Bootstrap

分布式专题-分布式服务治理05-Dubbo源码分析(下篇)

前言

关于Dubbo,本系列文章主要讲三方面内容。前面我们已经了解到Dubbo的基本特性,常用配置、自适应扩展点与服务发布,服务注册的过程。接下来我们主要讲服务端接收消息处理过程。

本节我们讲5、6、7、8

  1. Dubbo Extension扩展点
  2. 服务发布过程
  3. 消费端初始化过程
  4. 什么时候建立和服务端的连接
  5. 服务端接收消息处理过程
  6. Directory
  7. Cluster
  8. LoadBalance

Tips:本节文末有Dubbo中文注释版的源码哦~

服务端接收消息处理过程

服务端底层也是基于netty通信
NettyHandler. messageReceived
接收消息的时候,通过NettyHandler.messageReceived作为入口。

@Override
public void messageReceived(ChannelHandlerContext ctx, MessageEvent e) throws Exception {
    NettyChannel channel = NettyChannel.getOrAddChannel(ctx.getChannel(), url, handler);
    try {
        handler.received(channel, e.getMessage());
    } finally {
        NettyChannel.removeChannelIfDisconnected(ctx.getChannel());
    }
}

handler.received
这个handler是什么呢?还记得在服务发布的时候,组装了一系列的handler吗?代码如下
HeaderExchanger.bind

public ExchangeServer bind(URL url, ExchangeHandler handler) throws RemotingException {
    return new HeaderExchangeServer(Transporters.bind(url, new DecodeHandler(new HeaderExchangeHandler(handler))));
}

received这里实际调用的是MultiMessageHandler.received
MultiMessageHandler.received
将复合消息转化为单独消息处理

public class MultiMessageHandler extends AbstractChannelHandlerDelegate {

    public MultiMessageHandler(ChannelHandler handler) {
        super(handler);
    }

    @SuppressWarnings("unchecked")
	@Override
    public void received(Channel channel, Object message) throws RemotingException {
        if (message instanceof MultiMessage) {
            MultiMessage list = (MultiMessage)message;
          // deal message 
            for(Object obj : list) {
                handler.received(channel, obj);
            }
        } else {
            handler.received(channel, message);
        }
    }
}

这里的received实际上调用的是HeartbeatHandler.received
HeartbeatHandler.received
心跳机制

 public void received(Channel channel, Object message) throws RemotingException {
        setReadTimestamp(channel);
        //如果当前消息是心跳包
        if (isHeartbeatRequest(message)) {
            Request req = (Request) message;
            if (req.isTwoWay()) {
                Response res = new Response(req.getId(), req.getVersion());
                res.setEvent(Response.HEARTBEAT_EVENT);
                channel.send(res);
                if (logger.isInfoEnabled()) {
                    int heartbeat = channel.getUrl().getParameter(Constants.HEARTBEAT_KEY, 0);
                    if(logger.isDebugEnabled()) {
                        logger.debug("Received heartbeat from remote channel " + channel.getRemoteAddress()
                                        + ", cause: The channel has no data-transmission exceeds a heartbeat period"
                                        + (heartbeat > 0 ? ": " + heartbeat + "ms" : ""));
                    }
	            }
            }
            return;
        }
        if (isHeartbeatResponse(message)) {
            if (logger.isDebugEnabled()) {
            	logger.debug(
                    new StringBuilder(32)
                        .append("Receive heartbeat response in thread ")
                        .append(Thread.currentThread().getName())
                        .toString());
            }
            return;
        }
        handler.received(channel, message);
    }

接下来的 received实际上调用的是AllChannelHandler.received

AllChannelHandler.received
起一个线程处理任务

    public void received(Channel channel, Object message) throws RemotingException {
        ExecutorService cexecutor = getExecutorService();
        try {
            cexecutor.execute(new ChannelEventRunnable(channel, handler, ChannelState.RECEIVED, message));
        } catch (Throwable t) {
            throw new ExecutionException(message, channel, getClass() + " error when process received event .", t);
        }
    }

