4

SpringCloud(第 049 篇)Netflix Eureka 源码深入剖析(上)

-

一、大致介绍

1、鉴于一些朋友的提问并提议讲解下eureka的源码分析,由此应运而产生的本章节的内容;
2、所以我站在自我的理解角度试着整理了这篇Eureka源码的分析,希望对大家有所帮助;
3、由于篇幅太长不能在一篇里面发布出来,所以拆分了上下篇;

二、基本原理

1、Eureka Server 提供服务注册服务,各个节点启动后,会在Eureka Server中进行注册,这样Eureka Server中的服务注册表中将会存储所有可用服务节点的信息,服务节点的信息可以在界面中直观的看到。
2、Eureka Client 是一个Java 客户端,用于简化与Eureka Server的交互,客户端同时也具备一个内置的、使用轮询负载算法的负载均衡器。
3、在应用启动后,将会向Eureka Server发送心跳(默认周期为30秒),如果Eureka Server在多个心跳周期没有收到某个节点的心跳,Eureka Server 将会从服务注册表中把这个服务节点移除(默认90秒)。
4、Eureka Server之间将会通过复制的方式完成数据的同步;
5、Eureka Client具有缓存的机制,即使所有的Eureka Server 都挂掉的话,客户端依然可以利用缓存中的信息消费其它服务的API;

三、EurekaServer 启动流程分析

3.1 跑一下 springms-discovery-eureka 代码,不难发现,我们会看到一些有关 EurekaServer 启动的流程日志;

2017-10-22 18:14:17.635  INFO 5288 --- [           main] o.s.j.e.a.AnnotationMBeanExporter        : Located managed bean 'environmentManager': registering with JMX server as MBean [org.springframework.cloud.context.environment:name=environmentManager,type=EnvironmentManager]
2017-10-22 18:14:17.650  INFO 5288 --- [           main] o.s.j.e.a.AnnotationMBeanExporter        : Located managed bean 'restartEndpoint': registering with JMX server as MBean [org.springframework.cloud.context.restart:name=restartEndpoint,type=RestartEndpoint]
2017-10-22 18:14:17.661  INFO 5288 --- [           main] o.s.j.e.a.AnnotationMBeanExporter        : Located managed bean 'refreshScope': registering with JMX server as MBean [org.springframework.cloud.context.scope.refresh:name=refreshScope,type=RefreshScope]
2017-10-22 18:14:17.674  INFO 5288 --- [           main] o.s.j.e.a.AnnotationMBeanExporter        : Located managed bean 'configurationPropertiesRebinder': registering with JMX server as MBean [org.springframework.cloud.context.properties:name=configurationPropertiesRebinder,context=335b5620,type=ConfigurationPropertiesRebinder]
2017-10-22 18:14:17.683  INFO 5288 --- [           main] o.s.j.e.a.AnnotationMBeanExporter        : Located managed bean 'refreshEndpoint': registering with JMX server as MBean [org.springframework.cloud.endpoint:name=refreshEndpoint,type=RefreshEndpoint]
2017-10-22 18:14:17.926  INFO 5288 --- [           main] o.s.c.support.DefaultLifecycleProcessor  : Starting beans in phase 0
2017-10-22 18:14:17.927  INFO 5288 --- [           main] c.n.e.EurekaDiscoveryClientConfiguration : Registering application unknown with eureka with status UP
2017-10-22 18:14:17.927  INFO 5288 --- [      Thread-10] o.s.c.n.e.server.EurekaServerBootstrap   : Setting the eureka configuration..
2017-10-22 18:14:17.948  INFO 5288 --- [      Thread-10] o.s.c.n.e.server.EurekaServerBootstrap   : isAws returned false
2017-10-22 18:14:17.949  INFO 5288 --- [      Thread-10] o.s.c.n.e.server.EurekaServerBootstrap   : Initialized server context
2017-10-22 18:14:17.949  INFO 5288 --- [      Thread-10] c.n.e.r.PeerAwareInstanceRegistryImpl    : Got 1 instances from neighboring DS node
2017-10-22 18:14:17.949  INFO 5288 --- [      Thread-10] c.n.e.r.PeerAwareInstanceRegistryImpl    : Renew threshold is: 1
2017-10-22 18:14:17.949  INFO 5288 --- [      Thread-10] c.n.e.r.PeerAwareInstanceRegistryImpl    : Changing status to UP
2017-10-22 18:14:17.958  INFO 5288 --- [      Thread-10] e.s.EurekaServerInitializerConfiguration : Started Eureka Server
2017-10-22 18:14:18.019  INFO 5288 --- [           main] s.b.c.e.t.TomcatEmbeddedServletContainer : Tomcat started on port(s): 8761 (http)
2017-10-22 18:14:18.020  INFO 5288 --- [           main] c.n.e.EurekaDiscoveryClientConfiguration : Updating port to 8761
2017-10-22 18:14:18.023  INFO 5288 --- [           main] c.s.cloud.EurekaServerApplication        : Started EurekaServerApplication in 8.299 seconds (JVM running for 8.886)
【【【【【【 Eureka微服务 】】】】】】已启动.

【分析】:发现有这么一句日志打印“Setting the eureka configuration..”,eureka 开始进行配置,说不定也许就是Eureka Server 流程启动的开
始呢?我们抱着怀疑的心态进入这行日志打印的EurekaServerBootstrap类去看看。

3.2 进入 EurekaServerBootstrap 类看看,看这个类的名字,见名知意,应该就是 EurekaServer 的启动类了;

protected void initEurekaEnvironment() throws Exception {
    log.info("Setting the eureka configuration..");
    。。。
}

【分析一】:我们看到日志在 initEurekaEnvironment 方法中被打印出来,然后我顺着这个方法寻找该方法被调用的地方;

public void contextInitialized(ServletContext context) {
    try {
        initEurekaEnvironment();
        initEurekaServerContext();

        context.setAttribute(EurekaServerContext.class.getName(), this.serverContext);
    }
    catch (Throwable e) {
        log.error("Cannot bootstrap eureka server :", e);
        throw new RuntimeException("Cannot bootstrap eureka server :", e);
    }
}

【分析二】:接着发现 contextInitialized 这个方法里面调用了 initEurekaEnvironment 方法,接着我们再往上层寻找被调用的地方;

【分析三】:接着我们看到 EurekaServerInitializerConfiguration 类中有个 start 方法,该方法创建了一个线程来后台执行 EurekaServer 的初始化流程;

3.3 进入 EurekaServerInitializerConfiguration 方法,看看这个所谓的 EurekaServer 初始化配置做了哪些事情?

