RocketMQ主从同步机制源码解析及关键点总结(一)

我主要依照以下问题点及顺序来依次讲解:

  1. Slave角色配置以及基于配置下如何实现不同于Master的运行机制
  2. Master的数据传输以及Slave的进度上报

首先看第一点,Slave角色的关键配置:brokerIdbrokerRole,brokerId指定此Slave在Broker中的序号,0表示Master,1及之后的表示Slave,Broker中Slave可以有多个,当然一般一个就够了,所以brokerId的值一般为1。brokerRole表示其在Broker中的角色定位,有3个值可选:

public enum BrokerRole {
    ASYNC_MASTER,
    SYNC_MASTER,
    SLAVE;
}

ASYNC_MASTER:异步Master,也就是新的消息存储时不需要等Slave同步好;
SYNC_MASTER:同步Master,新消息存现时需要等Slave同步好,也就是返回的 Ack Offset >= 当前消息的CommitLog Offset

Slave角色其brokerRole=SLAVE。

brokerId和brokerRole是关键配置,其他的配置和Master一样,brokerName,brokerClusterName,namesrvAddr等相应配置都不能少。

当Slave启动时:

public class BrokerController {

    public boolean initialize() throws CloneNotSupportedException {
        ......
        //当当前角色是Slave时
        if (BrokerRole.SLAVE == this.messageStoreConfig.getBrokerRole()) {
            // 如果当前Broker配置中指定了haMasterAddress,则赋值 HAClient 的 masterAddress
            if (this.messageStoreConfig.getHaMasterAddress() != null && this.messageStoreConfig.getHaMasterAddress().length() >= 6) {
                // 将haMasterAddress的值设置到 HAService 的 HAClient 的masterAddress中
                this.messageStore.updateHaMasterAddress(this.messageStoreConfig.getHaMasterAddress());
                this.updateMasterHAServerAddrPeriodically = false;
            } else {
                this.updateMasterHAServerAddrPeriodically = true;   //如果配置中未指定Master的IP,则定期从Namesrv处更新获取
            }

            //Slave每隔60S从Master处同步TopicConfig,ConsumerOffset,DelayOffset,SubscriptionGroupConfig
            this.scheduledExecutorService.scheduleAtFixedRate(new Runnable() {

                @Override
                public void run() {
                    try {
                        BrokerController.this.slaveSynchronize.syncAll();
                    } catch (Throwable e) {
                        log.error("ScheduledTask syncAll slave exception", e);
                    }
                }
            }, 1000 * 10, 1000 * 60, TimeUnit.MILLISECONDS);
        }
        ......
    }

}

如果在配置文件中指定了haMasterAddress,那么这个就会作为Master的IP+端口,Slave会向此地址发送请求,建立长连接。

如果未指定,那么看如下:

public class BrokerController {

    /**
     * 向Namesrv注册Broker信息,每隔30S执行一次
     *
     * @param checkOrderConfig
     * @param oneway
     */
    public synchronized void registerBrokerAll(final boolean checkOrderConfig, boolean oneway) {
        ......
        if (registerBrokerResult != null) {
            //如果Slave在配置时没有指定Master的IP
            if (this.updateMasterHAServerAddrPeriodically && registerBrokerResult.getHaServerAddr() != null) {
                // Slave在注册Broker时,Namesrv会将Master的BrokerAddr返回来,更新HAClient上的 masterAddress
                this.messageStore.updateHaMasterAddress(registerBrokerResult.getHaServerAddr());
            }
            ......
        }
    }
}