既然是线程处理,找一下run方法:
AllChannelHandler.run

在这里插入图片描述
这里会判断状态,如果是received,则再次调用received,这里的received实际上调用的是DecodeHandler.received

  public void received(Channel channel, Object message) throws RemotingException {
        if (message instanceof Decodeable) {
            decode(message);
        }

        if (message instanceof Request) {
            decode(((Request)message).getData());
        }

        if (message instanceof Response) {
            decode( ((Response)message).getResult());
        }

        handler.received(channel, message);
    }

这里最后处理的received实际上调用的是
HeaderExchangeHandler.received

   public void received(Channel channel, Object message) throws RemotingException {
        channel.setAttribute(KEY_READ_TIMESTAMP, System.currentTimeMillis());
        ExchangeChannel exchangeChannel = HeaderExchangeChannel.getOrAddChannel(channel);
        try {
            if (message instanceof Request) {
                // handle request.
                Request request = (Request) message;
                if (request.isEvent()) {
                    handlerEvent(channel, request);
                } else {
                    if (request.isTwoWay()) {
                        Response response = handleRequest(exchangeChannel, request);
                        channel.send(response);
                    } else {
                        handler.received(exchangeChannel, request.getData());
                    }
                }
            } else if (message instanceof Response) {
                handleResponse(channel, (Response) message);
            } else if (message instanceof String) {
                if (isClientSide(channel)) {
                    Exception e = new Exception("Dubbo client can not supported string message: " + message + " in channel: " + channel + ", url: " + channel.getUrl());
                    logger.error(e.getMessage(), e);
                } else {
                    String echo = handler.telnet(channel, (String) message);
                    if (echo != null && echo.length() > 0) {
                        channel.send(echo);
                    }
                }
            } else {
                handler.received(exchangeChannel, message);
            }
        } finally {
            HeaderExchangeChannel.removeChannelIfDisconnected(channel);
        }
    }

NettyServer
回到主线,在Nettyserver中,wrap了多个handler,再次调用

public NettyServer(URL url, ChannelHandler handler) throws RemotingException {
    super(url, ChannelHandlers.wrap(handler, ExecutorUtil.setThreadName(url, SERVER_THREAD_POOL_NAME)));
}
protected ChannelHandler wrapInternal(ChannelHandler handler, URL url) {
    return new MultiMessageHandler(new HeartbeatHandler(ExtensionLoader.getExtensionLoader(Dispatcher.class)
            .getAdaptiveExtension().dispatch(handler, url)));
}

所以服务端的handler处理链为

MultiMessageHandler(HeartbeatHandler(AllChannelHandler(DecodeHandler)))
MultiMessageHandler: 复合消息处理
HeartbeatHandler:心跳消息处理,接收心跳并发送心跳响应
AllChannelHandler:业务线程转化处理器,把接收到的消息封装成
ChannelEventRunnable可执行任务给线程池处理
DecodeHandler:业务解码处理器

HeaderExchangeHandler.received

交互层请求响应处理,有三种处理方式

  1. handlerRequest,双向请求
  2. handler.received 单向请求
  3. handleResponse 响应消息
public void received(Channel channel, Object message) throws RemotingException {
    channel.setAttribute(KEY_READ_TIMESTAMP, System.currentTimeMillis());
    ExchangeChannel exchangeChannel = HeaderExchangeChannel.getOrAddChannel(channel);
    try {
        if (message instanceof Request) {
            // handle request.
            Request request = (Request) message;
            if (request.isEvent()) {
                handlerEvent(channel, request);
            } else {
                if (request.isTwoWay()) {
                    Response response = handleRequest(exchangeChannel, request);
                    channel.send(response);
                } else {
                    handler.received(exchangeChannel, request.getData());
                }
            }
        } else if (message instanceof Response) {
            handleResponse(channel, (Response) message);
        } else if (message instanceof String) {
            if (isClientSide(channel)) {
                Exception e = new Exception("Dubbo client can not supported string message: " + message + " in channel: " + channel + ", url: " + channel.getUrl());
                logger.error(e.getMessage(), e);
            } else {
                String echo = handler.telnet(channel, (String) message);
                if (echo != null && echo.length() > 0) {
                    channel.send(echo);
                }
            }
        } else {
            handler.received(exchangeChannel, message);
        }
    } finally {
        HeaderExchangeChannel.removeChannelIfDisconnected(channel);
    }
}

