseInt(cnxToValue);
    }


    this.self = self;


    this.mySid = mySid;
    this.socketTimeout = socketTimeout;
    this.view = view;
    this.listenOnAllIPs = listenOnAllIPs;


    initializeAuth(mySid, authServer, authLearner, quorumCnxnThreadsSize,
            quorumSaslAuthEnabled);
    // Starts listener thread that waits for connection requests 
    //创建选举监听线程 接收选举投票请求
    listener = new Listener();
    listener.setName("QuorumPeerListener");
}
//QuorumPeer.createElectionAlgorithm
protected Election createElectionAlgorithm(int electionAlgorithm){
    Election le=null;
    //TODO: use a factory rather than a switch
    switch (electionAlgorithm) {
    case 0:
        le = new LeaderElection(this);
        break;
    case 1:
        le = new AuthFastLeaderElection(this);
        break;
    case 2:
        le = new AuthFastLeaderElection(this, true);
        break;
    case 3:
        qcm = createCnxnManager();// new QuorumCnxManager(... new Listener())
        QuorumCnxManager.Listener listener = qcm.listener;
        if(listener != null){
            listener.start();//启动选举监听线程
            FastLeaderElection fle = new FastLeaderElection(this, qcm);
            fle.start();
            le = fle;
        } else {
            LOG.error("Null listener when initializing cnx manager");
        }
        break;
    default:
        assert false;
    }
return le;}





  1. 开启选举监听线程 QuorumCnxnManager.Listener;
  • 创建 ServerSockket 等待大于自己 sid 节点连接,连接信息存储到 senderWorkerMap;
  • sid>self.sid 才可以连接过来。
【Java】
//上面的listener.start()执行后,选择此方法
public void run() {
    int numRetries = 0;
    InetSocketAddress addr;
    Socket client = null;
    while((!shutdown) && (numRetries < 3)){
        try {
            ss = new ServerSocket();
            ss.setReuseAddress(true);
            if (self.getQuorumListenOnAllIPs()) {
                int port = self.getElectionAddress().getPort();
                addr = new InetSocketAddress(port);
            } else {
                // Resolve hostname for this server in case the
                // underlying ip address has changed.
                self.recreateSocketAddresses(self.getId());
                addr = self.getElectionAddress();
            }
            LOG.info("My election bind port: " + addr.toString());
            setName(addr.toString());
            ss.bind(addr);
            while (!shutdown) {
                client = ss.accept();
                setSockOpts(client);
                LOG.info("Received connection request "
                        + client.getRemoteSocketAddress());
                // Receive and handle the connection request
                // asynchronously if the quorum sasl authentication is
                // enabled. This is required because sasl server
                // authentication process may take few seconds to finish,
                // this may delay next peer connection requests.
                if (quorumSaslAuthEnabled) {
                    receiveConnectionAsync(client);
                } else {
//接收连接信息
                    receiveConnection(client);
                }
                numRetries = 0;
            }
        } catch (IOException e) {
            if (shutdown) {
                break;
            }
            LOG.error("Exception while listening", e);
            numRetries++;
            try {
                ss.close();
                Thread.sleep(1000);
            } catch (IOException ie) {
                LOG.error("Error closing server socket", ie);
            } catch (InterruptedException ie) {
                LOG.error("Interrupted while sleeping. " +
                    "Ignoring exception", ie);
            }
            closeSocket(client);
        }
    }
    LOG.info("Leaving listener");
    if (!shutdown) {
        LOG.error("As I'm leaving the listener thread, "
                + "I won't be able to participate in leader "
                + "election any longer: "
                + self.getElectionAddress());
    } else if (ss != null) {
        // Clean up for shutdown.
        try {
            ss.close();
        } catch (IOException ie) {
            // Don't log an error for shutdown.
            LOG.debug("Error closing server socket", ie);
        }
    }
}


//代码执行路径:receiveConnection()->handleConnection(...)
private void handleConnection(Socket sock, DataInputStream din)
            throws IOException {
//...省略
     if (sid < self.getId()) {
            /*
             * This replica might still believe that the connection to sid is
             * up, so we have to shut down the workers before trying to open a
             * new connection.
             */
            SendWorker sw = senderWorkerMap.get(sid);
            if (sw != null) {
                sw.finish();
            }


            /*
             * Now we start a new connection
             */
            LOG.debug("Create new connection to server: {}", sid);
            closeSocket(sock);


            if (electionAddr != null) {
                connectOne(sid, electionAddr);
            } else {
                connectOne(sid);
            }