Namesrv:
public class RouteInfoManager {

    /**
     * broker向Namesrv 注册自身信息
     *
     * 当 broker 为主节点时:不返回 haServerAddr、masterAddr
     * 当 broker 为从节点时:返回 haServerAddr、masterAddr
     */
    public RegisterBrokerResult registerBroker(
        final String clusterName,
        final String brokerAddr,
        final String brokerName,
        final long brokerId,
        final String haServerAddr,
        final TopicConfigSerializeWrapper topicConfigWrapper,
        final List filterServerList,
        final Channel channel) {
        ......
        // 当注册Slave时,返回Master的addr作为高可用的地址
        if (MixAll.MASTER_ID != brokerId) {
            String masterAddr = brokerData.getBrokerAddrs().get(MixAll.MASTER_ID);
            if (masterAddr != null) {
                // 获取master的brokerLiveInfo,将其haServerAddr(addr)赋值给Slave的haServerAddr
                BrokerLiveInfo brokerLiveInfo = this.brokerLiveTable.get(masterAddr);
                if (brokerLiveInfo != null) {
                    result.setHaServerAddr(brokerLiveInfo.getHaServerAddr());
                    result.setMasterAddr(masterAddr);
                }
            }
        }
    }

}

若Slave未显式指定haMasterAddress,则在向Namesrv注册Broker信息时,Namesrv会以brokerName相同为依据将此Broker下的Master Adress返回给Slave。

通过以上两种手段指定MasterAdress后,将其值赋给HAService里的HAClient的 masterAddress属性。Slave在start时,调用HAClient的start( ),当发现masterAddress有值时,就明白当前是Slave角色,就会启动Slave的运行流程。

Broker启动时,会相应启动HAService,用于主从复制,同时相应启动其内部的HAClient,下面看HAClient的启动过程:

class HAClient extends ServiceThread {

    public void run() {
        log.info(this.getServiceName() + " service started");

        while (!this.isStopped()) {
            try {
                //当存在masterAdress != null && 连接Master成功
                if (this.connectMaster()) {
                    // 若距离上次上报时间超过5S,上报到Master进度
                    if (this.isTimeToReportOffset()) {
                        boolean result = this.reportSlaveMaxOffset(this.currentReportedOffset);
                        if (!result) {
                            this.closeMaster();
                        }
                    }
                    //最多阻塞1S,直到Master有数据同步于过来。若1S满了还是没有接受到数据,中断阻塞,
                    // 执行processReadEvent(),但结果读入byteBufferRead的大小为0,然后循环到这步
                    this.selector.select(1000);

                    // 处理读取事件
                    boolean ok = this.processReadEvent();
                    if (!ok) {
                        this.closeMaster();
                    }

                    // 若进度有变化,上报到Master进度
                    if (!reportSlaveMaxOffsetPlus()) {
                        continue;
                    }

                    long interval = HAService.this.getDefaultMessageStore().getSystemClock().now() - this.lastWriteTimestamp;
                    // Master超过20S未返回数据,关闭连接
                    if (interval > HAService.this.getDefaultMessageStore().getMessageStoreConfig().getHaHousekeepingInterval()) {
                        log.warn("HAClient, housekeeping, found this connection[" + this.masterAddress
                            + "] expired, " + interval);
                        this.closeMaster();
                        log.warn("HAClient, master not response some time, so close connection");
                    }
                } else {
                    this.waitForRunning(1000 * 5);
                }
            } catch (Exception e) {
                log.warn(this.getServiceName() + " service has exception. ", e);
                this.waitForRunning(1000 * 5);
            }
        }

        log.info(this.getServiceName() + " service end");
    }


}

先看connectMaster( ) 方法,此方法就是向Master发送请求,通了之后建立SocketChannel:

    /**
     * 连接Master节点
     *
     * @return 是否连接成功
     * @throws ClosedChannelException 当注册读事件失败时
     */
    private boolean connectMaster() throws ClosedChannelException {
        if (null == socketChannel) {
            String addr = this.masterAddress.get();
            //当自身角色是Slave时,会将配置中的MessageStoreConfig中的haMasterAdress赋值给masterAddress,或者在registerBroker时的返回值赋值
            if (addr != null) {
                SocketAddress socketAddress = RemotingUtil.string2SocketAddress(addr);
                if (socketAddress != null) {
                    this.socketChannel = RemotingUtil.connect(socketAddress); // 连接Master节点。如果连接失败,直接关闭,不抛出异常。
                    if (this.socketChannel != null) {
                        this.socketChannel.register(this.selector, SelectionKey.OP_READ);
                    }
                }
            }
            //更新最近上报Offset,也就是当前CommitLog的maxOffset
            this.currentReportedOffset = HAService.this.defaultMessageStore.getMaxPhyOffset();
            //更新最近写入时间
            this.lastWriteTimestamp = System.currentTimeMillis();
        }

        return this.socketChannel != null;
    }

Slave向Master发送请求,那就需要去Master看看处理这个请求的方式,Broker在启动时会启动HAService内的AcceptSocketService,顾名思义,这个对象就是用来接收Socket请求的服务对象,看其对Slave的请求处理形式:

