javaWeb学习一些收藏微服务

微服务架构 | *2.5 Nacos 长轮询定时机制的源码分析

2022-01-24  本文已影响0人  多氯环己烷

前言

参考资料：
《Spring Microservices in Action》
《Spring Cloud Alibaba 微服务原理与实战》
《B站 尚硅谷 SpringCloud 框架开发教程 周阳》

为方便理解与表达，这里把 Nacos 控制台和 Nacos 注册中心称为 Nacos 服务器（就是 web 界面那个），我们编写的业务服务称为 Nacso 客户端；

由于篇幅有限，这里将源码分析分为上下两篇，其中上篇讲获取配置与事件订阅机制，下篇讲长轮询定时机制；在《微服务架构 | 2.2 Alibaba Nacos 的统一配置管理》中提到一张 Nacos 动态监听的长轮询机制原理图，本篇将围绕这张图剖析长轮询定时机制的原理；

Nacos 动态监听的长轮询机制.png

上篇《微服务架构 | *2.4 Nacos 配置中心的源码分析（获取配置与事件订阅机制）》中的 1.1 提到，ConfigService 是 Nacos 客户端提供的用于访问实现配置中心基本操作的类，我们将从 ConfigService 的实例化开始长轮询定时机制的源码之旅；

1. 客户端的长轮询定时机制

1.1 利用反射机制实例化 NacosConfigService 对象

public static ConfigService createConfigService(Properties properties) throws NacosException {
    //【断点步入】创建 ConfigService
    return ConfigFactory.createConfigService(properties);
}

public static ConfigService createConfigService(Properties properties) throws NacosException {
    try {
        //通过 Class.forName 来加载 NacosConfigService
        Class<?> driverImplClass = Class.forName("com.alibaba.nacos.client.config.NacosConfigService");
        //根据对象的属性创建对象构造器（反射）
        Constructor constructor = driverImplClass.getConstructor(Properties.class);
        //【断点步入 1.2】通过构造器实例化对象（反射）
        ConfigService vendorImpl = (ConfigService)constructor.newInstance(properties);
        return vendorImpl;
    } catch (Throwable var4) {
        throw new NacosException(-400, var4);
    }
}

1.2 NacosConfigService 的构造方法里启动长轮询定时任务

public NacosConfigService(Properties properties) throws NacosException {
    String encodeTmp = properties.getProperty("encode");
    if (StringUtils.isBlank(encodeTmp)) {
        this.encode = "UTF-8";
    } else {
        this.encode = encodeTmp.trim();
    }
    //初始化命名空间
    this.initNamespace(properties);
    //【断点步入 1.2.1】初始化 HttpAgent，用到了装饰器模式，实际工作的类是 ServerHttpAgent
    this.agent = new MetricsHttpAgent(new ServerHttpAgent(properties));
    this.agent.start();
    //【断点步入 1.2.2】ClientWorker 是客户端的一个工作类，agent 作为参数传入 ClientWorker
    this.worker = new ClientWorker(this.agent, this.configFilterChainManager, properties);
}

1.2.1 初始化 HttpAgent

1.2.2 初始化 ClientWorker

public ClientWorker(final HttpAgent agent, ConfigFilterChainManager configFilterChainManager, Properties properties) {
    this.agent = agent;
    this.configFilterChainManager = configFilterChainManager;
    this.init(properties);
    //创建 executor 线程池，只拥有一个核心线程，每隔 10ms 就会执行一次 checkConfiglnfo() 方法，检查配置信息
    this.executor = Executors.newScheduledThreadPool(1, new ThreadFactory() {
        public Thread newThread(Runnable r) {
            Thread t = new Thread(r);
            t.setName("com.alibaba.nacos.client.Worker." + agent.getName());
            t.setDaemon(true);
            return t;
        }
    });
    //创建 executorService 线程池，只完成了初始化，后续会用到，主要用于实现客户端的定时长轮询功能
    this.executorService = Executors.newScheduledThreadPool(Runtime.getRuntime().availableProcessors(), new ThreadFactory() {
        public Thread newThread(Runnable r) {
            Thread t = new Thread(r);
            t.setName("com.alibaba.nacos.client.Worker.longPolling." + agent.getName());
            t.setDaemon(true);
            return t;
        }
    });
    //使用 executor 启动一个每隔 10s 执行一次的定时任务
    this.executor.scheduleWithFixedDelay(new Runnable() {
        public void run() {
            try {
                //【断点步入】检查配置是否发生变化
                ClientWorker.this.checkConfigInfo();
            } catch (Throwable var2) {
                ClientWorker.LOGGER.error("[" + agent.getName() + "] [sub-check] rotate check error", var2);
            }
        }
    }, 1L, 10L, TimeUnit.MILLISECONDS);
}

