In the last article, we learned what the startup phase of the dubbo service reference does, and what the remote call phase did not look at. The Dubbo service invocation process is complex, including many steps, such as sending requests, encoding and decoding, service degradation, filter chain processing, serialization, thread dispatch, and responding to requests. Due to space limitations, this article cannot analyze all the steps one by one. This article will focus on analyzing the process of sending and receiving requests, encoding and decoding, thread dispatching, and sending and receiving responses. As for service degradation, filter chaining, and serialization, you can analyze them yourself.
1. The agent object
Dubbo supports both synchronous and asynchronous calling methods, among which asynchronous calls can be further subdivided into asynchronous calls with "return value" and asynchronous calls with "no return value". The so-called "no return value" asynchronous call means that the service consumer only calls, but does not care about the result of the call. At this time, Dubbo will directly return an empty RpcResult. To use the asynchronous feature requires manual configuration by the service consumer. By default, Dubbo uses synchronous invocation.
In the last article, we analyzed that dubbo actually calls remote methods through the proxy class, so let's first look at the specific content of the proxy class through decompilation:
public class proxy0 implements ClassGenerator.DC, EchoService, DemoService {
// 方法数组
public static Method[] methods;
private InvocationHandler handler;
public proxy0(InvocationHandler invocationHandler) {
this.handler = invocationHandler;
}
public proxy0() {
}
public String sayHello(String string) {
// 将参数存储到 Object 数组中
Object[] arrobject = new Object[]{string};
// 调用 InvocationHandler 实现类的 invoke 方法得到调用结果
Object object = this.handler.invoke(this, methods[0], arrobject);
// 返回调用结果
return (String)object;
}
/** 回声测试方法 */
public Object $echo(Object object) {
Object[] arrobject = new Object[]{object};
Object object2 = this.handler.invoke(this, methods[1], arrobject);
return object2;
}
}
The logic of the proxy class is relatively simple. First, the runtime parameters are stored in the array, and then the invoke method of the InvocationHandlerinterface implementation class is called to get the call result, and finally the result is converted and returned to the caller.
2. InvocationHandler
Then look at the InvocationHandler.invoke()method
//InvokerInvocationHandler.java
public class InvokerInvocationHandler implements InvocationHandler {
private final Invoker<?> invoker;
public InvokerInvocationHandler(Invoker<?> handler) {
this.invoker = handler;
}
@Override
public Object invoke(Object proxy, Method method, Object[] args) throws Throwable {
//调用的目标方法
String methodName = method.getName();
//参数类型
Class<?>[] parameterTypes = method.getParameterTypes();
// 拦截定义在 Object 类中的方法（未被子类重写），比如 wait/notify
if (method.getDeclaringClass() == Object.class) {
return method.invoke(invoker, args);
}
// 如果 toString、hashCode 和 equals 等方法被子类重写了，这里也直接调用
if ("toString".equals(methodName) && parameterTypes.length == 0) {
return invoker.toString();
}
if ("hashCode".equals(methodName) && parameterTypes.length == 0) {
return invoker.hashCode();
}
if ("equals".equals(methodName) && parameterTypes.length == 1) {
return invoker.equals(args[0]);
}
// 将 method 和 args 封装到 RpcInvocation 中，并执行后续的调用
return invoker.invoke(new RpcInvocation(method, args)).recreate();
}
}
3. Service downgrade
Next MockClusterInvoker.invoke(), the method will be called. This class is a processing class for service degradation. Let's look at the specific code:
//MockClusterInvoker.java
public Result invoke(Invocation invocation) throws RpcException {
Result result = null;
// 获取 mock 配置值
String value = directory.getUrl().getMethodParameter(invocation.getMethodName(), Constants.MOCK_KEY, Boolean.FALSE.toString()).trim();
if (value.length() == 0 || value.equalsIgnoreCase("false")) {
// 无 mock 逻辑，直接调用其他 Invoker 对象的 invoke 方法，
// 比如 FailoverClusterInvoker
result = this.invoker.invoke(invocation);
} else if (value.startsWith("force")) {
// force:xxx 直接执行 mock 逻辑，不发起远程调用
result = doMockInvoke(invocation, null);
} else {
// fail:xxx 表示消费方对调用服务失败后，再执行 mock 逻辑，不抛出异常
try {
// 调用其他 Invoker 对象的 invoke 方法
result = this.invoker.invoke(invocation);
} catch (RpcException e) {
if (e.isBiz()) {
throw e;
} else {
// 调用失败，执行 mock 逻辑
result = doMockInvoke(invocation, e);
}
}
}
return result;
}
4. Obtaining available service providers and routes
The notes here are already very clear, so I won't repeat them. Let's seethis.invoker.invoke()
//AbstractClusterInvoker.java
/**
*
* @param invocation RpcInvocation(method,args)
* @return
* @throws RpcException
*/
@Override
public Result invoke(final Invocation invocation) throws RpcException {
checkWhetherDestroyed();
LoadBalance loadbalance = null;
// binding attachments into invocation.