/**
 * Listens to slave connections to create {@link HAConnection}.
 */
class AcceptSocketService extends ServiceThread {

    public void run() {
        log.info(this.getServiceName() + " service started");

        while (!this.isStopped()) {
            try {
                this.selector.select(1000);
                Set selected = this.selector.selectedKeys();

                if (selected != null) {
                    for (SelectionKey k : selected) {
                        if ((k.readyOps() & SelectionKey.OP_ACCEPT) != 0) {
                            SocketChannel sc = ((ServerSocketChannel)k.channel()).accept();

                            //接收到Slave的请求,生成一个HAConnection,用户向Channle写数据以及读取Slave传过来的数据
                            if (sc != null) {
                                HAService.log.info("HAService receive new connection, "
                                    + sc.socket().getRemoteSocketAddress());

                                try {
                                    HAConnection conn = new HAConnection(HAService.this, sc);
                                    conn.start();
                                    HAService.this.addConnection(conn);
                                } catch (Exception e) {
                                    log.error("new HAConnection exception", e);
                                    sc.close();
                                }
                            }
                        } else {
                            log.warn("Unexpected ops in select " + k.readyOps());
                        }
                    }

                    selected.clear();
                }
            } catch (Exception e) {
                log.error(this.getServiceName() + " service has exception.", e);
            }
        }

        log.info(this.getServiceName() + " service end");
    }

}

AcceptSocketService在启动时会监听SelectionKey.OP_ACCEPT事件,当接收到请求时,创建一个SocketChannel长连接对象,然后用HAConnection去封装SocketChannel,之后启动HAConnection,可以先看看HAConnection的start( ) 方法做了什么:

public class HAConnection {

    public HAConnection(final HAService haService, final SocketChannel socketChannel) throws IOException {
        this.haService = haService;
        this.socketChannel = socketChannel;
        this.clientAddr = this.socketChannel.socket().getRemoteSocketAddress().toString();
        this.socketChannel.configureBlocking(false);
        this.socketChannel.socket().setSoLinger(false, -1);
        this.socketChannel.socket().setTcpNoDelay(true);
        this.socketChannel.socket().setReceiveBufferSize(1024 * 64);
        this.socketChannel.socket().setSendBufferSize(1024 * 64);
        this.writeSocketService = new WriteSocketService(this.socketChannel);
        this.readSocketService = new ReadSocketService(this.socketChannel);
        this.haService.getConnectionCount().incrementAndGet();
    }

    /**
     * 启动{@link #readSocketService}读取Slave传过来的数据
     * 启动{@link #writeSocketService}向Channel写入消息,同步给Slave
     */
    public void start() {
        this.readSocketService.start();
        this.writeSocketService.start();
    }

}

HAConnection 在实例化时生成了writeSocketServicereadSocketService 两个对象。
writeSocketService:负责向SocketChannel写入CommitLog的消息数据;
readSocketService:负责从SocketChannel中读取Slave传递的ACK进度,也就是其CommitLog的maxPhyOffset。

Master在接收到Slave的请求后,创建SocketChannel,封装成一个HAConnection ,执行start( )方法,也就是执行writeSocketService和readSocketService。但是仅仅是start( ),不会主动向Slave发送数据,Slave和Master之间的互相传输的第一步是Slave迈出的。下面我们看Slave迈出的这一步:

class HAClient extends ServiceThread {
    //Slave当前进度默认是0
    private long currentReportedOffset = 0;

    public void run() {
         ......      
         //当存在masterAdress != null && 连接Master成功
         if (this.connectMaster()) {
             // 若距离上次上报时间超过5S,上报到Master进度
             if (this.isTimeToReportOffset()) {
                 boolean result = this.reportSlaveMaxOffset(this.currentReportedOffset);
                 if (!result) {
                     this.closeMaster();
                 }
             }
             ......
              // 若进度有变化,上报到Master进度
              if (!reportSlaveMaxOffsetPlus()) {
                  continue;
              }
              ......
           }
     }