        } else { // Otherwise start worker threads to receive data.
            SendWorker sw = new SendWorker(sock, sid);
            RecvWorker rw = new RecvWorker(sock, din, sid, sw);
            sw.setRecv(rw);


            SendWorker vsw = senderWorkerMap.get(sid);


            if (vsw != null) {
                vsw.finish();
            }
  //存储连接信息<sid,SendWorker>
            senderWorkerMap.put(sid, sw);


            queueSendMap.putIfAbsent(sid,
                    new ArrayBlockingQueue<ByteBuffer>(SEND_CAPACITY));


            sw.start();
            rw.start();
     }
}





  1. 创建 FastLeaderElection 快速选举服务;
  • 初始选票发送队列 sendqueue(第一层队列)
  • 初始选票接收队列 recvqueue(第一层队列)
  • 创建线程 WorkerSender
  • 创建线程 WorkerReceiver
【Java】
//FastLeaderElection.starter
private void starter(QuorumPeer self, QuorumCnxManager manager) {
    this.self = self;
    proposedLeader = -1;
    proposedZxid = -1;
    //发送队列sendqueue(第一层队列)
    sendqueue = new LinkedBlockingQueue<ToSend>();
    //接收队列recvqueue(第一层队列)
    recvqueue = new LinkedBlockingQueue<Notification>();
    this.messenger = new Messenger(manager);
}
//new Messenger(manager)
Messenger(QuorumCnxManager manager) {
    //创建线程WorkerSender
    this.ws = new WorkerSender(manager);


    this.wsThread = new Thread(this.ws,
            "WorkerSender[myid=" + self.getId() + "]");
    this.wsThread.setDaemon(true);
    //创建线程WorkerReceiver
    this.wr = new WorkerReceiver(manager);


    this.wrThread = new Thread(this.wr,
            "WorkerReceiver[myid=" + self.getId() + "]");
    this.wrThread.setDaemon(true);
}





  1. 开启 WorkerSender 和 WorkerReceiver 线程。

WorkerSender 线程自旋获取 sendqueue 第一层队列元素

  • sendqueue 队列元素内容为相关选票信息详见 ToSend 类;
  • 首先判断选票 sid 是否和自己 sid 值相同,相等直接放入到 recvQueue 队列中;
  • 不相同将 sendqueue 队列元素转储到 queueSendMap第二层传输队列中。
【Java】//FastLeaderElection.Messenger.WorkerSenderclass WorkerSender extends ZooKeeperThread{
//...
  public void run() {
    while (!stop) {
        try {
            ToSend m = sendqueue.poll(3000, TimeUnit.MILLISECONDS);
            if(m == null) continue;
  //将投票信息发送出去
            process(m);
        } catch (InterruptedException e) {
            break;
        }
    }
    LOG.info("WorkerSender is down");
  }
}
//QuorumCnxManager#toSend
public void toSend(Long sid, ByteBuffer b) {
    /*
     * If sending message to myself, then simply enqueue it (loopback).
     */
    if (this.mySid == sid) {
         b.position(0);
         addToRecvQueue(new Message(b.duplicate(), sid));
        /*
         * Otherwise send to the corresponding thread to send.
         */
    } else {
         /*
          * Start a new connection if doesn't have one already.
          */
         ArrayBlockingQueue<ByteBuffer> bq = new ArrayBlockingQueue<ByteBuffer>(
            SEND_CAPACITY);
         ArrayBlockingQueue<ByteBuffer> oldq = queueSendMap.putIfAbsent(sid, bq);
         //转储到queueSendMap<sid,queue>第二层传输队列中
         if (oldq != null) {
             addToSendQueue(oldq, b);
         } else {
             addToSendQueue(bq, b);
         }
         connectOne(sid);     
    }
}





WorkerReceiver 线程自旋获取 recvQueue 第二层传输队列元素转存到 recvqueue 第一层队列中。

【Java】
//WorkerReceiver
public void run() {
    Message response;
    while (!stop) {
      // Sleeps on receive
      try {
          //自旋获取recvQueue第二层传输队列元素
          response = manager.pollRecvQueue(3000, TimeUnit.MILLISECONDS);
          if(response == null) continue;
          // The current protocol and two previous generations all send at least 28 bytes
          if (response.buffer.capacity() < 28) {
              LOG.error("Got a short response: " + response.buffer.capacity());
              continue;
          }
          //...
  if(self.getPeerState() == QuorumPeer.ServerState.LOOKING){
         //第二层传输队列元素转存到recvqueue第一层队列中
         recvqueue.offer(n);
         //...
      }
    }
//...
}