public void checkConfigInfo() {
    //分任务
    int listenerSize = ((Map)this.cacheMap.get()).size();
    //向上取整为批数
    int longingTaskCount = (int)Math.ceil((double)listenerSize / ParamUtil.getPerTaskConfigSize());
    //如果监听配置集超过 3000，就创建多个 LongPollingRunnable 线程
    if ((double)longingTaskCount > this.currentLongingTaskCount) {
        for(int i = (int)this.currentLongingTaskCount; i < longingTaskCount; ++i) {
            //【点进去】LongPollingRunnable 实际上是一个线程
            this.executorService.execute(new ClientWorker.LongPollingRunnable(i));
        }
        this.currentLongingTaskCount = (double)longingTaskCount;
    }
}

1.3 检查配置变更，读取变更配置 LongPollingRunnable.run()

public void run() {
    List<CacheData> cacheDatas = new ArrayList();
    ArrayList inInitializingCacheList = new ArrayList();
    try {
        //遍历 CacheData，检查本地配置
        Iterator var3 = ((Map)ClientWorker.this.cacheMap.get()).values().iterator();
        while(var3.hasNext()) {
            CacheData cacheData = (CacheData)var3.next();
            if (cacheData.getTaskId() == this.taskId) {
                cacheDatas.add(cacheData);
                try {
                    //检查本地配置
                    ClientWorker.this.checkLocalConfig(cacheData);
                    if (cacheData.isUseLocalConfigInfo()) {
                        cacheData.checkListenerMd5();
                    }
                } catch (Exception var13) {
                    ClientWorker.LOGGER.error("get local config info error", var13);
                }
            }
        }
        //【断点步入 1.3.1】通过长轮询请求检查服务端对应的配置是否发生变更
        List<String> changedGroupKeys = ClientWorker.this.checkUpdateDataIds(cacheDatas, inInitializingCacheList);
        //遍历存在变更的 groupKey，重新加载最新数据
        Iterator var16 = changedGroupKeys.iterator();
        while(var16.hasNext()) {
            String groupKey = (String)var16.next();
            String[] key = GroupKey.parseKey(groupKey);
            String dataId = key[0];
            String group = key[1];
            String tenant = null;
            if (key.length == 3) {
                tenant = key[2];
            }
            try {
                //【断点步入 1.3.2】读取变更配置，这里的 dataId、group 和 tenant 是【1.3.1】里获取的
                String content = ClientWorker.this.getServerConfig(dataId, group, tenant, 3000L);
                CacheData cache = (CacheData)((Map)ClientWorker.this.cacheMap.get()).get(GroupKey.getKeyTenant(dataId, group, tenant));
                cache.setContent(content);
                ClientWorker.LOGGER.info("[{}] [data-received] dataId={}, group={}, tenant={}, md5={}, content={}", new Object[]{ClientWorker.this.agent.getName(), dataId, group, tenant, cache.getMd5(), ContentUtils.truncateContent(content)});
            } catch (NacosException var12) {
                String message = String.format("[%s] [get-update] get changed config exception. dataId=%s, group=%s, tenant=%s", ClientWorker.this.agent.getName(), dataId, group, tenant);
                ClientWorker.LOGGER.error(message, var12);
            }
        }
        //触发事件通知
        var16 = cacheDatas.iterator();
        while(true) {
            CacheData cacheDatax;
            do {
                if (!var16.hasNext()) {
                    inInitializingCacheList.clear();
                    //继续定时执行当前线程
                    ClientWorker.this.executorService.execute(this);
                    return;
                }
                cacheDatax = (CacheData)var16.next();
            } while(cacheDatax.isInitializing() && !inInitializingCacheList.contains(GroupKey.getKeyTenant(cacheDatax.dataId, cacheDatax.group, cacheDatax.tenant)));
            cacheDatax.checkListenerMd5();
            cacheDatax.setInitializing(false);
        }
    } catch (Throwable var14) {
        ClientWorker.LOGGER.error("longPolling error : ", var14);
        ClientWorker.this.executorService.schedule(this, (long)ClientWorker.this.taskPenaltyTime, TimeUnit.MILLISECONDS);
    }
}