    /**
     * 是否满足上报进度的时间
     * 距离上一次上报进度时间超过5S , 也就是每5S上报一次进度
     *
     * @return 是否
     */
    private boolean isTimeToReportOffset() {
        long interval = HAService.this.defaultMessageStore.getSystemClock().now() - this.lastWriteTimestamp;
        return interval > HAService.this.defaultMessageStore.getMessageStoreConfig().getHaSendHeartbeatInterval();
    }

    /**
     * 上报Master进度。
     * 如果上报失败,关闭连接
     *
     * @return 是否上报成功
     */
    private boolean reportSlaveMaxOffsetPlus() {
        boolean result = true;
        long currentPhyOffset = HAService.this.defaultMessageStore.getMaxPhyOffset();
        if (currentPhyOffset > this.currentReportedOffset) {
            this.currentReportedOffset = currentPhyOffset;
            result = this.reportSlaveMaxOffset(this.currentReportedOffset);
            if (!result) {
                this.closeMaster();
                log.error("HAClient, reportSlaveMaxOffset error, " + this.currentReportedOffset);
            }
        }

        return result;
    }

    /**
     * 上报进度,传递进度的时候仅传递一个Long类型的Offset,8个字节,没有其他数据
     *
     * @param maxOffset 进度
     * @return 是否上报成功
     */
    private boolean reportSlaveMaxOffset(final long maxOffset) {
        this.reportOffset.position(0);
        this.reportOffset.limit(8);
        this.reportOffset.putLong(maxOffset);
        this.reportOffset.position(0);
        this.reportOffset.limit(8);

        for (int i = 0; i < 3 && this.reportOffset.hasRemaining(); i++) {
            try {
                this.socketChannel.write(this.reportOffset);
            } catch (IOException e) {
                log.error(this.getServiceName()
                    + "reportSlaveMaxOffset this.socketChannel.write exception", e);
                return false;
            }
        }

        return !this.reportOffset.hasRemaining();
    }

}

Slave在连接Master成功后,生成SocketChannel,注册SelectionKey.OP_READ事件,当Master有数据传输过来,
触发 SelectionKey.OP_READ,此时就能读取其数据。但Master启动时是不会主动传输数据的,因为其不知道Slave的CommitLog的maxPhyOffset,也就是不知道从哪个位置开始同步,需要Slave先上报当前CommitLog的maxPhyOffset。

reportSlaveMaxOffsetPlus( ) :Slave在启动时如果其maxPhyOffset > currentReportedOffset ,也就是大于0,也就是当前Slave是中途宕机了后再次启动的,那么马上向Master上报进度,从maxPhyOffset 处开始同步数据

isTimeToReportOffset() :Slave在启动时如果maxPhyOffset == 0 ,也就是说当前Slave是第一次启动,那么其会一直等待,等待5S中,触发isTimeToReportOffset(),然后向Master同步进度,从0开始同步数据。

为什么会有这种区别,在之后的代码中我会说说我自己的理解。

Slave将maxPhyOffset 传输给了Master,接下来看看Master收到这个进度后的处理流程。

/**
 * 读取Slave进度线程服务
 */
class ReadSocketService extends ServiceThread {

    public ReadSocketService(final SocketChannel socketChannel) throws IOException {
        this.selector = RemotingUtil.openSelector();
        this.socketChannel = socketChannel;
        //向SocketChannle注册OP_READ事件
        this.socketChannel.register(this.selector, SelectionKey.OP_READ);
        this.thread.setDaemon(true);
    }

    public void run() {

        while (!this.isStopped()) {
            try {
                //最多阻塞1S钟,当有一个Channel发送了数据后,立即中断阻塞
                //若1S时间内也没有Channel发送了数据,则processReadEvent()会立即返回true,然后再次阻塞,循环如此
                this.selector.select(1000);
                boolean ok = this.processReadEvent();
                if (!ok) {
                    HAConnection.log.error("processReadEvent error");
                    break;
                }