@Override
public void start() { // 打上断点
    new Thread(new Runnable() {
        @Override
        public void run() {
            try {
                //TODO: is this class even needed now?
                eurekaServerBootstrap.contextInitialized(EurekaServerInitializerConfiguration.this.servletContext);
                log.info("Started Eureka Server");

                publish(new EurekaRegistryAvailableEvent(getEurekaServerConfig()));
                EurekaServerInitializerConfiguration.this.running = true;
                publish(new EurekaServerStartedEvent(getEurekaServerConfig()));
            }
            catch (Exception ex) {
                // Help!
                log.error("Could not initialize Eureka servlet context", ex);
            }
        }
    }).start();
}

【分析一】:看到 log.info("Started Eureka Server"); 这行代码,相信大家已经释然了,这里就是所谓的启动了 EurekaServer 了,其实也就是 
eurekaServerBootstrap.contextInitialized(EurekaServerInitializerConfiguration.this.servletContext) 初始化了一些我们未知的东西;

【分析二】:当打印完启动Eureka Server日志后,调用了两次 publish 方法,该方法最终调用的是 this.applicationContext.publishEvent
(event) 方法,目的是利用Spring中ApplicationContext对事件传递性质,事件发布者(applicationContext)来发布事件(event),但是缺少的是监听
者,其实你仔细搜索下代码,发现好像没有地方对 EurekaServerStartedEvent、EurekaRegistryAvailableEvent 进行监听,奇了怪了,这是咋了呢?

【分析三】:然后找到 EurekaServerStartedEvent 所在的目录下,EurekaInstanceCanceledEvent、EurekaInstanceRegisteredEvent、
EurekaInstanceRenewedEvent、EurekaRegistryAvailableEvent、EurekaServerStartedEvent 有这么几个事件的类,服务下线事件、服务注册事
件、服务续约事件、注册中心启动事件、Eureka Server启动事件,这么几个事件都没有被监听,那么我们是不是给添加上监听,是不是就可以了呢?像这样
 @EventListener  public void listen(EurekaInstanceCanceledEvent event) { 。。。处下线逻辑 },添加 EventListener 监听注解,就可
以在我们自己的代码逻辑中收到这个事件的回调了,所以想想SpringCloud还是挺机制的,提供回调接口让我们自己实现自己的业务逻辑,真心不错;

【分析四】:那么反过来想想,为啥会无缘无故 start 方法就被调用了呢?那么反向继续向上找调用 start 方法的地方,结果找到了 
DefaultLifecycleProcessor类的doStart方法调用了 bean.start(); 这么一段代码;

3.4 进入 DefaultLifecycleProcessor 类看看,这个 EurekaServerInitializerConfiguration.start 方法是如何被触发的?

private void doStart(Map<String, ? extends Lifecycle> lifecycleBeans, String beanName, boolean autoStartupOnly) {
    // 打上断点
    Lifecycle bean = lifecycleBeans.remove(beanName);
    if (bean != null && !this.equals(bean)) {
        String[] dependenciesForBean = this.beanFactory.getDependenciesForBean(beanName);
        for (String dependency : dependenciesForBean) {
            doStart(lifecycleBeans, dependency, autoStartupOnly);
        }
        if (!bean.isRunning() &&
                (!autoStartupOnly || !(bean instanceof SmartLifecycle) || ((SmartLifecycle) bean).isAutoStartup())) {
            if (logger.isDebugEnabled()) {
                logger.debug("Starting bean '" + beanName + "' of type [" + bean.getClass() + "]");
            }
            try {
                bean.start();
            }
            catch (Throwable ex) {
                throw new ApplicationContextException("Failed to start bean '" + beanName + "'", ex);
            }
            if (logger.isDebugEnabled()) {
                logger.debug("Successfully started bean '" + beanName + "'");
            }
        }
    }
}

【分析一】:看到在 bean.isRunning 等一系列状态的判断下才去调用 bean.start() 方法的,我们再往上寻找被调用地方;

public void start() {
    // 打上断点
    if (this.members.isEmpty()) {
        return;
    }
    if (logger.isInfoEnabled()) {
        logger.info("Starting beans in phase " + this.phase);
    }
    Collections.sort(this.members);
    for (LifecycleGroupMember member : this.members) {
        if (this.lifecycleBeans.containsKey(member.name)) {
            doStart(this.lifecycleBeans, member.name, this.autoStartupOnly);
        }
    }
}

【分析二】:该类是DefaultLifecycleProcessor中内部类LifecycleGroup的一个方法,再往上寻找调用方;

private void startBeans(boolean autoStartupOnly) {
    Map<String, Lifecycle> lifecycleBeans = getLifecycleBeans();
    Map<Integer, LifecycleGroup> phases = new HashMap<Integer, LifecycleGroup>();
    for (Map.Entry<String, ? extends Lifecycle> entry : lifecycleBeans.entrySet()) {
        Lifecycle bean = entry.getValue();
        if (!autoStartupOnly || (bean instanceof SmartLifecycle && ((SmartLifecycle) bean).isAutoStartup())) {
            int phase = getPhase(bean);
            LifecycleGroup group = phases.get(phase);
            if (group == null) {
                group = new LifecycleGroup(phase, this.timeoutPerShutdownPhase, lifecycleBeans, autoStartupOnly);
                phases.put(phase, group);
            }
            group.add(entry.getKey(), bean);
        }
    }
    if (phases.size() > 0) {
        List<Integer> keys = new ArrayList<Integer>(phases.keySet());
        Collections.sort(keys);
        for (Integer key : keys) {
            phases.get(key).start();
        }
    }
}

【分析三】:startBeans 属于 DefaultLifecycleProcessor 类的一个私有方法,startBeans 方法第一行就是获取 getLifecycleBeans() 生命周期
Bean对象,由此可见似乎 Eureka Server 之所以会被启动,是不是实现了某个接口或者重写了某个方法,才会导致由于容易在初始化的过程中因调用某些特
殊方法或者某些类才启动的,因此我们回头去看看 EurekaServerInitializerConfiguration 这个类;

【分析四】:结果发现 EurekaServerInitializerConfiguration 这个类实现了 SmartLifecycle 这么样的一个接口,而 SmartLifecycle 接口又继
承了 Lifecycle 生命周期接口类,所以真想已经重见天日了,原来是实现了 Lifecycle 这样的一个接口,然后实现了 start 方法,因此 Eureka 
Server 就这么稀里糊涂的就被莫名其妙的启动起来了?

