broker是消息隊列的核心組建,承載了消息接收,存儲和轉(zhuǎn)發(fā)的職責(zé). 因此, broker需要具備各種基本功能和高階功能.
1.基本功能
- 承載消息堆積的能力
消息到達服務(wù)端如果不經(jīng)過任何處理就到接收者了, broker就失去了它的意義. 為了滿足我們錯峰/流控/最終可達等一系列需求, 把消息存儲下來, 然后選擇時機投遞就顯得是順理成章的了. 因此, broker必須具備強大的消息堆積能力. - 消費關(guān)系解偶
其實就是具備單播(點對點)和廣播(一點對多點)功能.
2.高階功能 - 消息去重
當(dāng)前的消息隊列還無法做到消息完全去重(除非允許消息丟失), 只能盡量減少消息重復(fù)的概率. - 順序消息
消息有序指的是一類消息消費時, 能按照發(fā)送的順序來消費. 例如: 一個訂單產(chǎn)生了3條消息, 分別是訂單創(chuàng)建, 訂單付款, 訂單完成. 消費時, 要按照這個順序消費才能有意義. 但是同時訂單之間是可以并行消費的. RocketMQ可以嚴(yán)格的保證消息有序.
我們從三個方面分析broker, 分別為broker的啟動, 接收消息和消費消息.
broker的啟動
broker的啟動過程是一個比較復(fù)雜的過程, 其中涉及到很多broker模塊自身的初始化, 例如:路由信息設(shè)置, 消息落地, 消息推送, 以及和namesvr模塊通信部分的初始化.
broker啟動可以分為兩個步驟, initialize和start.
initialize
1.當(dāng)broker啟動命令得到外部傳入的初始化參數(shù)以后, 將參數(shù)注入對應(yīng)的config類當(dāng)中, 這些config類包括:
- broker自身的配置:包括根目錄, namesrv地址, broker的IP和名稱, 消息隊列數(shù), 收發(fā)消息線程池數(shù)等參數(shù)
- netty啟動配置:包括netty監(jiān)聽端口, 工作線程數(shù), 異步發(fā)送消息信號量數(shù)量等網(wǎng)絡(luò)配置等參數(shù).
- 存儲層配置:包括存儲跟目錄, CommitLog配置, 持久化策略配置等參數(shù).
這一步驟的具體代碼在BrokerStartup的start方法和createBrokerController方法中, 較為簡單,此處不貼出.
2.當(dāng)配置信息設(shè)置完畢后, broker會將這些參數(shù)傳入borkerController控制器當(dāng)中, 這個控制器會初始加載很多的管理器, 如下:
- topicManager:用于管理broker中存儲的所有topic的配置
- consumerOffsetManager:管理Consumer的消費進度
- subscriptionGroupManager:用來管理訂閱組, 包括訂閱權(quán)限等
- messageStore:用于broker層的消息落地存儲.
這部分代碼在BrokerController.initialize中, 代碼片段如下:
public boolean initialize() throws CloneNotSupportedException {
boolean result = true;
//加載topicConfigManager
result = result && this.topicConfigManager.load();
//加載consumerOffsetManager
result = result && this.consumerOffsetManager.load();
//加載subscriptionGroupManager
result = result && this.subscriptionGroupManager.load();
//加載messageStore
if (result) {
try {
this.messageStore =
new DefaultMessageStore(this.messageStoreConfig, this.brokerStatsManager, this.messageArrivingListener,
this.brokerConfig);
this.brokerStats = new BrokerStats((DefaultMessageStore) this.messageStore);
//load plugin
MessageStorePluginContext context = new MessageStorePluginContext(messageStoreConfig, brokerStatsManager, messageArrivingListener, brokerConfig);
this.messageStore = MessageStoreFactory.build(context, this.messageStore);
} catch (IOException e) {
result = false;
e.printStackTrace();
}
}
result = result && this.messageStore.load();
...
}
下面看一下this.topicConfigManager.load()的代碼:
public boolean load() {
String fileName = null;
try {
//讀取文件路徑
fileName = this.configFilePath();
//讀取文件內(nèi)容
String jsonString = MixAll.file2String(fileName);
if (null == jsonString || jsonString.length() == 0) {
//讀取備份文件
return this.loadBak();
} else {
//將文件內(nèi)容讀入到topicConfigManager.topicConfigTable中
this.decode(jsonString);
PLOG.info("load {} OK", fileName);
return true;
}
} catch (Exception e) {
PLOG.error("load " + fileName + " Failed, and try to load backup file", e);
return this.loadBak();
}
}
consumerOffsetManager和subscriptionGroupManager的加載類似, 三個類都繼承自ConfigManager.
