详细介绍了RocketMQ的Broker与NameServer的心跳服务源码,主要包括三部分:
- Broker发送心跳注册请求源码;
- NameServer处理心跳注册请求源码;
- NameServer的心跳检测服务源码;
文章目录
1 Broker发送心跳注册请求
Broker启动过程中,会跟所有的NameServer建立并保持长连接,然后开启定时任务定时发送心跳包,心跳包中包含当前Broker信息,包括地址、名字、id等等,以及存储的所有Topic的信息。注册成功后,NameServer集群中就有Topic跟Broker的映射关系。
1.1 发送心跳包入口
具体的入口就是BrokerController#start方法:
/**
* BrokerController的方法
* 启动BrokerController
*/publicvoidstart()throwsException{//启动消息存储服务if(this.messageStore !=null){this.messageStore.start();}//启动netty远程服务if(this.remotingServer !=null){this.remotingServer.start();}//启动快速netty远程服务if(this.fastRemotingServer !=null){this.fastRemotingServer.start();}//文件监听器启动if(this.fileWatchService !=null){this.fileWatchService.start();}//broker对外api启动if(this.brokerOuterAPI !=null){this.brokerOuterAPI.start();}//长轮询拉取消息挂起服务启动if(this.pullRequestHoldService !=null){this.pullRequestHoldService.start();}//客户端连接心跳服务启动if(this.clientHousekeepingService !=null){this.clientHousekeepingService.start();}//过滤服务管理器启动if(this.filterServerManager !=null){this.filterServerManager.start();}//如果没有开启DLeger的相关设置,默认没有启动if(!messageStoreConfig.isEnableDLegerCommitLog()){//如果不是SLAVE,那么启动transactionalMessageCheckService事务消息检查服务startProcessorByHa(messageStoreConfig.getBrokerRole());//如果是SLAVE,那么启动定时任务每隔10s与master机器同步数据,采用slave主动拉取的方法//同步的内容包括topic配置,消费者消费位移、延迟消息偏移量、订阅组信息等handleSlaveSynchronize(messageStoreConfig.getBrokerRole());/*
* 强制注册当前broker信息到所有的nameserver
*/this.registerBrokerAll(true,false,true);}//启动定时任务,默认情况下每隔30s向nameServer进行一次注册,//时间间隔可以配置registerNameServerPeriod属性,允许的值是在1万到6万毫秒之间。this.scheduledExecutorService.scheduleAtFixedRate(newRunnable(){@Overridepublicvoidrun(){try{//定时发送心跳包并上报数据BrokerController.this.registerBrokerAll(true,false, brokerConfig.isForceRegister());}catch(Throwable e){
log.error("registerBrokerAll Exception", e);}}},1000*10,Math.max(10000,Math.min(brokerConfig.getRegisterNameServerPeriod(),60000)),TimeUnit.MILLISECONDS);//broker相关统计服务启动if(this.brokerStatsManager !=null){this.brokerStatsManager.start();}//broker快速失败服务启动if(this.brokerFastFailure !=null){this.brokerFastFailure.start();}}
在start方法中,可以看到在最后启动了一个定时任务,默认情况下每隔30s调用registerBrokerAll方法向所有的nameServer进行一次注册broker信息,时间间隔可以配置registerNameServerPeriod属性,允许的值是在1万到6万毫秒之间。这个定时任务就是Broker向nameserver发送的心跳包的定时任务,包括topic名、读、写队列个数、队列权限、是否有序等信息。
在这个定时任务之前,实际上还会调用一次registerBrokerAll方法,在broker首次启动时强制进行Broker注册。
1.2 registerBrokerAll注册broker信息
registerBrokerAll方法用于当前Broker将自身信息注册到所有的NameServer中。
内部调用的doRegisterBrokerAll方法执行注册,调用该方法之前,会判断是否需要注册,如果如果forceRegister为true,表示强制注册,或者如果当前broker应该注册,那么向nameServer进行注册。
在start方法中调用的registerBrokerAll方法,其forceRegister参数都为true,表示一定会强制注册的。
/**
* BrokerController的方法
* 注册Broker信息到NameServer,发送心跳包
*
* @param checkOrderConfig 是否检测顺序topic
* @param oneway 是否是单向
* @param forceRegister 是否强制注册