06 选举核心逻辑

  1. 启动线程 QuorumPeer

开始 Leader 选举投票 makeLEStrategy().lookForLeader();

sendNotifications() 向其它节点发送选票信息,选票信息存储到 sendqueue 队列中。sendqueue 队列由 WorkerSender 线程处理。

【plain】
//QuorunPeer.run
//...
try {
   reconfigFlagClear();
    if (shuttingDownLE) {
       shuttingDownLE = false;
       startLeaderElection();
       }
    //makeLEStrategy().lookForLeader() 发送投票
    setCurrentVote(makeLEStrategy().lookForLeader());
} catch (Exception e) {
    LOG.warn("Unexpected exception", e);
    setPeerState(ServerState.LOOKING);
}  
//...
//FastLeaderElection.lookLeader
public Vote lookForLeader() throws InterruptedException {
//...
  //向其他应用发送投票
sendNotifications();
//...
}


private void sendNotifications() {
    //获取应用节点
    for (long sid : self.getCurrentAndNextConfigVoters()) {
        QuorumVerifier qv = self.getQuorumVerifier();
        ToSend notmsg = new ToSend(ToSend.mType.notification,
                proposedLeader,
                proposedZxid,
                logicalclock.get(),
                QuorumPeer.ServerState.LOOKING,
                sid,
                proposedEpoch, qv.toString().getBytes());
        if(LOG.isDebugEnabled()){
            LOG.debug("Sending Notification: " + proposedLeader + " (n.leader), 0x"  +
                  Long.toHexString(proposedZxid) + " (n.zxid), 0x" + Long.toHexString(logicalclock.get())  +
                  " (n.round), " + sid + " (recipient), " + self.getId() +
                  " (myid), 0x" + Long.toHexString(proposedEpoch) + " (n.peerEpoch)");
        }
        //储存投票信息
        sendqueue.offer(notmsg);
    }
}


class WorkerSender extends ZooKeeperThread {
    //...
    public void run() {
    while (!stop) {
        try {
//提取已储存的投票信息
            ToSend m = sendqueue.poll(3000, TimeUnit.MILLISECONDS);
            if(m == null) continue;


            process(m);
        } catch (InterruptedException e) {
            break;
        }
    }
    LOG.info("WorkerSender is down");
  }
//...
}





自旋 recvqueue 队列元素获取投票过来的选票信息:

【Java】
public Vote lookForLeader() throws InterruptedException {
//...
/*
 * Loop in which we exchange notifications until we find a leader
 */
while ((self.getPeerState() == ServerState.LOOKING) &&
        (!stop)){
    /*
     * Remove next notification from queue, times out after 2 times
     * the termination time
     */
    //提取投递过来的选票信息
    Notification n = recvqueue.poll(notTimeout,
            TimeUnit.MILLISECONDS);
/*
 * Sends more notifications if haven't received enough.
 * Otherwise processes new notification.
 */
if(n == null){
    if(manager.haveDelivered()){
        //已全部连接成功,并且前一轮投票都完成,需要再次发起投票
        sendNotifications();
    } else {
        //如果未收到选票信息,manager.contentAll()自动连接其它socket节点
        manager.connectAll();
    }
    /*
     * Exponential backoff
     */
    int tmpTimeOut = notTimeout*2;
    notTimeout = (tmpTimeOut < maxNotificationInterval?
            tmpTimeOut : maxNotificationInterval);
    LOG.info("Notification time out: " + notTimeout);
         }
     //....
    }
  //...
}





【Java】
//manager.connectAll()->connectOne(sid)->initiateConnection(...)->startConnection(...)


private boolean startConnection(Socket sock, Long sid)
        throws IOException {
    DataOutputStream dout = null;
    DataInputStream din = null;
    try {
        // Use BufferedOutputStream to reduce the number of IP packets. This is
        // important for x-DC scenarios.
        BufferedOutputStream buf = new BufferedOutputStream(sock.getOutputStream());
        dout = new DataOutputStream(buf);


        // Sending id and challenge
        // represents protocol version (in other words - message type)
        dout.writeLong(PROTOCOL_VERSION);
        dout.writeLong(self.getId());
        String addr = self.getElectionAddress().getHostString() + ":" + self.getElectionAddress().getPort();
        byte[] addr_bytes = addr.getBytes();
        dout.writeInt(addr_bytes.length);
        dout.write(addr_bytes);
        dout.flush();


        din = new DataInputStream(
                new BufferedInputStream(sock.getInputStream()));
    } catch (IOException e) {
        LOG.warn("Ignoring exception reading or writing challenge: ", e);
        closeSocket(sock);
        return false;
    }