                // Slave超过20S没有返回数据,断开连接
                long interval = HAConnection.this.haService.getDefaultMessageStore().getSystemClock().now() - this.lastReadTimestamp;
                //超过20S时间没有收到Slave传递的消息,断开Slave连接
                if (interval > HAConnection.this.haService.getDefaultMessageStore().getMessageStoreConfig().getHaHousekeepingInterval()) {
                    log.warn("ha housekeeping, found this connection[" + HAConnection.this.clientAddr + "] expired, " + interval);
                    break;
                }
            } catch (Exception e) {
                HAConnection.log.error(this.getServiceName() + " service has exception.", e);
                break;
            }
        }
        ......
    }

}

当Slave传递了自身的maxPhyOffset时,马上中断 selector.select(1000),执行processReadEvent()方法。

private boolean processReadEvent() {
    int readSizeZeroTimes = 0;

    // 重置读取缓冲区以及processPosition
    if (!this.byteBufferRead.hasRemaining()) {
        this.byteBufferRead.flip();
        this.processPostion = 0;
    }

    while (this.byteBufferRead.hasRemaining()) {
        try {
            //从Channel里读取数据,可能Slave没有发送过来进度
            int readSize = this.socketChannel.read(this.byteBufferRead);
            if (readSize > 0) {
                readSizeZeroTimes = 0;

                // 设置最后读取时间
                this.lastReadTimestamp = HAConnection.this.haService.getDefaultMessageStore().getSystemClock().now();
                //传递过来的数据大于8字节,有效的同步,8个字节代表着有一个有效的CommitLog Offset
                if ((this.byteBufferRead.position() - this.processPostion) >= 8) {
                    // 读取Slave 请求来的CommitLog的最大位置
                    // position减去8的余数,这么做也是是为了防止流传输数据的粘包问题
                    int pos = this.byteBufferRead.position() - (this.byteBufferRead.position() % 8);
                    // 读取ByteBuffer中position内最大的8字节的倍数,比如 position = 571,则 571-(571%8)= 571-3 = 568
                    // 也就是读取 560-568位置的字节,因为Slave可能发送了多个进度过来,Master只读最末尾也就是最大的那个
                    long readOffset = this.byteBufferRead.getLong(pos - 8);
                    this.processPostion = pos;

                    // 设置Slave CommitLog的最大位置
                    HAConnection.this.slaveAckOffset = readOffset;

                    // 设置Slave 第一次请求的位置,初始化 slaveRequestOffset = Slave的请求位置
                    if (HAConnection.this.slaveRequestOffset < 0) {
                        HAConnection.this.slaveRequestOffset = readOffset;
                        log.info("slave[" + HAConnection.this.clientAddr + "] request offset " + readOffset);
                    }

                    // 通知目前Slave进度。主要用于Master节点为同步类型的。
                    HAConnection.this.haService.notifyTransferSome(HAConnection.this.slaveAckOffset);
                }
            } else if (readSize == 0) {
                if (++readSizeZeroTimes >= 3) {  //连续读取3此没有数据,break,返回true,休眠
                    break;
                }
            } else {
                log.error("read socket[" + HAConnection.this.clientAddr + "] < 0");
                return false;
            }
        } catch (IOException e) {
            log.error("processReadEvent exception", e);
            return false;
        }
    }

    return true;
}

HAConnection.this.slaveRequestOffset的默认值是-1,当Slave第一次同步进度时,更新其值为Slave的原始maxPhyOffset,writeSocketService启动后会定期循环,当发现slaveRequestOffset != -1 时,就知道从哪里开始将数据发送给Slave了,然后就会马上从此进度开始向SocketChannel写入数据。

接下来看WriteSocketService,也就是向SocketChannel写入数据的服务:

/**
 * 写入Message线程服务
 */
class WriteSocketService extends ServiceThread {
    ......
    private final int headerSize = 8 + 4;
    private final ByteBuffer byteBufferHeader = ByteBuffer.allocate(headerSize);