*/publicsynchronizedvoidregisterBrokerAll(finalboolean checkOrderConfig,boolean oneway,boolean forceRegister){//根据TopicConfigManager中的topic信息构建topic信息的传输协议对象,//在此前的topicConfigManager.load()方法中已经加载了所有topic信息,topic配置文件加载路径为{user.home}/store/config/topics.jsonTopicConfigSerializeWrapper topicConfigWrapper =this.getTopicConfigManager().buildTopicConfigSerializeWrapper();//如果当前broker权限不支持写或者读if(!PermName.isWriteable(this.getBrokerConfig().getBrokerPermission())||!PermName.isReadable(this.getBrokerConfig().getBrokerPermission())){ConcurrentHashMap<String,TopicConfig> topicConfigTable =newConcurrentHashMap<String,TopicConfig>();for(TopicConfig topicConfig : topicConfigWrapper.getTopicConfigTable().values()){//那么重新配置topic权限TopicConfig tmp =newTopicConfig(topicConfig.getTopicName(), topicConfig.getReadQueueNums(), topicConfig.getWriteQueueNums(),this.brokerConfig.getBrokerPermission());
topicConfigTable.put(topicConfig.getTopicName(), tmp);}
topicConfigWrapper.setTopicConfigTable(topicConfigTable);}/*
* 如果forceRegister为true,表示强制注册,或者如果当前broker应该注册,那么向nameServer进行注册
*/if(forceRegister ||needRegister(this.brokerConfig.getBrokerClusterName(),this.getBrokerAddr(),this.brokerConfig.getBrokerName(),this.brokerConfig.getBrokerId(),this.brokerConfig.getRegisterBrokerTimeoutMills())){/*
* 执行注册
*/doRegisterBrokerAll(checkOrderConfig, oneway, topicConfigWrapper);}}
1.2.1 needRegister是否需要注册
该方法用于判断当前broker是否需要向nameserver进行注册,当forceRegister参数为true的时候,表示强制注册,那么该方法的结果是无所谓的,如果forceRegister为false,那么borker是否需要向nameserver注册就得看这个方法的结果了。
其内部调用brokerOuterAPI#needRegister方法:
/**
* BrokerController的方法
* <p>
* broker是否需要向nemeserver中注册
*
* @param clusterName 集群名
* @param brokerAddr broker地址
* @param brokerName broker名字
* @param brokerId brkerid
* @param timeoutMills 超时时间
* @return broker是否需要向nemeserver中注册
*/privatebooleanneedRegister(finalString clusterName,finalString brokerAddr,finalString brokerName,finallong brokerId,finalint timeoutMills){//根据TopicConfigManager中的topic信息构建topic信息的传输协议对象,//在此前的topicConfigManager.load()方法中已经加载了所有topic信息,topic配置文件加载路径为{user.home}/store/config/topics.jsonTopicConfigSerializeWrapper topicConfigWrapper =this.getTopicConfigManager().buildTopicConfigSerializeWrapper();/*
* 获取所有nameServer的DataVersion数据,一一对比自身数据是否一致,如果有一个nameserver的DataVersion数据版本不一致则重新注册
*/List<Boolean> changeList = brokerOuterAPI.needRegister(clusterName, brokerAddr, brokerName, brokerId, topicConfigWrapper, timeoutMills);boolean needRegister =false;//如果和一个nameServer的数据版本不一致,则需要重新注册for(Boolean changed : changeList){if(changed){
needRegister =true;break;}}return needRegister;}
needRegister方法的逻辑也很简单,就是向所有nameServer发起请求(请求code为QUERY_DATA_VERSION,322),获取所有nameserver的DataVersion数据,然后一一对比自身的DataVersion数据是否一致,如果有一个nameserver的数据版本不一致则重新注册。
/**
* BrokerOuterAPI的方法