//隐式传参
Map<String, String> contextAttachments = RpcContext.getContext().getAttachments();
if (contextAttachments != null && contextAttachments.size() != 0) {
((RpcInvocation) invocation).addAttachments(contextAttachments);
}
//获取可用的服务提供者list
List<Invoker<T>> invokers = list(invocation);
if (invokers != null && !invokers.isEmpty()) {
//获取负载均衡策略
loadbalance = ExtensionLoader.getExtensionLoader(LoadBalance.class).getExtension(invokers.get(0).getUrl()
.getMethodParameter(RpcUtils.getMethodName(invocation), Constants.LOADBALANCE_KEY, Constants.DEFAULT_LOADBALANCE));
}
RpcUtils.attachInvocationIdIfAsync(getUrl(), invocation);
return doInvoke(invocation, invokers, loadbalance);
}
The main content of this code is to get the list of available service providers
list(invocation)The method willDirectoryfirst obtain the service provider invokers collection that we previously stored in the service reference.- Then call
Routerthe interface, route these invokers, filter out the list of available service providers and return them.
Here's a brief look at the RegistryDirectory.doList()method:
//RegistryDirectory.java
public List<Invoker<T>> doList(Invocation invocation) {
if (forbidden) {
// 服务提供者关闭或禁用了服务，此时抛出 No provider 异常
throw new RpcException(RpcException.FORBIDDEN_EXCEPTION,
"No provider available from registry ...");
}
List<Invoker<T>> invokers = null;
// 获取 Invoker 本地缓存
Map<String, List<Invoker<T>>> localMethodInvokerMap = this.methodInvokerMap;
if (localMethodInvokerMap != null && localMethodInvokerMap.size() > 0) {
// 获取方法名和参数列表
String methodName = RpcUtils.getMethodName(invocation);
Object[] args = RpcUtils.getArguments(invocation);
// 检测参数列表的第一个参数是否为 String 或 enum 类型
if (args != null && args.length > 0 && args[0] != null
&& (args[0] instanceof String || args[0].getClass().isEnum())) {
// 通过 方法名 + 第一个参数名称 查询 Invoker 列表，具体的使用场景暂时没想到
invokers = localMethodInvokerMap.get(methodName + "." + args[0]);
}
if (invokers == null) {
// 通过方法名获取 Invoker 列表
invokers = localMethodInvokerMap.get(methodName);
}
if (invokers == null) {
// 通过星号 * 获取 Invoker 列表
invokers = localMethodInvokerMap.get(Constants.ANY_VALUE);
}
// 冗余逻辑，pull request #2861 移除了下面的 if 分支代码
if (invokers == null) {
Iterator<List<Invoker<T>>> iterator = localMethodInvokerMap.values().iterator();
if (iterator.hasNext()) {
invokers = iterator.next();
}
}
}
// 返回 Invoker 列表
return invokers == null ? new ArrayList<Invoker<T>>(0) : invokers;
}
The logic here is also very simple, that is, get the method name from the invocation, and then localMethodInvokerMapget the invoker corresponding to the method name according to the key.