1.3.1 检查配置变更 ClientWorker.checkUpdateDataIds()

List<String> checkUpdateConfigStr(String probeUpdateString, boolean isInitializingCacheList) throws IOException {
    List<String> params = Arrays.asList("Listening-Configs", probeUpdateString);
    List<String> headers = new ArrayList(2);
    headers.add("Long-Pulling-Timeout");
    headers.add("" + this.timeout);
    if (isInitializingCacheList) {
        headers.add("Long-Pulling-Timeout-No-Hangup");
        headers.add("true");
    }
    if (StringUtils.isBlank(probeUpdateString)) {
        return Collections.emptyList();
    } else {
        try {
            //调用 /v1/cs/configs/listener 接口实现长轮询请求，返回的 HttpResult 里包含存在数据变更的 Data ID、Group、Tenant
            HttpResult result = this.agent.httpPost("/v1/cs/configs/listener", headers, params, this.agent.getEncode(), this.timeout);
            if (200 == result.code) {
                this.setHealthServer(true);
                //
                return this.parseUpdateDataIdResponse(result.content);
            }
            this.setHealthServer(false);
            LOGGER.error("[{}] [check-update] get changed dataId error, code: {}", this.agent.getName(), result.code);
        } catch (IOException var6) {
            this.setHealthServer(false);
            LOGGER.error("[" + this.agent.getName() + "] [check-update] get changed dataId exception", var6);
            throw var6;
        }
        return Collections.emptyList();
    }
}

1.3.2 读取变更配置 ClientWorker.getServerConfig()

public String getServerConfig(String dataId, String group, String tenant, long readTimeout) throws NacosException {
    if (StringUtils.isBlank(group)) {
        group = "DEFAULT_GROUP";
    }
    HttpResult result = null;
    try {
        List<String> params = null;
        if (StringUtils.isBlank(tenant)) {
            params = Arrays.asList("dataId", dataId, "group", group);
        } else {
            params = Arrays.asList("dataId", dataId, "group", group, "tenant", tenant);
        }
        //获取变更配置的接口调用
        result = this.agent.httpGet("/v1/cs/configs", (List)null, params, this.agent.getEncode(), readTimeout);
    } catch (IOException var9) {
        String message = String.format("[%s] [sub-server] get server config exception, dataId=%s, group=%s, tenant=%s", this.agent.getName(), dataId, group, tenant);
        LOGGER.error(message, var9);
        throw new NacosException(500, var9);
    }
    switch(result.code) {
    //获取变更的配置成功，添加进缓存里
    case 200:
        LocalConfigInfoProcessor.saveSnapshot(this.agent.getName(), dataId, group, tenant, result.content);
        //result.content 就是我们变更后的配置信息
        return result.content;
    case 403:
        LOGGER.error("[{}] [sub-server-error] no right, dataId={}, group={}, tenant={}", new Object[]{this.agent.getName(), dataId, group, tenant});
        throw new NacosException(result.code, result.content);
    case 404:
        LocalConfigInfoProcessor.saveSnapshot(this.agent.getName(), dataId, group, tenant, (String)null);
        return null;
    case 409:
        LOGGER.error("[{}] [sub-server-error] get server config being modified concurrently, dataId={}, group={}, tenant={}", new Object[]{this.agent.getName(), dataId, group, tenant});
        throw new NacosException(409, "data being modified, dataId=" + dataId + ",group=" + group + ",tenant=" + tenant);
    default:
        LOGGER.error("[{}] [sub-server-error]  dataId={}, group={}, tenant={}, code={}", new Object[]{this.agent.getName(), dataId, group, tenant, result.code});
        throw new NacosException(result.code, "http error, code=" + result.code + ",dataId=" + dataId + ",group=" + group + ",tenant=" + tenant);
    }
}
result.content.png