*/publicList<Boolean>needRegister(finalString clusterName,finalString brokerAddr,finalString brokerName,finallong brokerId,finalTopicConfigSerializeWrapper topicConfigWrapper,finalint timeoutMills){//创建一个CopyOnWriteArrayList类型的集合,用来保存请求的返回结果finalList<Boolean> changedList =newCopyOnWriteArrayList<>();//获取全部nameServer地址List<String> nameServerAddressList =this.remotingClient.getNameServerAddressList();if(nameServerAddressList !=null&& nameServerAddressList.size()>0){finalCountDownLatch countDownLatch =newCountDownLatch(nameServerAddressList.size());for(finalString namesrvAddr : nameServerAddressList){
brokerOuterExecutor.execute(newRunnable(){@Overridepublicvoidrun(){try{/*
* 构造请求头,将一些broker信息放入请求头
*/QueryDataVersionRequestHeader requestHeader =newQueryDataVersionRequestHeader();
requestHeader.setBrokerAddr(brokerAddr);
requestHeader.setBrokerId(brokerId);
requestHeader.setBrokerName(brokerName);
requestHeader.setClusterName(clusterName);//构建远程调用请求对象,code为QUERY_DATA_VERSION,322RemotingCommand request =RemotingCommand.createRequestCommand(RequestCode.QUERY_DATA_VERSION, requestHeader);
request.setBody(topicConfigWrapper.getDataVersion().encode());RemotingCommand response = remotingClient.invokeSync(namesrvAddr, request, timeoutMills);DataVersion nameServerDataVersion =null;Boolean changed =false;switch(response.getCode()){caseResponseCode.SUCCESS:{//QueryDataVersionResponseHeader queryDataVersionResponseHeader =(QueryDataVersionResponseHeader) response.decodeCommandCustomHeader(QueryDataVersionResponseHeader.class);
changed = queryDataVersionResponseHeader.getChanged();byte[] body = response.getBody();if(body !=null){//获取nameserver的dataversion
nameServerDataVersion =DataVersion.decode(body,DataVersion.class);//如果当前broker的dataversion与nameserver的dataversion不相等,则表示需要继续宁更新if(!topicConfigWrapper.getDataVersion().equals(nameServerDataVersion)){
changed =true;}}if(changed ==null|| changed){
changedList.add(Boolean.TRUE);}}default:break;}
log.warn("Query data version from name server {} OK,changed {}, broker {},name server {}", namesrvAddr, changed, topicConfigWrapper.getDataVersion(), nameServerDataVersion ==null?"": nameServerDataVersion);}catch(Exception e){
changedList.add(Boolean.TRUE);
log.error("Query data version from name server {} Exception, {}", namesrvAddr, e);}finally{
countDownLatch.countDown();}}});}try{
countDownLatch.await(timeoutMills,TimeUnit.MILLISECONDS);}catch(InterruptedException e){
log.error("query dataversion from nameserver countDownLatch await Exception", e);}}return changedList;}
1.2.1.1 DataVersion介绍
DataVersion是RocketMQ的数据版本控制机制。其结构比较简单,核心属性方法如下:
/**
* 时间戳毫秒值
*/privatelong timestamp =System.currentTimeMillis();/**
* 版本号
*/privateAtomicLong counter =newAtomicLong(0);/**
* 拷贝目标dataVersion的数据,在从文件恢复数据的时候会用到
*/publicvoidassignNewOne(finalDataVersion dataVersion){this.timestamp = dataVersion.timestamp;this.counter.set(dataVersion.counter.get());}/**
* 更新时间戳以及counter到下一个版本
*/publicvoidnextVersion(){this.timestamp =System.currentTimeMillis();this.counter.incrementAndGet();}
他的nextVersion方法被调用时,将会引起timestamp和counter的改变,一般来说,当新创建broker,或者更新topic的信息的时候nextVersion方法会被调用。