3.5 到这里难道就真的完了么?难道Eureka Server启动就干这么点点事情?不可能吧?

【分析一】:我们之前仅仅只是通过了日志来逆向分析,但是我们是不是忘了我们本应该标志是Eureka Server的这个注解了呢?没错,我们在分析的过程中
已经将 @EnableEurekaServer 这个注解遗忘了,那么我们现在先回到这个注解类来看看;

3.6 进入 EnableEurekaServer 类,看看究竟干了啥?

@Target(ElementType.TYPE)
@Retention(RetentionPolicy.RUNTIME)
@Documented
@Import(EurekaServerConfiguration.class)
public @interface EnableEurekaServer {

}

【分析一】:我们不难发现 EnableEurekaServer 类上有个 @Import 注解,引用了一个 class 文件,由此我们进入观察;

3.7 进入 EurekaServerConfiguration 类看看,看名称的话,理解的意思大概就是 Eureka Server 配置类;

【分析一】:果不其然,这个类有很多 @Bean、@Configuration 注解过的方法,那是不是我们可以认为刚才 3.1~3.4 的推论是不是就是由于被实例化了这么一个 Bean,然后就慢慢的调用到了 start 方法了呢?

【分析二】:搜索 “Bootstrap” 字样,还真发现了有这么一个方法;

@Bean
public EurekaServerBootstrap eurekaServerBootstrap(PeerAwareInstanceRegistry registry,
        EurekaServerContext serverContext) {
    return new EurekaServerBootstrap(this.applicationInfoManager,
            this.eurekaClientConfig, this.eurekaServerConfig, registry,
            serverContext);
}

【分析三】:既然有这么一个 Bean,那么是不是和刚开始顺着日志逆向分析也是有一定道理的,没有这么一个Bean的存在,那么 DefaultLifecycleProcessor.startBeans 方法中 getLifecycleBeans 的这个也就没那么顺畅被找到了呢?不过我的猜想是这样的,本人没有将源码下载下来,将 eurekaServerBootstrap 方法中的 @Bean 注解注释掉试试,不过推理起来也八九不离十,这个疑问悬念就留给大家尝试尝试吧;

【分析四】:既然找到了一个 @Bean 注解过的方法,那我们再找找其他的一些被注解过的方法,比如一些通用全局用的类似词眼,比如 Context,Bean,Init、Server 之类的;

@Bean
public EurekaServerContext eurekaServerContext(ServerCodecs serverCodecs,
        PeerAwareInstanceRegistry registry, PeerEurekaNodes peerEurekaNodes) {
    return new DefaultEurekaServerContext(this.eurekaServerConfig, serverCodecs,
            registry, peerEurekaNodes, this.applicationInfoManager);
}

@Bean
public PeerEurekaNodes peerEurekaNodes(PeerAwareInstanceRegistry registry,
        ServerCodecs serverCodecs) {
    return new PeerEurekaNodes(registry, this.eurekaServerConfig,
            this.eurekaClientConfig, serverCodecs, this.applicationInfoManager);
}

@Bean
public PeerAwareInstanceRegistry peerAwareInstanceRegistry(
        ServerCodecs serverCodecs) {
    this.eurekaClient.getApplications(); // force initialization
    return new InstanceRegistry(this.eurekaServerConfig, this.eurekaClientConfig,
            serverCodecs, this.eurekaClient,
            this.instanceRegistryProperties.getExpectedNumberOfRenewsPerMin(),
            this.instanceRegistryProperties.getDefaultOpenForTrafficCount());
}

@Bean
@ConditionalOnProperty(prefix = "eureka.dashboard", name = "enabled", matchIfMissing = true)
public EurekaController eurekaController() {
    return new EurekaController(this.applicationInfoManager);
}

【分析五】:DefaultEurekaServerContext.initialize 初始化了一些东西,现在还不知道干啥用的,先放这里,打上断点;

【分析六】:PeerEurekaNodes.start 方法,又是一个 start 方法,但是该类没有实现任何类,姑且先放这里,打上断点;

【分析七】:InstanceRegistry.register 方法,而且还有几个呢,可能是客户端注册用的,也先放这里,都打上断点,或者将 这个类的所有方法都断点上,断点打完后发现有注册的,有续约的,有注销的;

【分析八】:打完这些断点后,感觉没有思路了,索性就断点跑一把,看看有什么新的发现点;

3.8 停止服务,Debug 跑一下 springms-discovery-eureka 代码;

【分析一】:DefaultEurekaServerContext.initialize 方法被调用了,证实了刚才想法,EurekaServerConfiguration 不是白写的,还是有它的作用的;

@PostConstruct
@Override
public void initialize() throws Exception {
    logger.info("Initializing ...");
    peerEurekaNodes.start();
    registry.init(peerEurekaNodes);
    logger.info("Initialized");
}

【分析二】:进入 initialize 方法中 peerEurekaNodes.start();

public void start() {
    taskExecutor = Executors.newSingleThreadScheduledExecutor(
            new ThreadFactory() {
                @Override
                public Thread newThread(Runnable r) {
                    Thread thread = new Thread(r, "Eureka-PeerNodesUpdater");
                    thread.setDaemon(true);
                    return thread;
                }
            }
    );
    try {
        updatePeerEurekaNodes(resolvePeerUrls());
        Runnable peersUpdateTask = new Runnable() {
            @Override
            public void run() {
                try {
                    updatePeerEurekaNodes(resolvePeerUrls());
                } catch (Throwable e) {
                    logger.error("Cannot update the replica Nodes", e);
                }

            }
        };
        // 注释:间隔 600000 毫秒,即 10分钟 间隔执行一次服务集群数据同步;
        taskExecutor.scheduleWithFixedDelay(
                peersUpdateTask,
                serverConfig.getPeerEurekaNodesUpdateIntervalMs(),
                serverConfig.getPeerEurekaNodesUpdateIntervalMs(),
                TimeUnit.MILLISECONDS
        );
    } catch (Exception e) {
        throw new IllegalStateException(e);
    }
    for (PeerEurekaNode node : peerEurekaNodes) {
        logger.info("Replica node URL:  " + node.getServiceUrl());
    }
}