接下去就是messageStore的加載, 用于消息的持久化, 關(guān)于這一部分, 可以參考前文.
3.當(dāng)上述的管理器全部加載完成以后, 控制器將開始進入下一步的初始化:
- 啟動netty服務(wù)端, 包括處理消息的remotingServer和主從同步使用的fastRemotingServer.
- 初始化多個線程池, 包括:sendMessageExecutor(處理發(fā)送消息), pullMessageExecutor(處理拉取消息), adminBrokerExecutor(管理Broker), clientManageExecutor(管理client), consumerManageExecutor(管理消費者)
- 將上述的線程池注冊到netty消息處理器當(dāng)中
- 啟動定時調(diào)度線程來做很多事情, 包括:一天上報一次broker的所有狀態(tài), 10秒持久化一次Consumer消費進度, 60分鐘清理一次無效的topic訂閱信息, 60秒獲取一次namesrv的地址信息
代碼如下:
public boolean initialize() throws CloneNotSupportedException {
......
if (result) {
this.remotingServer = new NettyRemotingServer(this.nettyServerConfig, this.clientHousekeepingService);
NettyServerConfig fastConfig = (NettyServerConfig) this.nettyServerConfig.clone();
fastConfig.setListenPort(nettyServerConfig.getListenPort() - 2);
this.fastRemotingServer = new NettyRemotingServer(fastConfig, this.clientHousekeepingService);
this.sendMessageExecutor = new BrokerFixedThreadPoolExecutor(
this.brokerConfig.getSendMessageThreadPoolNums(),
this.brokerConfig.getSendMessageThreadPoolNums(),
1000 * 60,
TimeUnit.MILLISECONDS,
this.sendThreadPoolQueue,
new ThreadFactoryImpl("SendMessageThread_"));
this.pullMessageExecutor = new BrokerFixedThreadPoolExecutor(
this.brokerConfig.getPullMessageThreadPoolNums(),
this.brokerConfig.getPullMessageThreadPoolNums(),
1000 * 60,
TimeUnit.MILLISECONDS,
this.pullThreadPoolQueue,
new ThreadFactoryImpl("PullMessageThread_"));
this.adminBrokerExecutor =
Executors.newFixedThreadPool(this.brokerConfig.getAdminBrokerThreadPoolNums(), new ThreadFactoryImpl(
"AdminBrokerThread_"));
this.clientManageExecutor = new ThreadPoolExecutor(
this.brokerConfig.getClientManageThreadPoolNums(),
this.brokerConfig.getClientManageThreadPoolNums(),
1000 * 60,
TimeUnit.MILLISECONDS,
this.clientManagerThreadPoolQueue,
new ThreadFactoryImpl("ClientManageThread_"));
this.consumerManageExecutor =
Executors.newFixedThreadPool(this.brokerConfig.getConsumerManageThreadPoolNums(), new ThreadFactoryImpl(
"ConsumerManageThread_"));
this.registerProcessor();
// TODO remove in future
final long initialDelay = UtilAll.computNextMorningTimeMillis() - System.currentTimeMillis();
final long period = 1000 * 60 * 60 * 24;
this.scheduledExecutorService.scheduleAtFixedRate(new Runnable() {
@Override
public void run() {
try {
BrokerController.this.getBrokerStats().record();
} catch (Throwable e) {
log.error("schedule record error.", e);
}
}
}, initialDelay, period, TimeUnit.MILLISECONDS);
this.scheduledExecutorService.scheduleAtFixedRate(new Runnable() {
@Override
public void run() {
try {
BrokerController.this.consumerOffsetManager.persist();
} catch (Throwable e) {
log.error("schedule persist consumerOffset error.", e);
}
}
}, 1000 * 10, this.brokerConfig.getFlushConsumerOffsetInterval(), TimeUnit.MILLISECONDS);
this.scheduledExecutorService.scheduleAtFixedRate(new Runnable() {
@Override
public void run() {
try {
//清理出問題的consumer
BrokerController.this.protectBroker();
} catch (Exception e) {
log.error("protectBroker error.", e);
}
}
}, 3, 3, TimeUnit.MINUTES);
this.scheduledExecutorService.scheduleAtFixedRate(new Runnable() {
@Override
public void run() {
try {
//打印生產(chǎn), 消費隊列信息
BrokerController.this.printWaterMark();
} catch (Exception e) {
log.error("printWaterMark error.", e);
}
}
}, 10, 1, TimeUnit.SECONDS);