Dataversion和topic的配置都被持久化到topics.json文件中,其格式如下:
{"dataVersion":{"counter":3,"timestamp":1651398321850},"topicConfigTable":{"SCHEDULE_TOPIC_XXXX":{"order":false,"perm":6,"readQueueNums":18,"topicFilterType":"SINGLE_TAG","topicName":"SCHEDULE_TOPIC_XXXX","topicSysFlag":0,"writeQueueNums":18},"TopicTest":{"order":false,"perm":6,"readQueueNums":4,"topicFilterType":"SINGLE_TAG","topicName":"TopicTest","topicSysFlag":0,"writeQueueNums":4},"SELF_TEST_TOPIC":{"order":false,"perm":6,"readQueueNums":1,"topicFilterType":"SINGLE_TAG","topicName":"SELF_TEST_TOPIC","topicSysFlag":0,"writeQueueNums":1},"DefaultCluster":{"order":false,"perm":7,"readQueueNums":16,"topicFilterType":"SINGLE_TAG","topicName":"DefaultCluster","topicSysFlag":0,"writeQueueNums":16},"DefaultCluster_REPLY_TOPIC":{"order":false,"perm":6,"readQueueNums":1,"topicFilterType":"SINGLE_TAG","topicName":"DefaultCluster_REPLY_TOPIC","topicSysFlag":0,"writeQueueNums":1},"RMQ_SYS_TRANS_HALF_TOPIC":{"order":false,"perm":6,"readQueueNums":1,"topicFilterType":"SINGLE_TAG","topicName":"RMQ_SYS_TRANS_HALF_TOPIC","topicSysFlag":0,"writeQueueNums":1},"broker-a":{"order":false,"perm":7,"readQueueNums":1,"topicFilterType":"SINGLE_TAG","topicName":"broker-a","topicSysFlag":0,"writeQueueNums":1},"TBW102":{"order":false,"perm":7,"readQueueNums":8,"topicFilterType":"SINGLE_TAG","topicName":"TBW102","topicSysFlag":0,"writeQueueNums":8},"BenchmarkTest":{"order":false,"perm":6,"readQueueNums":1024,"topicFilterType":"SINGLE_TAG","topicName":"BenchmarkTest","topicSysFlag":0,"writeQueueNums":1024},"OFFSET_MOVED_EVENT":{"order":false,"perm":6,"readQueueNums":1,"topicFilterType":"SINGLE_TAG","topicName":"OFFSET_MOVED_EVENT","topicSysFlag":0,"writeQueueNums":1},"%RETRY%please_rename_unique_group_name_4":{"order":false,"perm":6,"readQueueNums":1,"topicFilterType":"SINGLE_TAG","topicName":"%RETRY%please_rename_unique_group_name_4","topicSysFlag":0,"writeQueueNums":1}}}
Dataversion和topic的配置被加载到内存之后,分别会解析成为topicConfigManager的topicConfigTablehe属性和dataVersion属性。
1.2.2 doRegisterBrokerAll注册broker信息
doRegisterBrokerAll方法的逻辑也很简单,就是向所有nameServer发起请求。
/**
* BrokerController的方法
*
* @param checkOrderConfig 是否检测顺序topic
* @param oneway 是否是单向
* @param topicConfigWrapper topic信息的传输协议包装对象
*/privatevoiddoRegisterBrokerAll(boolean checkOrderConfig,boolean oneway,TopicConfigSerializeWrapper topicConfigWrapper){/*
* 执行注册,broker作为客户端向所有的nameserver发起注册请求
*/List<RegisterBrokerResult> registerBrokerResultList =this.brokerOuterAPI.registerBrokerAll(this.brokerConfig.getBrokerClusterName(),this.getBrokerAddr(),this.brokerConfig.getBrokerName(),this.brokerConfig.getBrokerId(),this.getHAServerAddr(),//包含了携带topic信息的topicConfigTable,以及版本信息的dataVersion//这两个信息保存在持久化文件topics.json中
topicConfigWrapper,this.filterServerManager.buildNewFilterServerList(),
oneway,this.brokerConfig.getRegisterBrokerTimeoutMills(),this.brokerConfig.isCompressedRegister());/*