【分析三】: start 方法中会看到一个定时调度的任务,updatePeerEurekaNodes(resolvePeerUrls()); 间隔 600000 毫秒,即 10分钟 间隔执行一次服务集群数据同步;

【分析四】: 然后断点放走放下走,进入 initialize 方法中 registry.init(peerEurekaNodes);

@Override
public void init(PeerEurekaNodes peerEurekaNodes) throws Exception {
    this.numberOfReplicationsLastMin.start();
    this.peerEurekaNodes = peerEurekaNodes;
    // 注释:初始化 Eureka Server 响应缓存,默认缓存时间为30s
    initializedResponseCache();
    // 注释:定时任务,多久重置一下心跳阈值,900000 毫秒,即 15分钟 的间隔时间,会重置心跳阈值
    scheduleRenewalThresholdUpdateTask();
    // 注释:初始化远端注册
    initRemoteRegionRegistry();

    try {
        Monitors.registerObject(this);
    } catch (Throwable e) {
        logger.warn("Cannot register the JMX monitor for the InstanceRegistry :", e);
    }
}

【分析五】: 缓存也配置好了,定时任务也配置好了,似乎应该没啥了,那么我们把断点放开,看看下一步会走到哪里?

3.9 EurekaServerInitializerConfiguration.start 也进断点了。

【分析一】:先是 DefaultLifecycleProcessor.doStart 方法进断点,然后才是 EurekaServerInitializerConfiguration.start 方法进断点;

【分析二】:再一次证明刚刚的逆向分析仅仅只是缺了个从头EnableEurekaServer分析罢了,但是最终方法论分析思路还是对的,由于开始分析过这里,然而我们就跳过,继续放开断点向后继续看看;

3.10 InstanceRegistry.openForTraffic 也进断点了。

【分析一】:这不就是我们刚才在 “步骤3.7之分析七” 打的断点么?看下堆栈信息,正是 “步骤3.2之分析一” 中 initEurekaServerContext 方法中有
这么一句 this.registry.openForTraffic(this.applicationInfoManager, registryCount); 调用到了,因果轮回,代码千变万化,打上断点还有有好处的,结果还是回到了开始日志逆向分析的地方。

【分析二】:进入 super.openForTraffic 方法;

@Override
public void openForTraffic(ApplicationInfoManager applicationInfoManager, int count) {
    // Renewals happen every 30 seconds and for a minute it should be a factor of 2.
    // 注释:每30秒续约一次,那么每分钟续约就是2次,所以才是 count * 2 的结果;
    this.expectedNumberOfRenewsPerMin = count * 2;
    this.numberOfRenewsPerMinThreshold =
            (int) (this.expectedNumberOfRenewsPerMin * serverConfig.getRenewalPercentThreshold());
    logger.info("Got " + count + " instances from neighboring DS node");
    logger.info("Renew threshold is: " + numberOfRenewsPerMinThreshold);
    this.startupTime = System.currentTimeMillis();
    if (count > 0) {
        this.peerInstancesTransferEmptyOnStartup = false;
    }
    DataCenterInfo.Name selfName = applicationInfoManager.getInfo().getDataCenterInfo().getName();
    boolean isAws = Name.Amazon == selfName;
    if (isAws && serverConfig.shouldPrimeAwsReplicaConnections()) {
        logger.info("Priming AWS connections for all replicas..");
        primeAwsReplicas(applicationInfoManager);
    }
    logger.info("Changing status to UP");
    // 注释:修改 Eureka Server 为上电状态,就是说设置 Eureka Server 已经处于活跃状态了,那就是意味着 EurekaServer 基本上说可以正常使用了;
    applicationInfoManager.setInstanceStatus(InstanceStatus.UP);
    // 注释:定时任务,60000 毫秒,即 1分钟 的间隔时间,Eureke Server定期进行失效节点的清理
    super.postInit();
}

【分析三】:这里主要设置了服务状态,以及开启了定时清理失效节点的定时任务,每分钟扫描一次;

3.11 继续放开断点,来到了日志打印 “main] c.n.e.EurekaDiscoveryClientConfiguration : Updating port to 8761” 的EurekaDiscoveryClientConfiguration 类中 onApplicationEvent 方法。

@EventListener(EmbeddedServletContainerInitializedEvent.class)
public void onApplicationEvent(EmbeddedServletContainerInitializedEvent event) {
    // TODO: take SSL into account when Spring Boot 1.2 is available
    int localPort = event.getEmbeddedServletContainer().getPort();
    if (this.port.get() == 0) {
        log.info("Updating port to " + localPort);
        this.port.compareAndSet(0, localPort);
        start();
    }
}

【分析一】:设置端口,当看到 Updating port to 8761 这样的日志打印出来的话,说明 Eureka Server 整个启动也就差不多Over了。现在回头看看,
发现分析了不少的方法和流程,有种感觉被掏空的感觉了。

3.12 总结 EurekaServer 启动时候大概干了哪些事情?

1、初始化Eureka环境,Eureka上下文;
2、初始化EurekaServer的缓存
3、启动了一些定时任务,比如充值心跳阈值定时任务,清理失效节点定时任务;
4、更新EurekaServer上电状态,更新EurekaServer端口;

虽然我从列举的流程里面大概总结了这么几点,但是还是有些是我没关注到的,如果大家有关注到的,可以和我共同讨论分析分析。

四、EurekaServer 处理服务注册、集群数据复制

4.1 EurekaClient 是如何注册到 EurekaServer 的?

【分析一】:由于我们刚才在 org.springframework.cloud.netflix.eureka.server.InstanceRegistry 的每个方法都打了一个断点,而且现在 
EurekaServer 已经处于 Debug 运行状态,那么我们就随便找一个被 @EnableEurekaClient 的微服务启动试试,要么就拿 springms-provider-user
微服务来试试吧,直接 Run。

【分析二】:猜测,如果如我们分析所想,当 springms-provider-user 启动后,就一定会调用注册register方法,那么就接着往下看,拭目以待;

4.2 InstanceRegistry.register(final InstanceInfo info, final boolean isReplication) 方法进断点了。

【分析一】:由于 InstanceRegistry.register 是我们刚刚打断点的地方,那么我们顺着堆栈信息往上看,原来是 ApplicationResource.addInstance 方法被调用了,那么我们就看看 addInstance 这个方法,并在 addInstance 这里打上断点;接着我们重新杀死 springms-provider-user 服务,然后再重启 springms-provider-user 服务;