5. Cluster fault tolerance
Let's go back to AbstractClusterInvoker.doInvoke()the method. After we have the service provider list and load balancing strategy, the next step is the fault tolerance stage. There are 5 types of dubbo fault tolerance mechanisms, such as Failove、Failfastetc. The default is Failoverto retry on failure. Take a look next FailoverClusterInvoker.doInvoke():
//FailoverClusterInvoker.java
@Override
public Result doInvoke(Invocation invocation, final List<Invoker<T>> invokers, LoadBalance loadbalance) throws RpcException {
List<Invoker<T>> copyinvokers = invokers;
checkInvokers(copyinvokers, invocation);
// 获取重试次数 len
int len = getUrl().getMethodParameter(invocation.getMethodName(), Constants.RETRIES_KEY, Constants.DEFAULT_RETRIES) + 1;
if (len <= 0) {
len = 1;
}
RpcException le = null;
List<Invoker<T>> invoked = new ArrayList<Invoker<T>>(copyinvokers.size());
Set<String> providers = new HashSet<String>(len);
// 循环调用，失败重试
for (int i = 0; i < len; i++) {
if (i > 0) {
checkWhetherDestroyed();
// 在进行重试前重新列举 Invoker，这样做的好处是，如果某个服务挂了，
// 通过调用 list 可得到最新可用的 Invoker 列表
copyinvokers = list(invocation);
// 对 copyinvokers 进行判空检查
checkInvokers(copyinvokers, invocation);
}
// 通过负载均衡选择 Invoker
Invoker<T> invoker = select(loadbalance, invocation, copyinvokers, invoked);
// 添加到 invoker 到 invoked 列表中
invoked.add(invoker);
// 设置 invoked 到 RPC 上下文中
RpcContext.getContext().setInvokers((List) invoked);
try {
// 调用目标 Invoker 的 invoke 方法
Result result = invoker.invoke(invocation);
return result;
} catch (RpcException e) {
//业务异常，不重试
if (e.isBiz()) {
throw e;
}
//记录异常
le = e;
} catch (Throwable e) {
//记录异常
le = new RpcException(e.getMessage(), e);
} finally {
providers.add(invoker.getUrl().getAddress());
}
}
// 若重试失败，则抛出异常
throw new RpcException(..., "Failed to invoke the method ...");
}
}
Get the number of retries, and then make a loop call based on the number of retries, and retry if it fails.
In the for loop, first select an Invoker through the load balancing component, and then make a remote call through the Invoker's invoke method. If it fails, log the exception and try again. When retrying, the list method of the parent class will be called again to enumerate the Invoker.
6. Load balancing
Invoker<T> invoker = select(loadbalance, invocation, copyinvokers, invoked)This code is the code for load balancing to select the invoker, so it will not be analyzed in detail here.