* 对执行结果进行处理,选择抵押给调用的结果作为默认数据设置
*/if(registerBrokerResultList.size()>0){RegisterBrokerResult registerBrokerResult = registerBrokerResultList.get(0);if(registerBrokerResult !=null){if(this.updateMasterHAServerAddrPeriodically && registerBrokerResult.getHaServerAddr()!=null){this.messageStore.updateHaMasterAddress(registerBrokerResult.getHaServerAddr());}this.slaveSynchronize.setMasterAddr(registerBrokerResult.getMasterAddr());if(checkOrderConfig){this.getTopicConfigManager().updateOrderTopicConfig(registerBrokerResult.getKvTable());}}}}
内部调用BrokerOuterAPI#registerBrokerAll方法,BrokerOuterAPI类专门提供了broker向外部发起请求的api方法。
因为broker要向所有的nameServer进行注册,为了提升性能,registerBrokerAll方法里面使用了多线程机制,使用brokerOuterExecutor线程池并行的发起对于每个nameserver的注册请求。
有了多线程提升效率,自然就需要保证线程安全和控制并发:
- 因为需要在多线程中将执行结果并行存入集合中, RocketMQ使用了CopyOnWriteArrayList这个并发集合来保证线程安全。CopyOnWriteArrayList采用COW(Copy On Write)机制,即写是复制,读数据时完全没有控制,即不会加锁。写数据时加独占锁,并且会复制出一个新的List,在新的List中写入数据,写完了之后使用新的List替换旧的List。
- 虽然对于注册的请求使用了线程池异步执行,但是主线程却需要等待这些请求都执行完毕,所有的结果才能继续向下执行,对于这种并发控制, RocketMQ使用了CountDownLatch倒计数器,它能够使得主线程阻塞,确保在其他线程任务执行完毕之后,才会唤醒主线程继续执行后续逻辑。
/**
* BrokerOuterAPI的方法
* <p>
* broker向nameserver进行注册
*/publicList<RegisterBrokerResult>registerBrokerAll(finalString clusterName,finalString brokerAddr,finalString brokerName,finallong brokerId,finalString haServerAddr,finalTopicConfigSerializeWrapper topicConfigWrapper,finalList<String> filterServerList,finalboolean oneway,finalint timeoutMills,finalboolean compressed){//创建一个CopyOnWriteArrayList类型的集合,用来保存请求的返回结果finalList<RegisterBrokerResult> registerBrokerResultList =newCopyOnWriteArrayList<>();//获取全部nameServer地址List<String> nameServerAddressList =this.remotingClient.getNameServerAddressList();if(nameServerAddressList !=null&& nameServerAddressList.size()>0){/*
* 构造请求头,将一些broker信息放入请求头
*/finalRegisterBrokerRequestHeader requestHeader =newRegisterBrokerRequestHeader();
requestHeader.setBrokerAddr(brokerAddr);
requestHeader.setBrokerId(brokerId);
requestHeader.setBrokerName(brokerName);
requestHeader.setClusterName(clusterName);
requestHeader.setHaServerAddr(haServerAddr);
requestHeader.setCompressed(compressed);/*
* 构造请求体,将携带topic信息的topicConfigTable,以及版本信息的dataVersion,以及消费过滤信息集合放入请求体
*/RegisterBrokerBody requestBody =newRegisterBrokerBody();
requestBody.setTopicConfigSerializeWrapper(topicConfigWrapper);
requestBody.setFilterServerList(filterServerList);finalbyte[] body = requestBody.encode(compressed);finalint bodyCrc32 =UtilAll.crc32(body);
requestHeader.setBodyCrc32(bodyCrc32);/*
* 使用CountDownLatch作为倒计数器,用于并发控制
* CountDownLatch可以让主线程等待,直到任务全部执行完毕之后,再唤醒主线程继续后面的逻辑
*/finalCountDownLatch countDownLatch =newCountDownLatch(nameServerAddressList.size());/*
* 采用线程池的方式,即多线程并发的向所有的nameserver发起注册请求
*/for(finalString namesrvAddr : nameServerAddressList){//并发的执行线程任务