4.2 断点再次来到了 ApplicationResource 类,这个类呢,主要是处理接收 Http 的服务请求。

@POST
@Consumes({"application/json", "application/xml"})
public Response addInstance(InstanceInfo info,
                            @HeaderParam(PeerEurekaNode.HEADER_REPLICATION) String isReplication) {
    logger.debug("Registering instance {} (replication={})", info.getId(), isReplication);
    // validate that the instanceinfo contains all the necessary required fields
    if (isBlank(info.getId())) {
        return Response.status(400).entity("Missing instanceId").build();
    } else if (isBlank(info.getHostName())) {
        return Response.status(400).entity("Missing hostname").build();
    } else if (isBlank(info.getAppName())) {
        return Response.status(400).entity("Missing appName").build();
    } else if (!appName.equals(info.getAppName())) {
        return Response.status(400).entity("Mismatched appName, expecting " + appName + " but was " + info.getAppName()).build();
    } else if (info.getDataCenterInfo() == null) {
        return Response.status(400).entity("Missing dataCenterInfo").build();
    } else if (info.getDataCenterInfo().getName() == null) {
        return Response.status(400).entity("Missing dataCenterInfo Name").build();
    }

    // handle cases where clients may be registering with bad DataCenterInfo with missing data
    DataCenterInfo dataCenterInfo = info.getDataCenterInfo();
    if (dataCenterInfo instanceof UniqueIdentifier) {
        String dataCenterInfoId = ((UniqueIdentifier) dataCenterInfo).getId();
        if (isBlank(dataCenterInfoId)) {
            boolean experimental = "true".equalsIgnoreCase(serverConfig.getExperimental("registration.validation.dataCenterInfoId"));
            if (experimental) {
                String entity = "DataCenterInfo of type " + dataCenterInfo.getClass() + " must contain a valid id";
                return Response.status(400).entity(entity).build();
            } else if (dataCenterInfo instanceof AmazonInfo) {
                AmazonInfo amazonInfo = (AmazonInfo) dataCenterInfo;
                String effectiveId = amazonInfo.get(AmazonInfo.MetaDataKey.instanceId);
                if (effectiveId == null) {
                    amazonInfo.getMetadata().put(AmazonInfo.MetaDataKey.instanceId.getName(), info.getId());
                }
            } else {
                logger.warn("Registering DataCenterInfo of type {} without an appropriate id", dataCenterInfo.getClass());
            }
        }
    }

    registry.register(info, "true".equals(isReplication));
    return Response.status(204).build();  // 204 to be backwards compatible
}

【分析一】:这里的写法貌似看起来和我们之前 Controller 的 RESTFUL 写法有点不一样,仔细一看,原来是Jersey RESTful 框架,是一个产品级的 
RESTful service 和 client 框架。与Struts类似,它同样可以和hibernate,spring框架整合。

【分析二】:紧接着,我们看到 registry.register(info, "true".equals(isReplication)); 这么一段代码,注册啊,原来EurekaClient客户端启
动后会调用会通过Http(s)请求,直接调到 ApplicationResource.addInstance 方法,那么总算明白了,只要是和注册有关的,都会调用这个方法。

【分析三】:接着我们深入 registry.register(info, "true".equals(isReplication)) 查看;

@Override
public void register(final InstanceInfo info, final boolean isReplication) {
    handleRegistration(info, resolveInstanceLeaseDuration(info), isReplication);
    super.register(info, isReplication);
}

【分析四】:看看 handleRegistration(info, resolveInstanceLeaseDuration(info), isReplication) 方法;

private void handleRegistration(InstanceInfo info, int leaseDuration,
        boolean isReplication) {
    log("register " + info.getAppName() + ", vip " + info.getVIPAddress()
            + ", leaseDuration " + leaseDuration + ", isReplication "
            + isReplication);
    publishEvent(new EurekaInstanceRegisteredEvent(this, info, leaseDuration,
            isReplication));
}

【分析五】:该方法仅仅只是打了一个日志,然后通过 ApplicationContext 发布了一个事件 EurekaInstanceRegisteredEvent 服务注册事件,正如
“步骤3.3之分析三” 所提到的,用户可以给 EurekaInstanceRegisteredEvent 添加监听事件,那么用户就可以在此刻实现自己想要的一些业务逻辑。
然后我们再来看看 super.register(info, isReplication) 方法,该方法是 InstanceRegistry 的父类 PeerAwareInstanceRegistryImpl 的方法。

4.3 进入 PeerAwareInstanceRegistryImpl 类的 register(final InstanceInfo info, final boolean isReplication) 方法;

@Override
public void register(final InstanceInfo info, final boolean isReplication) {
    // 注释:续约时间,默认时间是常量值 90 秒
    int leaseDuration = Lease.DEFAULT_DURATION_IN_SECS;
    // 注释:续约时间,当然也可以从配置文件中取出来,所以说续约时间值也是可以让我们自己自定义配置的
    if (info.getLeaseInfo() != null && info.getLeaseInfo().getDurationInSecs() > 0) {
        leaseDuration = info.getLeaseInfo().getDurationInSecs();
    }
    // 注释:将注册方的信息写入 EurekaServer 的注册表,父类为 AbstractInstanceRegistry
    super.register(info, leaseDuration, isReplication);
    // 注释:EurekaServer 节点之间的数据同步,复制到其他Peer
    replicateToPeers(Action.Register, info.getAppName(), info.getId(), info, null, isReplication);
}