7. DubboInvoker
Let's look at the Result result = invoker.invoke(invocation);code above, where the method will eventually be executed DubboInvoker.doInvoke():
//DubboInvoker.java
public class DubboInvoker<T> extends AbstractInvoker<T> {
private final ExchangeClient[] clients;
protected Result doInvoke(final Invocation invocation) throws Throwable {
RpcInvocation inv = (RpcInvocation) invocation;
final String methodName = RpcUtils.getMethodName(invocation);
// 设置 path 和 version 到 attachment 中
inv.setAttachment(Constants.PATH_KEY, getUrl().getPath());
inv.setAttachment(Constants.VERSION_KEY, version);
ExchangeClient currentClient;
if (clients.length == 1) {
// 从 clients 数组中获取 ExchangeClient
currentClient = clients[0];
} else {
currentClient = clients[index.getAndIncrement() % clients.length];
}
try {
// 获取异步配置
boolean isAsync = RpcUtils.isAsync(getUrl(), invocation);
// isOneway 当前方法是否有返回值
boolean isOneway = RpcUtils.isOneway(getUrl(), invocation);
int timeout = getUrl().getMethodParameter(methodName, Constants.TIMEOUT_KEY, Constants.DEFAULT_TIMEOUT);
// 异步无返回值
if (isOneway) {
boolean isSent = getUrl().getMethodParameter(methodName, Constants.SENT_KEY, false);
// 发送请求
currentClient.send(inv, isSent);
// 设置上下文中的 future 字段为 null
RpcContext.getContext().setFuture(null);
// 返回一个空的 RpcResult
return new RpcResult();
}
// 异步有返回值
else if (isAsync) {
// 发送请求，并得到一个 ResponseFuture 实例
ResponseFuture future = currentClient.request(inv, timeout);
// 设置 future 到上下文中
RpcContext.getContext().setFuture(new FutureAdapter<Object>(future));
// 暂时返回一个空结果
return new RpcResult();
}
// 同步调用
else {
RpcContext.getContext().setFuture(null);
// 发送请求，得到一个 ResponseFuture 实例，并调用该实例的 get 方法进行等待
return (Result) currentClient.request(inv, timeout).get();
}
} catch (TimeoutException e) {
throw new RpcException(..., "Invoke remote method timeout....");
} catch (RemotingException e) {
throw new RpcException(..., "Failed to invoke remote method: ...");
}
}
// 省略其他方法
}
Here clientsis the encapsulation of the netty communication client we built when the service was referenced earlier.
The above code includes Dubbo's processing logic for synchronous and asynchronous calls, so I won't go into details here.
Next, look at currentClient.request(inv, timeout)the method. This method will eventually execute the following method through the execution chain of ReferenceCountExchangeClient.request -> HeaderExchangeClient.request -> HeaderExchangeChannel.request:
//HeaderExchangeChannel.java
@Override
public ResponseFuture request(Object request, int timeout) throws RemotingException {
if (closed) {
throw new RemotingException(..., "Failed to send request ...);
}
// 创建 Request 对象
Request req = new Request();
req.setVersion(Version.getProtocolVersion());
// 设置双向通信标志为 true
req.setTwoWay(true);
// 这里的 request 变量类型为 RpcInvocation
req.setData(request);
// 创建 DefaultFuture 对象
DefaultFuture future = new DefaultFuture(channel, req, timeout);
try {
// 调用 NettyClient 的 send 方法发送请求
channel.send(req);
} catch (RemotingException e) {
future.cancel();
throw e;
}
// 返回 DefaultFuture 对象
return future;
}
At this point, everyone finally sees the Request semantics. The above method first defines a Request object, and then passes the object to NettyClientthe send method of , for subsequent encoding, serialization, and netty calls.
There is no detailed analysis here, and interested students can learn more about it.
8. The service provider accepts the call
8.1
The process of initiating remote calls by service consumers has been analyzed earlier. Next, let's take a look at the process of receiving calls by service providers.
As mentioned earlier, Dubbo uses Netty as the underlying communication framework by default. After Netty detects that there is data inbound, it will first decode the data through the decoder, and pass the decoded data to the specified method of the next inbound processor.
Call DubboCodec.decodeBody()the method to decode and encapsulate the decoded fields into Request.
The DecodeableRpcInvocation.decode()method for subsequent decoding work.
After this work is completed, the calling method name, attachment, and calling parameters can be parsed out, and then these parameters are filled into DecodeableRpcInvocation, and then DecodeableRpcInvocation is filled into the Requestobject.
After the decoder parses the data packet into a Request object, the messageReceived method of NettyHandler will receive the object and pass it down.
During this period, the object will be passed to NettyServer, MultiMessageHandler, HeartbeatHandler and AllChannelHandler in turn.