brokerOuterExecutor.execute(newRunnable(){@Overridepublicvoidrun(){try{//发起注册请求RegisterBrokerResult result =registerBroker(namesrvAddr, oneway, timeoutMills, requestHeader, body);if(result !=null){
registerBrokerResultList.add(result);}
log.info("register broker[{}]to name server {} OK", brokerId, namesrvAddr);}catch(Exception e){
log.warn("registerBroker Exception, {}", namesrvAddr, e);}finally{/*
* 每一个请求执行完毕,无论是正常还是异常,都需要减少一个计数
*/
countDownLatch.countDown();}}});}try{/*
* 主线程在此限时等待6000ms,直到上面的任务全部执行完毕之后,计数变为0,会唤醒主线程继续后面的逻辑
*/
countDownLatch.await(timeoutMills,TimeUnit.MILLISECONDS);}catch(InterruptedException e){}}return registerBrokerResultList;}
1.2.2.1 registerBroker注册broker
在上面的多线程代码中,线程任务是调用另一个registerBroker方法,该方法真正的执行向一个nameserver发起注册的请求。
broker注册请求为同步请求,code为REGISTER_BROKER,103,注册的信息主要包括自身的所有topic数据、dataVersion、filterServerList、以及包括集群名、broker地址、broker名、brokerId等等在内的一些broker自身的信息。
/**
* BrokerOuterAPI的方法
* <p>
* 注册broker
*/privateRegisterBrokerResultregisterBroker(finalString namesrvAddr,finalboolean oneway,finalint timeoutMills,finalRegisterBrokerRequestHeader requestHeader,finalbyte[] body
)throwsRemotingCommandException,MQBrokerException,RemotingConnectException,RemotingSendRequestException,RemotingTimeoutException,InterruptedException{//构建远程调用请求对象,code为REGISTER_BROKER,103RemotingCommand request =RemotingCommand.createRequestCommand(RequestCode.REGISTER_BROKER, requestHeader);
request.setBody(body);//如果是单向请求,则broker发起异步请求即可返回,不必关心执行结果,注册请求不是单向请求if(oneway){try{this.remotingClient.invokeOneway(namesrvAddr, request, timeoutMills);}catch(RemotingTooMuchRequestException e){// Ignore}returnnull;}/*
* 通过remotingClient发起同步调用,非单向请求,即需要同步的获取结果
*/RemotingCommand response =this.remotingClient.invokeSync(namesrvAddr, request, timeoutMills);assert response !=null;switch(response.getCode()){caseResponseCode.SUCCESS:{/*
* 解析响应数据,封装结果
*/RegisterBrokerResponseHeader responseHeader =(RegisterBrokerResponseHeader) response.decodeCommandCustomHeader(RegisterBrokerResponseHeader.class);RegisterBrokerResult result =newRegisterBrokerResult();
result.setMasterAddr(responseHeader.getMasterAddr());
result.setHaServerAddr(responseHeader.getHaServerAddr());if(response.getBody()!=null){
result.setKvTable(KVTable.decode(response.getBody(),KVTable.class));}return result;}default:break;}thrownewMQBrokerException(response.getCode(), response.getRemark(), requestHeader ==null?null: requestHeader.getBrokerAddr());}
2 NameServer处理心跳注册请求
Broker发送了心跳包之后,nameserver会进行专门的处理,保存或者更新broker上报的心跳包数据。
2.1 处理心跳包入口
NameServer的默认网络处理器是DefaultRequestProcessor,因此心跳请求的入口也就是DefaultRequestProcessor#processRequest方法。
processRequest方法是一个通用的请求处理入口方法,内部会根据请求的不同requestCode进入分发处理,心跳请求的requestCode就是REGISTER_BROKER,103。
caseRequestCode.REGISTER_BROKER:/*
* 处理broker心跳请求的逻辑