    /**
     * Slave下一次同步CommigLog的Offset
     */
    private long nextTransferFromWhere = -1;

    public void run() {

        while (!this.isStopped()) {
            try {
                this.selector.select(1000);

                // 从未获得Slave读取进度请求,sleep等待。Slave第一次请求时会将初始化 slaveRequestOffset = Slave原始进度,可能为0,也可能是宕机重启,大于0
                // 只有Slave同步了其原始的maxPhyOffset时,Master才知道从哪里开始将数据同步给Slave
                if (-1 == HAConnection.this.slaveRequestOffset) {
                    Thread.sleep(10);
                    continue;
                }

                // 计算初始化nextTransferFromWhere
                if (-1 == this.nextTransferFromWhere) {
                    //Slave的CommitLog里没有任何数据,当前请求是其第一次请求
                    if (0 == HAConnection.this.slaveRequestOffset) {
                        long masterOffset = HAConnection.this.haService.getDefaultMessageStore().getCommitLog().getMaxOffset();
                        //从lastMappedFile的0位开始,也就是会忽略之前的MappedFile,只从最后一个文件开始同步
                        masterOffset = masterOffset - (masterOffset % HAConnection.this.haService.getDefaultMessageStore().getMessageStoreConfig()
                            .getMapedFileSizeCommitLog());
                        if (masterOffset < 0) {
                            masterOffset = 0;
                        }
                        this.nextTransferFromWhere = masterOffset;
                    } else {
                        //如果CommitLog是中途宕机了,重启,那么会接着上次的位置继续传输
                        this.nextTransferFromWhere = HAConnection.this.slaveRequestOffset;
                    }

                    log.info("master transfer data from " + this.nextTransferFromWhere + " to slave[" + HAConnection.this.clientAddr
                        + "], and slave request " + HAConnection.this.slaveRequestOffset);
                }

                if (this.lastWriteOver) {
                    long interval = HAConnection.this.haService.getDefaultMessageStore().getSystemClock().now() - this.lastWriteTimestamp;
                    //如果距离上次同步(写数据给Slave)超过5S,也就是5S内没有新的消息,发送一个空包过去,就是没有实际数据的,刷新Slave的最后写入时间
                    if (interval > HAConnection.this.haService.getDefaultMessageStore().getMessageStoreConfig().getHaSendHeartbeatInterval()) {
                        // Build Header
                        this.byteBufferHeader.position(0);
                        this.byteBufferHeader.limit(headerSize);
                        this.byteBufferHeader.putLong(this.nextTransferFromWhere);
                        this.byteBufferHeader.putInt(0);
                        this.byteBufferHeader.flip();

                        this.lastWriteOver = this.transferData();
                        if (!this.lastWriteOver) {
                            continue;
                        }
                    }
                } else { // 上次传输未完成,,继续传输
                    this.lastWriteOver = this.transferData();
                    if (!this.lastWriteOver) {
                        continue;
                    }
                }

                // 选择新的CommitLog内容进行传输
                SelectMappedBufferResult selectResult =
                    HAConnection.this.haService.getDefaultMessageStore().getCommitLogData(this.nextTransferFromWhere);
                if (selectResult != null) {
                    int size = selectResult.getSize();
                    //一次最多同步传输32KB
                    if (size > HAConnection.this.haService.getDefaultMessageStore().getMessageStoreConfig().getHaTransferBatchSize()) {
                        size = HAConnection.this.haService.getDefaultMessageStore().getMessageStoreConfig().getHaTransferBatchSize();
                    }

                    long thisOffset = this.nextTransferFromWhere;
                    this.nextTransferFromWhere += size;   //更新下次同步的起始位置,当前位置 + 修正过的size

                    selectResult.getByteBuffer().limit(size);
                    this.selectMappedBufferResult = selectResult;

                    // Build Header
                    this.byteBufferHeader.position(0);
                    this.byteBufferHeader.limit(headerSize);    // 头部大小12字节,前8 Offset ; 后4 是 totalSize
                    this.byteBufferHeader.putLong(thisOffset);  // 将同步的CommitLog的Offset存入
                    this.byteBufferHeader.putInt(size);         // 将同步的数据size存入
                    this.byteBufferHeader.flip();  //转为写模式

                    this.lastWriteOver = this.transferData();
                } else { // 没新的消息,挂起等待
                    HAConnection.this.haService.getWaitNotifyObject().allWaitForRunning(100);
                }
            } catch (Exception e) {
                HAConnection.log.error(this.getServiceName() + " service has exception.", e);
                break;
            }
        }