    // authenticate learner
    QuorumPeer.QuorumServer qps = self.getVotingView().get(sid);
    if (qps != null) {
        // TODO - investigate why reconfig makes qps null.
        authLearner.authenticate(sock, qps.hostname);
    }


    // If lost the challenge, then drop the new connection
    //保证集群中所有节点之间只有一个通道连接
    if (sid > self.getId()) {
        LOG.info("Have smaller server identifier, so dropping the " +
                "connection: (" + sid + ", " + self.getId() + ")");
        closeSocket(sock);
        // Otherwise proceed with the connection
    } else {
        SendWorker sw = new SendWorker(sock, sid);
        RecvWorker rw = new RecvWorker(sock, din, sid, sw);
        sw.setRecv(rw);


        SendWorker vsw = senderWorkerMap.get(sid);


        if(vsw != null)
            vsw.finish();


        senderWorkerMap.put(sid, sw);
        queueSendMap.putIfAbsent(sid, new ArrayBlockingQueue<ByteBuffer>(
                SEND_CAPACITY));


        sw.start();
        rw.start();


        return true;


    }
    return false;
}





如上述代码中所示,sid>self.sid 才可以创建连接 Socket 和 SendWorker,RecvWorker 线程,存储到 senderWorkerMap中。对应第 2 步中的 sid<self.sid 逻辑,保证集群中所有节点之间只有一个通道连接。

图 4 节点之间连接方式

【Java】


public Vote lookForLeader() throws InterruptedException {
//...
    if (n.electionEpoch > logicalclock.get()) {
        //当前选举周期小于选票周期,重置recvset选票池
        //大于当前周期更新当前选票信息,再次发送投票
        logicalclock.set(n.electionEpoch);
        recvset.clear();
        if(totalOrderPredicate(n.leader, n.zxid, n.peerEpoch,
                getInitId(), getInitLastLoggedZxid(), getPeerEpoch())) {
            updateProposal(n.leader, n.zxid, n.peerEpoch);
        } else {
            updateProposal(getInitId(),
                    getInitLastLoggedZxid(),
                    getPeerEpoch());
        }
        sendNotifications();
    } else if (n.electionEpoch < logicalclock.get()) {
        if(LOG.isDebugEnabled()){
            LOG.debug("Notification election epoch is smaller than logicalclock. n.electionEpoch = 0x"
                    + Long.toHexString(n.electionEpoch)
                    + ", logicalclock=0x" + Long.toHexString(logicalclock.get()));
        }
        break;
    } else if (totalOrderPredicate(n.leader, n.zxid, n.peerEpoch,
            proposedLeader, proposedZxid, proposedEpoch)) {//相同选举周期
        //接收的选票与当前选票PK成功后,替换当前选票
        updateProposal(n.leader, n.zxid, n.peerEpoch);
        sendNotifications();
    }
//...


}





在上代码中,自旋从 recvqueue 队列中获取到选票信息。开始进行选举:

  • 判断当前选票和接收过来的选票周期是否一致
  • 大于当前周期更新当前选票信息,再次发送投票
  • 周期相等:当前选票信息和接收的选票信息进行 PK
【Java】
//接收的选票与当前选票PK
protected boolean totalOrderPredicate(long newId, long newZxid, long newEpoch, long curId, long curZxid, long curEpoch) {
        LOG.debug("id: " + newId + ", proposed id: " + curId + ", zxid: 0x" +
                Long.toHexString(newZxid) + ", proposed zxid: 0x" + Long.toHexString(curZxid));
        if(self.getQuorumVerifier().getWeight(newId) == 0){
            return false;
        }


        /*
         * We return true if one of the following three cases hold:
         * 1- New epoch is higher
         * 2- New epoch is the same as current epoch, but new zxid is higher
         * 3- New epoch is the same as current epoch, new zxid is the same
         *  as current zxid, but server id is higher.
         */
        return ((newEpoch > curEpoch) ||
                ((newEpoch == curEpoch) &&
                ((newZxid > curZxid) || ((newZxid == curZxid) && (newId > curId)))));wId > curId)))));
  }





在上述代码中的 totalOrderPredicate 方法逻辑如下:

  • 竞选周期大于当前周期为 true
  • 竞选周期相等,竞选 zxid 大于当前 zxid 为 true
  • 竞选周期相等,竞选 zxid 等于当前 zxid,竞选 sid 大于当前 sid 为 true
  • 经过上述条件判断为 true 将当前选票信息替换为竞选成功的选票,同时再次将新的选票投出去。
【Java】
public Vote lookForLeader() throws InterruptedException {
//...
   //存储节点对应的选票信息
    // key:选票来源sid  value:选票推举的Leader sid
    recvset.put(n.sid, new Vote(n.leader, n.zxid, n.electionEpoch, n.peerEpoch));


    //半数选举开始
    if (termPredicate(recvset,
            new Vote(proposedLeader, proposedZxid,
                    logicalclock.get(), proposedEpoch))) {
        // Verify if there is any change in the proposed leader
        while((n = recvqueue.poll(finalizeWait,
                TimeUnit.MILLISECONDS)) != null){
            if(totalOrderPredicate(n.leader, n.zxid, n.peerEpoch,
                    proposedLeader, proposedZxid, proposedEpoch)){
                recvqueue.put(n);
                break;
            }
        }
        /*WorkerSender
         * This predicate is true once we don't read any new
         * relevant message from the reception queue
         */
        if (n == null) {
            //已选举出leader 更新当前节点是否为leader 
            self.setPeerState((proposedLeader == self.getId()) ?
                    ServerState.LEADING: learningState());


            Vote endVote = new Vote(proposedLeader,
                    proposedZxid, proposedEpoch);
            leaveInstance(endVote);
            return endVote;
        }
    }
//...
}
/**
     * Termination predicate. Given a set of votes, determines if have
     * sufficient to declare the end of the election round.
     *
     * @param votes
     *            Set of votes
     * @param vote
     *            Identifier of the vote received last  PK后的选票
     */
private boolean termPredicate(HashMap<Long, Vote> votes, Vote vote) {
    SyncedLearnerTracker voteSet = new SyncedLearnerTracker();
    voteSet.addQuorumVerifier(self.getQuorumVerifier());
    if (self.getLastSeenQuorumVerifier() != null
            && self.getLastSeenQuorumVerifier().getVersion() > self
                    .getQuorumVerifier().getVersion()) {
        voteSet.addQuorumVerifier(self.getLastSeenQuorumVerifier());
    }
    /*
     * First make the views consistent. Sometimes peers will have different
     * zxids for a server depending on timing.
     */
    //votes 来源于recvset 存储各个节点推举出来的选票信息
    for (Map.Entry<Long, Vote> entry : votes.entrySet()) {
//选举出的sid和其它节点选择的sid相同存储到voteSet变量中。
        if (vote.equals(entry.getValue())) {
//保存推举出来的sid
            voteSet.addAck(entry.getKey());
        }
    }
    //判断选举出来的选票数量是否过半
    return voteSet.hasAllQuorums();
}
//QuorumMaj#containsQuorum
public boolean containsQuorum(Set<Long> ackSet) {
    return (ackSet.size() > half);
   }





在上述代码中:recvset 是存储每个 sid 推举的选票信息。

第一轮 sid1:vote(1,0,1) ,sid2:vote(2,0,1);

第二轮 sid1:vote(2,0,1) ,sid2:vote(2,0,1)。

最终经过选举信息 vote(2,0,1) 为推荐 leader,并用推荐 leader 在 recvset 选票池里比对持相同票数量为 2 个。因为总共有 3 个节点参与选举,sid1 和 sid2 都选举 sid2 为 leader,满足票数过半要求,故确认 sid2 为 leader。

  • setPeerState 更新当前节点角色;
  • proposedLeader 选举出来的 sid 和自己 sid 相等,设置为 Leader;
  • 上述条件不相等,设置为 Follower 或 Observing;
  • 更新 currentVote 当前选票为 Leader 的选票 vote(2,0,1)。

07 总结

通过对 Leader 选举源码的解析,可以了解到:

  1. 多个应用节点之间网络通信采用 BIO 方式进行相互投票,同时保证每个节点之间只使用一个通道,减少网络资源的消耗,足以见得在 BIO 分布式中间件开发中的技术重要性。

  2. 基于 BIO 的基础上,灵活运用多线程和内存消息队列完好实现多层队列架构,每层队列由不同的线程分工协作,提高快速选举性能目的。

  3. 为 BIO 在多线程技术上的实践带来了宝贵的经验。

暂无回复。
需要 登录 后方可回复, 如果你还没有账号请点击这里 注册