handleRequest
处理请求并返回response

Response handleRequest(ExchangeChannel channel, Request req) throws RemotingException {
    Response res = new Response(req.getId(), req.getVersion());
    if (req.isBroken()) {
        Object data = req.getData();
        String msg;
        if (data == null) msg = null;
        else if (data instanceof Throwable) msg = StringUtils.toString((Throwable) data);
        else msg = data.toString();
        res.setErrorMessage("Fail to decode request due to: " + msg);
        res.setStatus(Response.BAD_REQUEST);

        return res;
    }
    // find handler by message class.
    Object msg = req.getData();
    try {
        // handle data.
        Object result = handler.reply(channel, msg);
        res.setStatus(Response.OK);
        res.setResult(result);
    } catch (Throwable e) {
        res.setStatus(Response.SERVICE_ERROR);
        res.setErrorMessage(StringUtils.toString(e));
    }
    return res;
}

ExchangeHandlerAdaptive.replay(DubboProtocol)
调用DubboProtocol中定义的ExchangeHandlerAdaptive.replay方法处理消息

private ExchangeHandler requestHandler = new ExchangeHandlerAdapter() {
    
    public Object reply(ExchangeChannel channel, Object message) throws RemotingException {
     invoker.invoke(inv);
}

那接下来invoker.invoke会调用哪个类中的方法呢?还记得在RegistryDirectory中发布本地方法的时候,对invoker做的包装吗?通过InvokerDelegete对原本的invoker做了一层包装,而原本的invoker是什么呢?是一个JavassistProxyFactory生成的动态代理吧。所以此处的invoker应该是
Filter(Listener(InvokerDelegete(AbstractProxyInvoker (Wrapper.invokeMethod)))
RegistryDirectory生成invoker的代码如下

private <T> ExporterChangeableWrapper<T>  doLocalExport(final Invoker<T> originInvoker){
    String key = getCacheKey(originInvoker);
    ExporterChangeableWrapper<T> exporter = (ExporterChangeableWrapper<T>) bounds.get(key);
    if (exporter == null) {
        synchronized (bounds) {
            exporter = (ExporterChangeableWrapper<T>) bounds.get(key);
            if (exporter == null) {
                final Invoker<?> invokerDelegete = new InvokerDelegete<T>(originInvoker, getProviderUrl(originInvoker));
                exporter = new ExporterChangeableWrapper<T>((Exporter<T>)protocol.export(invokerDelegete), originInvoker);
                bounds.put(key, exporter);
            }
        }
    }
    return (ExporterChangeableWrapper<T>) exporter;
}

Directory

集群目录服务Directory, 代表多个Invoker, 可以看成List<Invoker>,它的值可能是动态变化的比如注册中心推送变更。集群选择调用服务时通过目录服务找到所有服务
StaticDirectory: 静态目录服务, 它的所有Invoker通过构造函数传入, 服务消费方引用服务的时候, 服务对多注册中心的引用,将Invokers集合直接传入 StaticDirectory构造器,再由Cluster伪装成一个Invoker;StaticDirectory的list方法直接返回所有invoker集合;
RegistryDirectory: 注册目录服务, 它的Invoker集合是从注册中心获取的, 它实现了NotifyListener接口实现了回调接口notify(List<Url>)