2. 服务端的长轮询定时机制

2.1 服务器接收请求 ConfigController.listener()

@PostMapping("/listener")
@Secured(action = ActionTypes.READ, parser = ConfigResourceParser.class)
public void listener(HttpServletRequest request, HttpServletResponse response) throws ServletException, IOException {
    request.setAttribute("org.apache.catalina.ASYNC_SUPPORTED", true);
    String probeModify = request.getParameter("Listening-Configs");
    if (StringUtils.isBlank(probeModify)) {
        throw new IllegalArgumentException("invalid probeModify");
    }
    
    probeModify = URLDecoder.decode(probeModify, Constants.ENCODE);
    
    Map<String, String> clientMd5Map;
    try {
        //计算 MD5 值
        clientMd5Map = MD5Util.getClientMd5Map(probeModify);
    } catch (Throwable e) {
        throw new IllegalArgumentException("invalid probeModify");
    }
    
    //【断点步入 2.2】执行长轮询请求
    inner.doPollingConfig(request, response, clientMd5Map, probeModify.length());
}

2.2 执行长轮询请求 ConfigServletInner.doPollingConfig()

public String doPollingConfig(HttpServletRequest request, HttpServletResponse response, Map<String, String> clientMd5Map, int probeRequestSize) throws IOException {
    //长轮询
    if (LongPollingService.isSupportLongPolling(request)) {
        //【断点步入】长轮询逻辑
        longPollingService.addLongPollingClient(request, response, clientMd5Map, probeRequestSize);
        return HttpServletResponse.SC_OK + "";
    }
    //兼容短轮询逻辑
    List<String> changedGroups = MD5Util.compareMd5(request, response, clientMd5Map);
    //兼容短轮询 result
    String oldResult = MD5Util.compareMd5OldResult(changedGroups);
    String newResult = MD5Util.compareMd5ResultString(changedGroups);
    String version = request.getHeader(Constants.CLIENT_VERSION_HEADER);
    if (version == null) {
        version = "2.0.0";
    }
    int versionNum = Protocol.getVersionNumber(version);
    
    //在 2.0.4 版本之前，返回值放入表头
    if (versionNum < START_LONG_POLLING_VERSION_NUM) {
        response.addHeader(Constants.PROBE_MODIFY_RESPONSE, oldResult);
        response.addHeader(Constants.PROBE_MODIFY_RESPONSE_NEW, newResult);
    } else {
        request.setAttribute("content", newResult);
    }
    Loggers.AUTH.info("new content:" + newResult);
    //禁用缓存
    response.setHeader("Pragma", "no-cache");
    response.setDateHeader("Expires", 0);
    response.setHeader("Cache-Control", "no-cache,no-store");
    response.setStatus(HttpServletResponse.SC_OK);
    return HttpServletResponse.SC_OK + "";
}

public void addLongPollingClient(HttpServletRequest req, HttpServletResponse rsp, Map<String, String> clientMd5Map, int probeRequestSize) {
    //获取客户端设置的请求超时时间
    String str = req.getHeader(LongPollingService.LONG_POLLING_HEADER);
    String noHangUpFlag = req.getHeader(LongPollingService.LONG_POLLING_NO_HANG_UP_HEADER);
    String appName = req.getHeader(RequestUtil.CLIENT_APPNAME_HEADER);
    String tag = req.getHeader("Vipserver-Tag");
    int delayTime = SwitchService.getSwitchInteger(SwitchService.FIXED_DELAY_TIME, 500);
    