this.scheduledExecutorService.scheduleAtFixedRate(new Runnable() {
@Override
public void run() {
try {
log.info("dispatch behind commit log {} bytes", BrokerController.this.getMessageStore().dispatchBehindBytes());
} catch (Throwable e) {
log.error("schedule dispatchBehindBytes error.", e);
}
}
}, 1000 * 10, 1000 * 60, TimeUnit.MILLISECONDS);
//更新namesrv的地址信息
if (this.brokerConfig.getNamesrvAddr() != null) {
this.brokerOuterAPI.updateNameServerAddressList(this.brokerConfig.getNamesrvAddr());
log.info("Set user specified name server address: {}", this.brokerConfig.getNamesrvAddr());
} else if (this.brokerConfig.isFetchNamesrvAddrByAddressServer()) {
this.scheduledExecutorService.scheduleAtFixedRate(new Runnable() {
@Override
public void run() {
try {
BrokerController.this.brokerOuterAPI.fetchNameServerAddr();
} catch (Throwable e) {
log.error("ScheduledTask fetchNameServerAddr exception", e);
}
}
}, 1000 * 10, 1000 * 60 * 2, TimeUnit.MILLISECONDS);
}
//slave更新master地址, 從master同步消息數(shù)據(jù)
if (BrokerRole.SLAVE == this.messageStoreConfig.getBrokerRole()) {
if (this.messageStoreConfig.getHaMasterAddress() != null && this.messageStoreConfig.getHaMasterAddress().length() >= 6) {
this.messageStore.updateHaMasterAddress(this.messageStoreConfig.getHaMasterAddress());
this.updateMasterHAServerAddrPeriodically = false;
} else {
this.updateMasterHAServerAddrPeriodically = true;
}
this.scheduledExecutorService.scheduleAtFixedRate(new Runnable() {
@Override
public void run() {
try {
BrokerController.this.slaveSynchronize.syncAll();
} catch (Throwable e) {
log.error("ScheduledTask syncAll slave exception", e);
}
}
}, 1000 * 10, 1000 * 60, TimeUnit.MILLISECONDS);
} else {
//master打印slave落后的信息
this.scheduledExecutorService.scheduleAtFixedRate(new Runnable() {
@Override
public void run() {
try {
BrokerController.this.printMasterAndSlaveDiff();
} catch (Throwable e) {
log.error("schedule printMasterAndSlaveDiff error.", e);
}
}
}, 1000 * 10, 1000 * 60, TimeUnit.MILLISECONDS);
}
}
return result;
}
上述的全部過程中, 比較關(guān)鍵的操作是: 啟動netty服務(wù)端, 注冊多個消息處理器, 初始化線程池用來做消息的并發(fā)處理.
start
當(dāng)broker初始化了配置參數(shù)以后, 就可以開始啟動部分了, 啟動的部分會簡單一些:
- 啟動剛剛初始化的各個管理器:topicManager, consumerOffsetManager, subscriptionGroupManager, messageStore
- 開啟定時調(diào)度線程30秒向namesrv不斷上報自己的信息
public void start() throws Exception {
if (this.messageStore != null) {
this.messageStore.start();
}
if (this.remotingServer != null) {
this.remotingServer.start();
}
if (this.fastRemotingServer != null) {
this.fastRemotingServer.start();
}
if (this.brokerOuterAPI != null) {
this.brokerOuterAPI.start();
}
if (this.pullRequestHoldService != null) {
this.pullRequestHoldService.start();
}
if (this.clientHousekeepingService != null) {
this.clientHousekeepingService.start();
}
if (this.filterServerManager != null) {
this.filterServerManager.start();
}
this.registerBrokerAll(true, false);
this.scheduledExecutorService.scheduleAtFixedRate(new Runnable() {
@Override
public void run() {
try {
//向nameserver上報信息
BrokerController.this.registerBrokerAll(true, false);
} catch (Throwable e) {
log.error("registerBrokerAll Exception", e);
}
}
}, 1000 * 10, 1000 * 30, TimeUnit.MILLISECONDS);
if (this.brokerStatsManager != null) {
this.brokerStatsManager.start();
}
if (this.brokerFastFailure != null) {
this.brokerFastFailure.start();
}
}
后文介紹消息的接收.
參考資料
1.rocketMQ的broker模塊