*///获取broker版本Version brokerVersion =MQVersion.value2Version(request.getVersion());//如果大于3.0.11版本则调用registerBrokerWithFilterServer,否则调用registerBrokerif(brokerVersion.ordinal()>=MQVersion.Version.V3_0_11.ordinal()){returnthis.registerBrokerWithFilterServer(ctx, request);}else{returnthis.registerBroker(ctx, request);}caseRequestCode.UNREGISTER_BROKER:/*
* 处理解除broker注册的逻辑
*/returnthis.unregisterBroker(ctx, request);
可以看到nameserver会调用registerBrokerWithFilterServer方法来处理大于3.0.11版本的broker的注册请求。
我们进入registerBrokerWithFilterServer方法,可以看到,该方法会解析请求头、请求体中的内容,内容包括topic信息、版本信息dataVersion、消息过滤信息filterServerList、以及broker基本信息,例如broker地址、名字等等。最后会调用routeInfoManager#registerBroker方法实现broker信息的注册。
/**
* DefaultRequestProcessor的方法
* <p>
* 处理broker的心跳请求
* 心跳请求的内容包括topic信息、版本信息dataVersion、消息过滤信息filterServerList、以及broker基本信息,例如broker地址、名字等等
*/publicRemotingCommandregisterBrokerWithFilterServer(ChannelHandlerContext ctx,RemotingCommand request)throwsRemotingCommandException{//创建返回数据finalRemotingCommand response =RemotingCommand.createResponseCommand(RegisterBrokerResponseHeader.class);//构建响应头finalRegisterBrokerResponseHeader responseHeader =(RegisterBrokerResponseHeader) response.readCustomHeader();//获取请求头finalRegisterBrokerRequestHeader requestHeader =(RegisterBrokerRequestHeader) request.decodeCommandCustomHeader(RegisterBrokerRequestHeader.class);//校验crc32if(!checksum(ctx, request, requestHeader)){
response.setCode(ResponseCode.SYSTEM_ERROR);
response.setRemark("crc32 not match");return response;}/*
* 解析请求体的信息成为RegisterBrokerBody对象,内部包含发送请求时封装的filterServerList和topicConfigSerializeWrapper对象
* 包括topic信息、版本信息dataVersion、消息过滤信息filterServerList
*/RegisterBrokerBody registerBrokerBody =newRegisterBrokerBody();if(request.getBody()!=null){try{//解析请求体的信息
registerBrokerBody =RegisterBrokerBody.decode(request.getBody(), requestHeader.isCompressed());}catch(Exception e){thrownewRemotingCommandException("Failed to decode RegisterBrokerBody", e);}}else{
registerBrokerBody.getTopicConfigSerializeWrapper().getDataVersion().setCounter(newAtomicLong(0));
registerBrokerBody.getTopicConfigSerializeWrapper().getDataVersion().setTimestamp(0);}/*
* broker信息注册
*/RegisterBrokerResult result =this.namesrvController.getRouteInfoManager().registerBroker(
requestHeader.getClusterName(),
requestHeader.getBrokerAddr(),
requestHeader.getBrokerName(),
requestHeader.getBrokerId(),
requestHeader.getHaServerAddr(),
registerBrokerBody.getTopicConfigSerializeWrapper(),
registerBrokerBody.getFilterServerList(),
ctx.channel());
responseHeader.setHaServerAddr(result.getHaServerAddr());
responseHeader.setMasterAddr(result.getMasterAddr());//从configTable获取顺序消息的配置,configTable可用于存储一些配置信息,实现匹配的namespace隔离。//目前版本似乎不太起作用,或许是当初设想但未利用起来的设计,返回nullbyte[] jsonValue =this.namesrvController.getKvConfigManager().getKVListByNamespace(NamesrvUtil.NAMESPACE_ORDER_TOPIC_CONFIG);
response.setBody(jsonValue);
response.setCode(ResponseCode.SUCCESS);
response.setRemark(null);return response;}
2.2 Nameserver注册broker信息
2.3.1 RouteInfoManager的介绍