Finally, AllChannelHandlerthe object is encapsulated into the Runnable implementation class object, and the Runnable is put into the thread pool to execute the subsequent calling logic. The entire call stack is as follows:
NettyHandler#messageReceived(ChannelHandlerContext, MessageEvent)
—> AbstractPeer#received(Channel, Object)
—> MultiMessageHandler#received(Channel, Object)
—> HeartbeatHandler#received(Channel, Object)
—> AllChannelHandler#received(Channel, Object)
—> ExecutorService#execute(Runnable) // 由线程池执行后续的调用逻辑
8.2、AllChannelHandler
Every method in the call stack is not analyzed here. Here we directly analyze the logic of the last call method in the call stack. as follows:
//AllChannelHandler.java
/** 处理请求和响应消息，这里的 message 变量类型可能是 Request，也可能是 Response */
@Override
public void received(Channel channel, Object message) throws RemotingException {
ExecutorService cexecutor = getExecutorService();
try {
// 将请求和响应消息派发到线程池中处理
cexecutor.execute(new ChannelEventRunnable(channel, handler, ChannelState.RECEIVED, message));
} catch (Throwable t) {
if(message instanceof Request && t instanceof RejectedExecutionException){
Request request = (Request)message;
// 如果通信方式为双向通信，此时将 Server side ... threadpool is exhausted
// 错误信息封装到 Response 中，并返回给服务消费方。
if(request.isTwoWay()){
String msg = "Server side(" + url.getIp() + "," + url.getPort()
+ ") threadpool is exhausted ,detail msg:" + t.getMessage();
Response response = new Response(request.getId(), request.getVersion());
response.setStatus(Response.SERVER_THREADPOOL_EXHAUSTED_ERROR);
response.setErrorMessage(msg);
// 返回包含错误信息的 Response 对象
channel.send(response);
return;
}
}
throw new ExecutionException(..., " error when process received event .", t);
}
}
8.3、ChannelEventRunnable
As above, the request object will be encapsulated in ChannelEventRunnable, which will be a new starting point for the service invocation process. So let's explore downwards ChannelEventRunnablefrom the starting point.
//ChannelEventRunnable.java
@Override
public void run() {
// 检测通道状态，对于请求或响应消息，此时 state = RECEIVED
if (state == ChannelState.RECEIVED) {
try {
// 将 channel 和 message 传给 ChannelHandler 对象，进行后续的调用
handler.received(channel, message);
} catch (Exception e) {
logger.warn("... operation error, channel is ... message is ...");
}
}
// 其他消息类型通过 switch 进行处理
else {
switch (state) {
case CONNECTED:
case DISCONNECTED:
// ...
case SENT:
// ...
case CAUGHT:
// ...
default:
logger.warn("unknown state: " + state + ", message is " + message);
}
}
8.4、DecodeHandler
ChannelEventRunnable is only a transit station, its run method does not contain specific calling logic, it is only used to pass parameters to other ChannelHandler objects for processing, the object type is DecodeHandler.
//DecodeHandler.java
@Override
public void received(Channel channel, Object message) throws RemotingException {
if (message instanceof Decodeable) {
// 对 Decodeable 接口实现类对象进行解码
decode(message);
}
if (message instanceof Request) {
// 对 Request 的 data 字段进行解码
decode(((Request) message).getData());
}
if (message instanceof Response) {
// 对 Request 的 result 字段进行解码
decode(((Response) message).getResult());
}
// 执行后续逻辑
handler.received(channel, message);
}
8.5、HeaderExchangeHandler
DecodeHandler mainly contains some decoding logic. When analyzing request decoding in Section 8.1, it was said that request decoding can be performed on the IO thread or in the thread pool, depending on the runtime configuration.
The meaning of DecodeHandler is to ensure that the request or response object can be decoded in the thread pool. After decoding, the fully decoded Request object will continue to be passed backwards, and the next stop is HeaderExchangeHandler.