    //为 LoadBalance 添加延迟时间，并提前 500ms 返回响应，避免客户端超时（即超时时间减 500ms 后赋值给 timeout 变量）
    long timeout = Math.max(10000, Long.parseLong(str) - delayTime);
    //判断是否为固定轮询，是则 30s 后执行；否则 29.5s 后执行
    if (isFixedPolling()) {
        timeout = Math.max(10000, getFixedPollingInterval());
        // Do nothing but set fix polling timeout.
    } else {
        long start = System.currentTimeMillis();
        //和服务端的数据进行 MD5 对比，发生变化则直接返回
        List<String> changedGroups = MD5Util.compareMd5(req, rsp, clientMd5Map);
        if (changedGroups.size() > 0) {
            generateResponse(req, rsp, changedGroups);
            LogUtil.CLIENT_LOG.info("{}|{}|{}|{}|{}|{}|{}", System.currentTimeMillis() - start, "instant", RequestUtil.getRemoteIp(req), "polling", clientMd5Map.size(), probeRequestSize, changedGroups.size());
            return;
        } else if (noHangUpFlag != null && noHangUpFlag.equalsIgnoreCase(TRUE_STR)) {
            LogUtil.CLIENT_LOG.info("{}|{}|{}|{}|{}|{}|{}", System.currentTimeMillis() - start, "nohangup", RequestUtil.getRemoteIp(req), "polling", clientMd5Map.size(), probeRequestSize, changedGroups.size());
            return;
        }
    }
    String ip = RequestUtil.getRemoteIp(req);
    
    //一定要由 HTTP 线程调用，否则离开容器会立即发送响应
    final AsyncContext asyncContext = req.startAsync();
    
    //AsyncContext.setTimeout()的超时时间不准，所以自己控制
    asyncContext.setTimeout(0L);
    
    //【点进去】调用 scheduler.execute 执行 ClientLongPolling 线程
    ConfigExecutor.executeLongPolling(new ClientLongPolling(asyncContext, clientMd5Map, ip, probeRequestSize, timeout, appName, tag));
}

2.3 创建线程执行定时任务 ClientLongPolling.run()

@Override
public void run() {
    //启动定时任务，延时时间为 29.5s
    asyncTimeoutFuture = ConfigExecutor.scheduleLongPolling(new Runnable() {
        @Override
        public void run() {
            try {
                //将 ClientLongPolling 实例本身添加到 allSubs 队列中，它主要维护一个长轮询的订阅关系
                getRetainIps().put(ClientLongPolling.this.ip, System.currentTimeMillis());
                //定时任务执行后，先把 ClientLongPolling 实例本身从 allSubs 队列中移除
                allSubs.remove(ClientLongPolling.this);
                //判断是否为固定轮询
                if (isFixedPolling()) {
                    LogUtil.CLIENT_LOG.info("{}|{}|{}|{}|{}|{}", (System.currentTimeMillis() - createTime), "fix", RequestUtil.getRemoteIp((HttpServletRequest) asyncContext.getRequest()),"polling", clientMd5Map.size(), probeRequestSize);
                    //比较数据的 MD5 值，判断是否发生变更
                    List<String> changedGroups = MD5Util.compareMd5((HttpServletRequest) asyncContext.getRequest(), (HttpServletResponse) asyncContext.getResponse(), clientMd5Map);
                    if (changedGroups.size() > 0) {
                        //并将变更的结果通过response返回给客户端
                        sendResponse(changedGroups);
                    } else {
                        sendResponse(null);
                    }
                } else {
                    LogUtil.CLIENT_LOG.info("{}|{}|{}|{}|{}|{}", (System.currentTimeMillis() - createTime), "timeout", RequestUtil.getRemoteIp((HttpServletRequest) asyncContext.getRequest()),"polling", clientMd5Map.size(), probeRequestSize);
                    sendResponse(null);
                }
            } catch (Throwable t) {
                LogUtil.DEFAULT_LOG.error("long polling error:" + t.getMessage(), t.getCause());
            }
        }
    }, timeoutTime, TimeUnit.MILLISECONDS);
    allSubs.add(this);
}