【分析一】:进入 super.register(info, leaseDuration, isReplication) 看看是如何写入 EurekaServer 的注册表的,即进入 AbstractInstanceRegistry.register(InstanceInfo registrant, int leaseDuration, boolean isReplication) 方法。

public void register(InstanceInfo registrant, int leaseDuration, boolean isReplication) {
    try {
        read.lock();
        // 注释:registry 这个变量,就是我们所谓的注册表,注册表是保存在内存中的;
        Map<String, Lease<InstanceInfo>> gMap = registry.get(registrant.getAppName());
        REGISTER.increment(isReplication);
        if (gMap == null) {
            final ConcurrentHashMap<String, Lease<InstanceInfo>> gNewMap = new ConcurrentHashMap<String, Lease<InstanceInfo>>();
            gMap = registry.putIfAbsent(registrant.getAppName(), gNewMap);
            if (gMap == null) {
                gMap = gNewMap;
            }
        }
        Lease<InstanceInfo> existingLease = gMap.get(registrant.getId());
        // Retain the last dirty timestamp without overwriting it, if there is already a lease
        if (existingLease != null && (existingLease.getHolder() != null)) {
            Long existingLastDirtyTimestamp = existingLease.getHolder().getLastDirtyTimestamp();
            Long registrationLastDirtyTimestamp = registrant.getLastDirtyTimestamp();
            logger.debug("Existing lease found (existing={}, provided={}", existingLastDirtyTimestamp, registrationLastDirtyTimestamp);
            if (existingLastDirtyTimestamp > registrationLastDirtyTimestamp) {
                logger.warn("There is an existing lease and the existing lease's dirty timestamp {} is greater" +
                        " than the one that is being registered {}", existingLastDirtyTimestamp, registrationLastDirtyTimestamp);
                logger.warn("Using the existing instanceInfo instead of the new instanceInfo as the registrant");
                registrant = existingLease.getHolder();
            }
        } else {
            // The lease does not exist and hence it is a new registration
            synchronized (lock) {
                if (this.expectedNumberOfRenewsPerMin > 0) {
                    // Since the client wants to cancel it, reduce the threshold
                    // (1
                    // for 30 seconds, 2 for a minute)
                    this.expectedNumberOfRenewsPerMin = this.expectedNumberOfRenewsPerMin + 2;
                    this.numberOfRenewsPerMinThreshold =
                            (int) (this.expectedNumberOfRenewsPerMin * serverConfig.getRenewalPercentThreshold());
                }
            }
            logger.debug("No previous lease information found; it is new registration");
        }
        Lease<InstanceInfo> lease = new Lease<InstanceInfo>(registrant, leaseDuration);
        if (existingLease != null) {
            lease.setServiceUpTimestamp(existingLease.getServiceUpTimestamp());
        }
        gMap.put(registrant.getId(), lease);
        synchronized (recentRegisteredQueue) {
            recentRegisteredQueue.add(new Pair<Long, String>(
                    System.currentTimeMillis(),
                    registrant.getAppName() + "(" + registrant.getId() + ")"));
        }
        // This is where the initial state transfer of overridden status happens
        if (!InstanceStatus.UNKNOWN.equals(registrant.getOverriddenStatus())) {
            logger.debug("Found overridden status {} for instance {}. Checking to see if needs to be add to the "
                            + "overrides", registrant.getOverriddenStatus(), registrant.getId());
            if (!overriddenInstanceStatusMap.containsKey(registrant.getId())) {
                logger.info("Not found overridden id {} and hence adding it", registrant.getId());
                overriddenInstanceStatusMap.put(registrant.getId(), registrant.getOverriddenStatus());
            }
        }
        InstanceStatus overriddenStatusFromMap = overriddenInstanceStatusMap.get(registrant.getId());
        if (overriddenStatusFromMap != null) {
            logger.info("Storing overridden status {} from map", overriddenStatusFromMap);
            registrant.setOverriddenStatus(overriddenStatusFromMap);
        }

        // Set the status based on the overridden status rules
        InstanceStatus overriddenInstanceStatus = getOverriddenInstanceStatus(registrant, existingLease, isReplication);
        registrant.setStatusWithoutDirty(overriddenInstanceStatus);

        // If the lease is registered with UP status, set lease service up timestamp
        if (InstanceStatus.UP.equals(registrant.getStatus())) {
            lease.serviceUp();
        }
        registrant.setActionType(ActionType.ADDED);
        recentlyChangedQueue.add(new RecentlyChangedItem(lease));
        registrant.setLastUpdatedTimestamp();
        invalidateCache(registrant.getAppName(), registrant.getVIPAddress(), registrant.getSecureVipAddress());
        logger.info("Registered instance {}/{} with status {} (replication={})",
                registrant.getAppName(), registrant.getId(), registrant.getStatus(), isReplication);
    } finally {
        read.unlock();
    }
}

【分析二】:发现这个方法有点长,大致阅读,主要更新了注册表的时间之外,还更新了缓存等其它东西,大家有兴趣的可以深究阅读该方法;

4.4 跳出来我们接着看上面的 replicateToPeers(Action.Register, info.getAppName(), info.getId(), info, null, isReplication) 的这个方法。

private void replicateToPeers(Action action, String appName, String id,
                              InstanceInfo info /* optional */,
                              InstanceStatus newStatus /* optional */, boolean isReplication) {
    Stopwatch tracer = action.getTimer().start();
    try {
        if (isReplication) {
            numberOfReplicationsLastMin.increment();
        }
        // If it is a replication already, do not replicate again as this will create a poison replication
        // 注释:如果已经复制过,就不再复制  
        if (peerEurekaNodes == Collections.EMPTY_LIST || isReplication) {
            return;
        }

        // 遍历Eureka Server集群中的所有节点,进行复制操作 
        for (final PeerEurekaNode node : peerEurekaNodes.getPeerEurekaNodes()) {
            // If the url represents this host, do not replicate to yourself.
            if (peerEurekaNodes.isThisMyUrl(node.getServiceUrl())) {
                continue;
            }
            // 没有复制过,遍历Eureka Server集群中的node节点,依次操作,包括取消、注册、心跳、状态更新等。
            replicateInstanceActionsToPeers(action, appName, id, info, newStatus, node);
        }
    } finally {
        tracer.stop();
    }
}

【分析一】:走到这里,我不难理解,每当有注册请求,首先更新 EurekaServer 的注册表,然后再将信息同步到其它EurekaServer的节点上去;

【分析二】:接下来我们看看 node 节点是如何进行复制操作的,进入 replicateInstanceActionsToPeers 方法。

private void replicateInstanceActionsToPeers(Action action, String appName,
                                             String id, InstanceInfo info, InstanceStatus newStatus,
                                             PeerEurekaNode node) {
    try {
        InstanceInfo infoFromRegistry = null;
        CurrentRequestVersion.set(Version.V2);
        switch (action) {
            case Cancel:
                node.cancel(appName, id);
                break;
            case Heartbeat:
                InstanceStatus overriddenStatus = overriddenInstanceStatusMap.get(id);
                infoFromRegistry = getInstanceByAppAndId(appName, id, false);
                node.heartbeat(appName, id, infoFromRegistry, overriddenStatus, false);
                break;
            case Register:
                node.register(info);
                break;
            case StatusUpdate:
                infoFromRegistry = getInstanceByAppAndId(appName, id, false);
                node.statusUpdate(appName, id, newStatus, infoFromRegistry);
                break;
            case DeleteStatusOverride:
                infoFromRegistry = getInstanceByAppAndId(appName, id, false);
                node.deleteStatusOverride(appName, id, infoFromRegistry);
                break;
        }
    } catch (Throwable t) {
        logger.error("Cannot replicate information to {} for action {}", node.getServiceUrl(), action.name(), t);
    }
}