//HeaderExchangeHandler.java
public class HeaderExchangeHandler implements ChannelHandlerDelegate {
private final ExchangeHandler handler;
@Override
public void received(Channel channel, Object message) throws RemotingException {
channel.setAttribute(KEY_READ_TIMESTAMP, System.currentTimeMillis());
ExchangeChannel exchangeChannel = HeaderExchangeChannel.getOrAddChannel(channel);
try {
// 处理请求对象
if (message instanceof Request) {
Request request = (Request) message;
if (request.isEvent()) {
// 处理事件
handlerEvent(channel, request);
}
// 处理普通的请求
else {
// 双向通信
if (request.isTwoWay()) {
// 向后调用服务，并得到调用结果
Response response = handleRequest(exchangeChannel, request);
// 将调用结果返回给服务消费端
channel.send(response);
}
// 如果是单向通信，仅向后调用指定服务即可，无需返回调用结果
else {
handler.received(exchangeChannel, request.getData());
}
}
}
// 处理响应对象，服务消费方会执行此处逻辑，后面分析
else if (message instanceof Response) {
handleResponse(channel, (Response) message);
} else if (message instanceof String) {
// telnet 相关，忽略
} else {
handler.received(exchangeChannel, message);
}
} finally {
HeaderExchangeChannel.removeChannelIfDisconnected(channel);
}
}
Response handleRequest(ExchangeChannel channel, Request req) throws RemotingException {
Response res = new Response(req.getId(), req.getVersion());
// 检测请求是否合法，不合法则返回状态码为 BAD_REQUEST 的响应
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);
// 设置 BAD_REQUEST 状态
res.setStatus(Response.BAD_REQUEST);
return res;
}
// 获取 data 字段值，也就是 RpcInvocation 对象
Object msg = req.getData();
try {
//重点！！ 继续向下调用
Object result = handler.reply(channel, msg);
// 设置 OK 状态码
res.setStatus(Response.OK);
// 设置调用结果
res.setResult(result);
} catch (Throwable e) {
// 若调用过程出现异常，则设置 SERVICE_ERROR，表示服务端异常
res.setStatus(Response.SERVICE_ERROR);
res.setErrorMessage(StringUtils.toString(e));
}
return res;
}
}
At this point, we have seen relatively clear request and response logic. For two-way communication, HeaderExchangeHandler first calls backwards to get the result of the call. Then encapsulate the call result into the Response object, and finally return the object to the service consumer.
If the request is invalid, or the call fails, the error information will be encapsulated in the Response object and returned to the service consumer.
8.6、DubboProtocol.reply()
Let's look at the method next , where the method handler.reply(channel, msg)will eventually be calledDubboProtocol.reply()
//DubboProtocol.java
public class DubboProtocol extends AbstractProtocol {
public static final String NAME = "dubbo";
private ExchangeHandler requestHandler = new ExchangeHandlerAdapter() {
@Override
public Object reply(ExchangeChannel channel, Object message) throws RemotingException {
if (message instanceof Invocation) {
Invocation inv = (Invocation) message;
// 获取 Invoker 实例
Invoker<?> invoker = getInvoker(channel, inv);
if (Boolean.TRUE.toString().equals(inv.getAttachments().get(IS_CALLBACK_SERVICE_INVOKE))) {
// 回调相关，忽略
}
RpcContext.getContext().setRemoteAddress(channel.getRemoteAddress());
// 通过 Invoker 调用具体的服务
return invoker.invoke(inv);
}
throw new RemotingException(channel, "Unsupported request: ...");
}
// 忽略其他方法
}
Invoker<?> getInvoker(Channel channel, Invocation inv) throws RemotingException {
// 忽略回调和本地存根相关逻辑
// ...
int port = channel.getLocalAddress().getPort();
// 计算 service key，格式为 groupName/serviceName:serviceVersion:port。比如：
// dubbo/com.alibaba.dubbo.demo.DemoService:1.0.0:20880
String serviceKey = serviceKey(port, path, inv.getAttachments().get(Constants.VERSION_KEY), inv.getAttachments().get(Constants.GROUP_KEY));
// 从 exporterMap 查找与 serviceKey 相对应的 DubboExporter 对象，
// 服务导出过程中会将 <serviceKey, DubboExporter> 映射关系存储到 exporterMap 集合中
DubboExporter<?> exporter = (DubboExporter<?>) exporterMap.get(serviceKey);
if (exporter == null)
throw new RemotingException(channel, "Not found exported service ...");
// 获取 Invoker 对象，并返回
return exporter.getInvoker();
}
// 忽略其他方法
}
The above logic is used to obtain the Invoker instance corresponding to the specified service, and invoke the service logic through the invoke method of the Invoker.