Directory目录服务的更新过程

RegistryProtocol.doRefer方法,也就是消费端在初始化的时候,这里涉及到了RegistryDirectory这个类。然后执行cluster.join(directory)方法。这些代码在上节课有分析过。
cluster.join其实就是将Directory中的多个Invoker伪装成一个Invoker, 对上层透明,包含集群的容错机制

private <T> Invoker<T> doRefer(Cluster cluster, Registry registry, Class<T> type, URL url) {
    RegistryDirectory<T> directory = new RegistryDirectory<T>(type, url);//对多个invoker进行组装
    directory.setRegistry(registry); //ZookeeperRegistry
    directory.setProtocol(protocol); //protocol=Protocol$Adaptive
    //url=consumer://192.168.111....
    URL subscribeUrl = new URL(Constants.CONSUMER_PROTOCOL, NetUtils.getLocalHost(), 0, type.getName(), directory.getUrl().getParameters());
    //会把consumer://192...  注册到注册中心
    if (! Constants.ANY_VALUE.equals(url.getServiceInterface())
            && url.getParameter(Constants.REGISTER_KEY, true)) {
        //zkClient.create()
        registry.register(subscribeUrl.addParameters(Constants.CATEGORY_KEY, Constants.CONSUMERS_CATEGORY,
                Constants.CHECK_KEY, String.valueOf(false)));
    }
    directory.subscribe(subscribeUrl.addParameter(Constants.CATEGORY_KEY, 
            Constants.PROVIDERS_CATEGORY 
            + "," + Constants.CONFIGURATORS_CATEGORY 
            + "," + Constants.ROUTERS_CATEGORY));
    //Cluster$Adaptive
    return cluster.join(directory);
}

directory.subscribe

订阅节点的变化

  1. 当zookeeper上指定节点发生变化以后,会通知到RegistryDirectory的notify方法
  2. 将url转化为invoker对象

调用过程中invokers的使用

再调用过程中,AbstractClusterInvoker.invoke方法中,

public Result invoke(final Invocation invocation) throws RpcException {

    checkWhetherDestroyed();

    LoadBalance loadbalance;
    
    List<Invoker<T>> invokers = list(invocation); 
    if (invokers != null && invokers.size() > 0) {
        loadbalance = ExtensionLoader.getExtensionLoader(LoadBalance.class).getExtension(invokers.get(0).getUrl()
                .getMethodParameter(invocation.getMethodName(),Constants.LOADBALANCE_KEY, Constants.DEFAULT_LOADBALANCE));
    } else {
        loadbalance = ExtensionLoader.getExtensionLoader(LoadBalance.class).getExtension(Constants.DEFAULT_LOADBALANCE);
    }
    RpcUtils.attachInvocationIdIfAsync(getUrl(), invocation);
    return doInvoke(invocation, invokers, loadbalance);
}

list方法

从directory中获得invokers的集合,这里获取的方式基于netty,这一点我们简单的在上一节有所分析

protected  List<Invoker<T>> list(Invocation invocation) throws RpcException {
   List<Invoker<T>> invokers = directory.list(invocation);
   return invokers;
}

什么是Cluster?

关于cluster,我们之前已经多次提到,这里就不必重复造轮子了,引用网友的一篇文章:
Dubbo—Cluster剖析

LoadBalance

LoadBalance负载均衡, 负责从多个 Invokers中选出具体的一个Invoker用于本次调用,调用过程中包含了负载均衡的算法。

负载均衡代码访问入口

在AbstractClusterInvoker.invoke中代码如下,通过名称获得指定的扩展点。RandomLoadBalance

public Result invoke(final Invocation invocation) throws RpcException {

    checkWhetherDestroyed();

    LoadBalance loadbalance;
    
    List<Invoker<T>> invokers = list(invocation);
    if (invokers != null && invokers.size() > 0) {
        loadbalance = ExtensionLoader.getExtensionLoader(LoadBalance.class).getExtension(invokers.get(0).getUrl()
                .getMethodParameter(invocation.getMethodName(),Constants.LOADBALANCE_KEY, Constants.DEFAULT_LOADBALANCE));
    } else {
        loadbalance = ExtensionLoader.getExtensionLoader(LoadBalance.class).getExtension(Constants.DEFAULT_LOADBALANCE);
    }
    RpcUtils.attachInvocationIdIfAsync(getUrl(), invocation);
    return doInvoke(invocation, invokers, loadbalance);
}