【分析三】:节点之间的复制状态操作,都在这里体现的淋漓尽致,那么我们就拿 Register 类型 node.register(info) 来看,我们来看看 node 究竟是
如何做到同步信息的,进入 node.register(info) 方法看看;

4.5 进入 PeerEurekaNode.register(final InstanceInfo info) 方法,一窥究竟如何同步数据。

public void register(final InstanceInfo info) throws Exception {
    // 注释:任务过期时间给任务分发器处理,默认时间偏移当前时间 30秒
    long expiryTime = System.currentTimeMillis() + getLeaseRenewalOf(info);
    batchingDispatcher.process(
            taskId("register", info),
            new InstanceReplicationTask(targetHost, Action.Register, info, null, true) {
                public EurekaHttpResponse<Void> execute() {
                    return replicationClient.register(info);
                }
            },
            expiryTime
    );
}

【分析一】:这里涉及到了 Eureka 的任务批处理,通常情况下Peer之间的同步需要调用多次,如果EurekaServer一多的话,那么将会有很多http请求,所
以自然而然的孕育出了任务批处理,但是也在一定程度上导致了注册和下线的一些延迟,突出优势的同时也势必会造成一些劣势,但是这些延迟情况还是能符合
常理在容忍范围之内的。

【分析二】:在 expiryTime 超时时间之内,批次处理要做的事情就是合并任务为一个List,然后发送请求的时候,将这个批次List直接打包发送请求出去,这样的话,在这个批次的List里面,可能包含取消、注册、心跳、状态等一系列状态的集合List。

【分析三】:我们再接着看源码,batchingDispatcher.process 这么一调用,然后我们就直接看这个 TaskDispatchers.createBatchingTaskDispatcher 方法。

public static <ID, T> TaskDispatcher<ID, T> createBatchingTaskDispatcher(String id,
                                                                             int maxBufferSize,
                                                                             int workloadSize,
                                                                             int workerCount,
                                                                             long maxBatchingDelay,
                                                                             long congestionRetryDelayMs,
                                                                             long networkFailureRetryMs,
                                                                             TaskProcessor<T> taskProcessor) {
        final AcceptorExecutor<ID, T> acceptorExecutor = new AcceptorExecutor<>(
                id, maxBufferSize, workloadSize, maxBatchingDelay, congestionRetryDelayMs, networkFailureRetryMs
        );
        final TaskExecutors<ID, T> taskExecutor = TaskExecutors.batchExecutors(id, workerCount, taskProcessor, acceptorExecutor);
        return new TaskDispatcher<ID, T>() {
            @Override
            public void process(ID id, T task, long expiryTime) {
                acceptorExecutor.process(id, task, expiryTime);
            }

            @Override
            public void shutdown() {
                acceptorExecutor.shutdown();
                taskExecutor.shutdown();
            }
        };
    }

【分析四】:这里的 process 方法会将任务添加到队列中,有入队列自然有出队列,具体怎么取任务,我就不一一给大家讲解了,我就讲讲最后是怎么触发任务的。进入 final TaskExecutors<ID, T> taskExecutor = TaskExecutors.batchExecutors(id, workerCount, taskProcessor, acceptorExecutor) 这句代码的 TaskExecutors.batchExecutors 方法。

static <ID, T> TaskExecutors<ID, T> batchExecutors(final String name,
                                                   int workerCount,
                                                   final TaskProcessor<T> processor,
                                                   final AcceptorExecutor<ID, T> acceptorExecutor) {
    final AtomicBoolean isShutdown = new AtomicBoolean();
    final TaskExecutorMetrics metrics = new TaskExecutorMetrics(name);
    return new TaskExecutors<>(new WorkerRunnableFactory<ID, T>() {
        @Override
        public WorkerRunnable<ID, T> create(int idx) {
            return new BatchWorkerRunnable<>("TaskBatchingWorker-" +name + '-' + idx, isShutdown, metrics, processor, acceptorExecutor);
        }
    }, workerCount, isShutdown);
}

【分析五】:我们发现 TaskExecutors 类中的 batchExecutors 这个静态方法,有个 BatchWorkerRunnable 返回的实现类,因此我们再次进入 BatchWorkerRunnable 类看看究竟,而且既然是 Runnable,那么势必会有 run 方法。

@Override
public void run() {
    try {
        while (!isShutdown.get()) {
            // 注释:获取信号量释放 batchWorkRequests.release(),返回任务集合列表
            List<TaskHolder<ID, T>> holders = getWork();
            metrics.registerExpiryTimes(holders);

            List<T> tasks = getTasksOf(holders);
            // 注释:将批量任务打包请求Peer节点
            ProcessingResult result = processor.process(tasks);
            switch (result) {
                case Success:
                    break;
                case Congestion:
                case TransientError:
                    taskDispatcher.reprocess(holders, result);
                    break;
                case PermanentError:
                    logger.warn("Discarding {} tasks of {} due to permanent error", holders.size(), workerName);
            }
            metrics.registerTaskResult(result, tasks.size());
        }
    } catch (InterruptedException e) {
        // Ignore
    } catch (Throwable e) {
        // Safe-guard, so we never exit this loop in an uncontrolled way.
        logger.warn("Discovery WorkerThread error", e);
    }
}

【分析六】:这就是我们 BatchWorkerRunnable 类的 run 方法,这里面首先要获取信号量释放,才能获得任务集合,一旦获取到了任务集合的话,那么就直接调用 processor.process(tasks) 方法请求 Peer 节点同步数据,接下来我们看看 ReplicationTaskProcessor.process 方法;