The invoke method is defined in AbstractProxyInvoker with the following code:
public abstract class AbstractProxyInvoker<T> implements Invoker<T> {
@Override
public Result invoke(Invocation invocation) throws RpcException {
try {
// 调用 doInvoke 执行后续的调用，并将调用结果封装到 RpcResult 中，并
return new RpcResult(doInvoke(proxy, invocation.getMethodName(), invocation.getParameterTypes(), invocation.getArguments()));
} catch (InvocationTargetException e) {
return new RpcResult(e.getTargetException());
} catch (Throwable e) {
throw new RpcException("Failed to invoke remote proxy method ...");
}
}
protected abstract Object doInvoke(T proxy, String methodName, Class<?>[] parameterTypes, Object[] arguments) throws Throwable;
}
As above, doInvoke is an abstract method, which needs to be implemented by a concrete Invoker instance. The Invoker instance is created at runtime through JavassistProxyFactory, and the creation logic is as follows:
public class JavassistProxyFactory extends AbstractProxyFactory {
// 省略其他方法
@Override
public <T> Invoker<T> getInvoker(T proxy, Class<T> type, URL url) {
final Wrapper wrapper = Wrapper.getWrapper(proxy.getClass().getName().indexOf('$') < 0 ? proxy.getClass() : type);
// 创建匿名类对象
return new AbstractProxyInvoker<T>(proxy, type, url) {
@Override
protected Object doInvoke(T proxy, String methodName,
Class<?>[] parameterTypes,
Object[] arguments) throws Throwable {
// 调用 invokeMethod 方法进行后续的调用
return wrapper.invokeMethod(proxy, methodName, parameterTypes, arguments);
}
};
}
}
Wrapper is an abstract class, where invokeMethod is an abstract method.
Dubbo will generate implementation classes for Wrapper through the Javassist framework at runtime, and implement the invokeMethod method, which will eventually call specific services according to the call information.
8.7、wrapper.invokeMethod()
Taking DemoServiceImpl as an example, the proxy class generated by Javassist is as follows.
/** Wrapper0 是在运行时生成的，大家可使用 Arthas 进行反编译 */
public class Wrapper0 extends Wrapper implements ClassGenerator.DC {
public static String[] pns;
public static Map pts;
public static String[] mns;
public static String[] dmns;
public static Class[] mts0;
// 省略其他方法
public Object invokeMethod(Object object, String string, Class[] arrclass, Object[] arrobject) throws InvocationTargetException {
DemoService demoService;
try {
// 类型转换
demoService = (DemoService)object;
}
catch (Throwable throwable) {
throw new IllegalArgumentException(throwable);
}
try {
// 根据方法名调用指定的方法
if ("sayHello".equals(string) && arrclass.length == 1) {
return demoService.sayHello((String)arrobject[0]);
}
}
catch (Throwable throwable) {
throw new InvocationTargetException(throwable);
}
throw new NoSuchMethodException(new StringBuffer().append("Not found method \"").append(string).append("\" in class com.alibaba.dubbo.demo.DemoService.").toString());
}
}
The following figure is the attribute information viewed by debug, and the structure of invoker, wrapper, and proxy can be clearly seen.
At this point, the entire service invocation process is analyzed.
If it is still not clear, it is recommended to read the article Talking about RPC, the analysis is very thorough.
最后
Arranged a flow chart of dubbo service reference for easy understanding.
Reference article:
Dubbo official website