AbstractClusterInvoker.doselect

调用LoadBalance.select方法,讲invokers按照指定算法进行负载

private Invoker<T> doselect(LoadBalance loadbalance, Invocation invocation, List<Invoker<T>> invokers, List<Invoker<T>> selected) throws RpcException {
    if (invokers == null || invokers.size() == 0)
        return null;
    if (invokers.size() == 1)
        return invokers.get(0);
    // 如果只有两个invoker,退化成轮循
    if (invokers.size() == 2 && selected != null && selected.size() > 0) {
        return selected.get(0) == invokers.get(0) ? invokers.get(1) : invokers.get(0);
    }
    Invoker<T> invoker = loadbalance.select(invokers, getUrl(), invocation);
    
    //如果 selected中包含(优先判断) 或者 不可用&&availablecheck=true 则重试.
    if( (selected != null && selected.contains(invoker))
            ||(!invoker.isAvailable() && getUrl()!=null && availablecheck)){
        try{
            Invoker<T> rinvoker = reselect(loadbalance, invocation, invokers, selected, availablecheck);
            if(rinvoker != null){
                invoker =  rinvoker;
            }else{
                //看下第一次选的位置,如果不是最后,选+1位置.
                int index = invokers.indexOf(invoker);
                try{
                    //最后在避免碰撞
                    invoker = index <invokers.size()-1?invokers.get(index+1) :invoker;
                }catch (Exception e) {
                    logger.warn(e.getMessage()+" may because invokers list dynamic change, ignore.",e);
                }
            }
        }catch (Throwable t){
            logger.error("clustor relselect fail reason is :"+t.getMessage() +" if can not slove ,you can set cluster.availablecheck=false in url",t);
        }
    }
    return invoker;
} 

默认情况下,LoadBalance使用的是Random算法,但是这个随机和我们理解上的随机还是不一样的,因为他还有个概念叫weight(权重)

RandomLoadBalance

假设有四个集群节点A,B,C,D,对应的权重分别是1,2,3,4,那么请求到A节点的概率就为1/(1+2+3+4) = 10%.B,C,D节点依次类推为20%,30%,40%.

protected <T> Invoker<T> doSelect(List<Invoker<T>> invokers, URL url, Invocation invocation) {
    int length = invokers.size(); // 总个数
    int totalWeight = 0; // 总权重
    boolean sameWeight = true; // 权重是否都一样
    for (int i = 0; i < length; i++) {
        int weight = getWeight(invokers.get(i), invocation);
        totalWeight += weight; // 累计总权重
        if (sameWeight && i > 0
                && weight != getWeight(invokers.get(i - 1), invocation)) {
            sameWeight = false; // 计算所有权重是否一样
        }
    }
    if (totalWeight > 0 && ! sameWeight) {
        // 如果权重不相同且权重大于0则按总权重数随机
        int offset = random.nextInt(totalWeight);
        // 并确定随机值落在哪个片断上
        for (int i = 0; i < length; i++) {
            offset -= getWeight(invokers.get(i), invocation);
            if (offset < 0) {
                return invokers.get(i);
            }
        }
    }
    // 如果权重相同或权重为0则均等随机
    return invokers.get(random.nextInt(length));
}

后记

Dubbo源码讲到这里就完结了,最后再给一张官方经典图:
在这里插入图片描述
Dubbo中文注释版:DubboV2.5.4下载地址

至此,关于Dubbo的介绍就告以段落了,下节预告:

分布式专题-分布式消息通信之ActiveMQ01-初识ActiveMQ
分布式专题-分布式消息通信之ActiveMQ02-ActiveMQ原理分析

更多架构知识,欢迎关注本套Java系列文章-地址导航Java架构师成长之路

;