@Override
public ProcessingResult process(List<ReplicationTask> tasks) {
    ReplicationList list = createReplicationListOf(tasks);
    try {
        // 注释:这里通过 JerseyReplicationClient 客户端对象直接发送list请求数据
        EurekaHttpResponse<ReplicationListResponse> response = replicationClient.submitBatchUpdates(list);
        int statusCode = response.getStatusCode();
        if (!isSuccess(statusCode)) {
            if (statusCode == 503) {
                logger.warn("Server busy (503) HTTP status code received from the peer {}; rescheduling tasks after delay", peerId);
                return ProcessingResult.Congestion;
            } else {
                // Unexpected error returned from the server. This should ideally never happen.
                logger.error("Batch update failure with HTTP status code {}; discarding {} replication tasks", statusCode, tasks.size());
                return ProcessingResult.PermanentError;
            }
        } else {
            handleBatchResponse(tasks, response.getEntity().getResponseList());
        }
    } catch (Throwable e) {
        if (isNetworkConnectException(e)) {
            logNetworkErrorSample(null, e);
            return ProcessingResult.TransientError;
        } else {
            logger.error("Not re-trying this exception because it does not seem to be a network exception", e);
            return ProcessingResult.PermanentError;
        }
    }
    return ProcessingResult.Success;
}

【分析七】:感觉快要见到真相了,所以我们迫不及待的进入 JerseyReplicationClient.submitBatchUpdates(ReplicationList replicationList) 方法一窥究竟。

@Override
public EurekaHttpResponse<ReplicationListResponse> submitBatchUpdates(ReplicationList replicationList) {
    ClientResponse response = null;
    try {
        response = jerseyApacheClient.resource(serviceUrl)
                // 注释:这才是重点,请求目的相对路径,peerreplication/batch/
                .path(PeerEurekaNode.BATCH_URL_PATH)
                .accept(MediaType.APPLICATION_JSON_TYPE)
                .type(MediaType.APPLICATION_JSON_TYPE)
                .post(ClientResponse.class, replicationList);
        if (!isSuccess(response.getStatus())) {
            return anEurekaHttpResponse(response.getStatus(), ReplicationListResponse.class).build();
        }
        ReplicationListResponse batchResponse = response.getEntity(ReplicationListResponse.class);
        return anEurekaHttpResponse(response.getStatus(), batchResponse).type(MediaType.APPLICATION_JSON_TYPE).build();
    } finally {
        if (response != null) {
            response.close();
        }
    }
}

【分析八】:看到了相对路径地址,我们搜索下"batch"这样的字符串看看有没有对应的接收方法或者被@Path注解进入的;在 eureka-core-1.4.12.jar 这个包下面,果然搜到到了 @Path("batch") 这样的字样,直接进入,发现这是 PeerReplicationResource 类的方法 batchReplication,我们进入这方法看看。

@Path("batch")
@POST
public Response batchReplication(ReplicationList replicationList) {
    try {
        ReplicationListResponse batchResponse = new ReplicationListResponse();
        // 注释:这里将收到的任务列表,依次循环解析处理,主要核心方法在 dispatch 方法中。
        for (ReplicationInstance instanceInfo : replicationList.getReplicationList()) {
            try {
                batchResponse.addResponse(dispatch(instanceInfo));
            } catch (Exception e) {
                batchResponse.addResponse(new ReplicationInstanceResponse(Status.INTERNAL_SERVER_ERROR.getStatusCode(), null));
                logger.error(instanceInfo.getAction() + " request processing failed for batch item "
                        + instanceInfo.getAppName() + '/' + instanceInfo.getId(), e);
            }
        }
        return Response.ok(batchResponse).build();
    } catch (Throwable e) {
        logger.error("Cannot execute batch Request", e);
        return Response.status(Status.INTERNAL_SERVER_ERROR).build();
    }
}

【分析九】:看到了循环一次遍历任务进行处理,不知不觉觉得心花怒放,胜利的重点马上就要到来了,我们进入 PeerReplicationResource.dispatch 方法看看。

private ReplicationInstanceResponse dispatch(ReplicationInstance instanceInfo) {
    ApplicationResource applicationResource = createApplicationResource(instanceInfo);
    InstanceResource resource = createInstanceResource(instanceInfo, applicationResource);

    String lastDirtyTimestamp = toString(instanceInfo.getLastDirtyTimestamp());
    String overriddenStatus = toString(instanceInfo.getOverriddenStatus());
    String instanceStatus = toString(instanceInfo.getStatus());

    Builder singleResponseBuilder = new Builder();
    switch (instanceInfo.getAction()) {
        case Register:
            singleResponseBuilder = handleRegister(instanceInfo, applicationResource);
            break;
        case Heartbeat:
            singleResponseBuilder = handleHeartbeat(resource, lastDirtyTimestamp, overriddenStatus, instanceStatus);
            break;
        case Cancel:
            singleResponseBuilder = handleCancel(resource);
            break;
        case StatusUpdate:
            singleResponseBuilder = handleStatusUpdate(instanceInfo, resource);
            break;
        case DeleteStatusOverride:
            singleResponseBuilder = handleDeleteStatusOverride(instanceInfo, resource);
            break;
    }
    return singleResponseBuilder.build();
}

【分析十】:随便抓一个类型,那我们也拿 Register 类型来看,进入 PeerReplicationResource.handleRegister 看看。

private static Builder handleRegister(ReplicationInstance instanceInfo, ApplicationResource applicationResource) {
    // 注释:private static final String REPLICATION = "true"; 定义的一个常量值,而且还是回调 ApplicationResource.addInstance 方法
    applicationResource.addInstance(instanceInfo.getInstanceInfo(), REPLICATION);
    return new Builder().setStatusCode(Status.OK.getStatusCode());
}

【分析十一】:Peer节点的同步旅程终于结束了,最终又回调到了 ApplicationResource.addInstance 这个方法,这个方法在最终是EurekaClient启动后注册调用的方法,然而Peer节点的信息同步也调用了这个方法,仅仅只是通过一个变量 isReplication 为true还是false来判断是否是节点复制。剩下的ApplicationResource.addInstance流程前面已经提到过了,相信大家已经明白了注册的流程是如何扭转的,包括批量任务是如何处理EurekaServer节点之间的信息同步的了。

五、EurekaClient 启动流程分析

详见 SpringCloud(第 050 篇)Netflix Eureka 源码深入剖析(下)

六、下载地址

https://gitee.com/ylimhhmily/SpringCloudTutorial.git

SpringCloudTutorial交流QQ群: 235322432

SpringCloudTutorial交流微信群: 微信沟通群二维码图片链接

欢迎关注,您的肯定是对我最大的支持!!!


HMILYYLIMH
325 声望314 粉丝