        ......
    }

}

以上代码的注释挺详细的了,看注释加代码基本能看明白,有几点需要额外讲解下:

  1. 当Slave是第一次启动时,也就是其CommitLog里没有任何数据,发送的maxPhyOffset==0,那么Master在同步数据时会仅同步lastMappedFile,也就是从最后一个MappedFile的起始位置开始同步,忽略其前面的所有MappedFile。如果是中途宕机后的再次启动,那么会接着其CommitLog的进度继续同步,否则其CommitLog就会出现数据断层的情况。也正是基于这种机制,当Slave再次启动时,马上就上报进度,开始数据同步,因为其进度与Master的maxPhyOffset可能间隔了一个或多个MappedFile了,要抓紧时间跟上。
  2. Master每次同步CommitLog的数据大小最大为32KB,不包括Header,也就是8byte的offset,4byte的bodySize。
  3. 由以上可知,Master传输数据并不是按消息整体来传输的,也就是说很多时候会传递消息的一部分
  4. Master会在一次数据写入SocketChannel后紧接着就进行下次写入,而不需要等待Slave返回ACK,也就是说不需要等待Slave返回确认收到
  5. 当没有新消息时,WriteSocketService每休眠100ms运行一次

下面看WriteSocketService 的transferData( )方法,也就是时间数据传输方法:

/**
 * 传输数据
 */
private boolean transferData() throws Exception {
    int writeSizeZeroTimes = 0;
    // Write Header
    while (this.byteBufferHeader.hasRemaining()) {
        int writeSize = this.socketChannel.write(this.byteBufferHeader);
        if (writeSize > 0) {
            writeSizeZeroTimes = 0;
            this.lastWriteTimestamp = HAConnection.this.haService.getDefaultMessageStore().getSystemClock().now();
        } else if (writeSize == 0) {
            if (++writeSizeZeroTimes >= 3) {
                break;
            }
        } else {
            throw new Exception("ha master write header error < 0");
        }
    }

    if (null == this.selectMappedBufferResult) {
        return !this.byteBufferHeader.hasRemaining();
    }

    writeSizeZeroTimes = 0;

    // Write Body
    if (!this.byteBufferHeader.hasRemaining()) {
        while (this.selectMappedBufferResult.getByteBuffer().hasRemaining()) {
            int writeSize = this.socketChannel.write(this.selectMappedBufferResult.getByteBuffer());
            if (writeSize > 0) {
                writeSizeZeroTimes = 0;
                this.lastWriteTimestamp = HAConnection.this.haService.getDefaultMessageStore().getSystemClock().now();
            } else if (writeSize == 0) {
                if (++writeSizeZeroTimes >= 3) {
                    break;
                }
            } else {
                throw new Exception("ha master write body error < 0");
            }
        }
    }

    boolean result = !this.byteBufferHeader.hasRemaining() && !this.selectMappedBufferResult.getByteBuffer().hasRemaining();

    if (!this.selectMappedBufferResult.getByteBuffer().hasRemaining()) {
        this.selectMappedBufferResult.release();
        this.selectMappedBufferResult = null;
    }

    return result;
}

第一次传输byteBufferHeader,也就是8个字节的起始offset,加上4个字节的bodySize。如果bodySize == 0 ( 也就是 selectMappedBufferResult == null ) ,也就是Slave的进度等于Master的进度了,此时仅传递一个offset,当长时间没有消息生成时需要这样,为了更新Master和Slave的最后同步时间。当Slave进度落后于Master进度,提取Slave的maxPhyOffset到Master的maxPhyOffset之间的数据,写入SocketChannel。

下篇博客我们接着讲Slave接收到数据的处理过程以及SYNC_MASTER下的数据同步流程。

你可能感兴趣的:(rocketmq)