2.4 监听配置变更事件

2.4.1 监听 LocalDataChangeEvent 事件的实现

@Service
public class LongPollingService extends AbstractEventListener {

   //省略其他代码

    @Override
    public void onEvent(Event event) {
        if (isFixedPolling()) {
            // Ignore.
        } else {
            if (event instanceof LocalDataChangeEvent) {
                LocalDataChangeEvent evt = (LocalDataChangeEvent) event;
                //【点进去 2.4.2】通过线程池执行 DataChangeTask 任务
                scheduler.execute(new DataChangeTask(evt.groupKey, evt.isBeta, evt.betaIps));
            }
        }
    }
}    

NotifyCenter.registerSubscriber(new Subscriber() {
    
    @Override
    public void onEvent(Event event) {
        if (isFixedPolling()) {
            // Ignore.
        } else {
            if (event instanceof LocalDataChangeEvent) {
                LocalDataChangeEvent evt = (LocalDataChangeEvent) event;
                //【点进去 2.4.2】通过线程池执行 DataChangeTask 任务
                ConfigExecutor.executeLongPolling(new DataChangeTask(evt.groupKey, evt.isBeta, evt.betaIps));
            }
        }
    }
    @Override
    public Class<? extends Event> subscribeType() {
        return LocalDataChangeEvent.class;
    }
});

2.4.2 监听事件后的处理逻辑 DataChangeTask.run()

@Override
public void run() {
    try {
        ConfigCacheService.getContentBetaMd5(groupKey);
        //遍历 allSubs 中的客户端长轮询请求
        for (Iterator<ClientLongPolling> iter = allSubs.iterator(); iter.hasNext(); ) {
            ClientLongPolling clientSub = iter.next();
            //比较每一个客户端长轮询请求携带的groupKey，如果服务端变更的配置和客户端请求关注的配置一致，则直接返回
            if (clientSub.clientMd5Map.containsKey(groupKey)) {
                //如果 beta 发布且不在 beta 列表，则直接跳过
                if (isBeta && !CollectionUtils.contains(betaIps, clientSub.ip)) {
                    continue;
                }
                //如果 tag 发布且不在 tag 列表，则直接跳过
                if (StringUtils.isNotBlank(tag) && !tag.equals(clientSub.tag)) {
                    continue;
                }
                getRetainIps().put(clientSub.ip, System.currentTimeMillis());
                iter.remove(); //删除订阅关系
                LogUtil.CLIENT_LOG.info("{}|{}|{}|{}|{}|{}|{}", (System.currentTimeMillis() - changeTime), "in-advance", RequestUtil.getRemoteIp((HttpServletRequest) clientSub.asyncContext.getRequest()), "polling", clientSub.clientMd5Map.size(), clientSub.probeRequestSize, groupKey);
                //发送响应
                clientSub.sendResponse(Arrays.asList(groupKey));
            }
        }
    } catch (Throwable t) {
        LogUtil.DEFAULT_LOG.error("data change error: {}", ExceptionUtil.getStackTrace(t));
    }
}

3. 源码结构图小结

3.1 客户端的长轮询定时机制

3.2 服务端的长轮询定时机制

3.3 Nacos 服务器配置变更的事件监听

最后

\color{blue}{\rm\small{新人制作，如有错误，欢迎指出，感激不尽！}}

\color{blue}{\rm\small{欢迎关注我，并与我交流！}}

\color{blue}{\rm\small{如需转载，请标注出处！}}

上一篇下一篇

猜你喜欢

热点阅读