Nameserver通过routeInfoManager#registerBroker方法实现broker信息的注册。我们先来看看RouteInfoManager的基本属性,这个类位于namesrv模块中,用来管理nameServer上的关于真个RocketMQ集群的各种路由信息,nameServer作为轻量级的注册中心,RouteInfoManager这个类非常的重要。
/**
* Broker过期时间,默认120秒,如果当前时间大于最后修改时间加上Broker过期时间,那么就剔除该Broker
*/privatefinalstaticlong BROKER_CHANNEL_EXPIRED_TIME =1000*60*2;/**
* 读写锁,用于在获取路由信息时保证并发安全的同时提升效率
*/privatefinalReadWriteLock lock =newReentrantReadWriteLock();/**
* Topic到Topic下面的队列集合的路由信息
*/privatefinalHashMap<String/* topic */,List<QueueData>> topicQueueTable;/**
* brokerName到BrokerData的路由信息,BrokerData包含broker的基础信息,例如brokerName、brokerAddr、cluster等
*/privatefinalHashMap<String/* brokerName */,BrokerData> brokerAddrTable;/**
* clusterName到cluster下面的brokerName的路由信息
*/privatefinalHashMap<String/* clusterName */,Set<String/* brokerName */>> clusterAddrTable;/**
* brokerAddr名到BrokerLiveInfo的路由信息,BrokerLiveInfo存储broker的状态信息,包括上次接收心跳时间,数据版本号等
*/privatefinalHashMap<String/* brokerAddr */,BrokerLiveInfo> brokerLiveTable;/**
* brokerAddr名到Filter Server集合的路由信息,用于类模式消息过滤。
*/privatefinalHashMap<String/* brokerAddr */,List<String>/* Filter Server */> filterServerTable;
2.3.2 registerBroker注册broker
该方法用于注册broker,也就是对broker的各种路由信息进行更新或者注册。
其主要步骤为:
- 加写锁,保证线程安全。
- 存入或者更新brokerName信息集合clusterAddrTable。
- 存入或者更新broker基本信息集合brokerAddrTable。存入一个brokerData对象。
- 如果当前broker是主broker节点。更新或者创建topic的队列配置信息集合topicQueueTable。
- 存入或者更新中broker状态信息集合brokerLiveTable。存入或者更新的信息包括最新的更新时间戳设置为当前时间,brokerLiveTable被nameServer用于执行心跳检测操作。
- 存入或者更新消费过滤信息集合filterServerList。ClassFilter模式的消费过滤集合的操作。
- 如果当前broker不是主broker节点。对返回结果result设置HaServerAddr以及MasterAddr的地址。
- 释放写锁。
/**
* RouteInfoManager的方法
*
* @param clusterName 集群名
* @param brokerAddr broker地址
* @param brokerName broker名
* @param brokerId broker Id
* @param haServerAddr 高可用服务地址
* @param topicConfigWrapper topic配置
* @param filterServerList 消费过滤
* @param channel 通道
* @return 注册的结果
*/publicRegisterBrokerResultregisterBroker(finalString clusterName,finalString brokerAddr,finalString brokerName,finallong brokerId,finalString haServerAddr,finalTopicConfigSerializeWrapper topicConfigWrapper,finalList<String> filterServerList,finalChannel channel){//构建注册结果对象RegisterBrokerResult result =newRegisterBrokerResult();try{try{/*
* 加本地写锁,防止并发
*/this.lock.writeLock().lockInterruptibly();/*
* 1 存入或者更新brokerName信息集合clusterAddrTable
* brokerName的操作
*///获取集群下所有的brokerName集合Set<String> brokerNames =this.clusterAddrTable.get(clusterName);if(null== brokerNames){//如果此前没有此集群的broker的话就新创建一个,并且将当前brokerName加入brokerName集合中
brokerNames =newHashSet<String>();this.clusterAddrTable.put(clusterName, brokerNames);}//将当前brokerName加入brokerName集合中,set集合不会重复
brokerNames.add(brokerName);/*
* 2 存入或者更新broker基本信息集合brokerAddrTable
* brokerId和brokerAddr的操作
*///是否是第一次注册的标志位boolean registerFirst =false;//从brokerAddrTable获取当前brokerName对应的BrokerData信息BrokerData brokerData =this.brokerAddrTable.get(brokerName);if(null== brokerData){//如果brokerData为null,则表示是第一次注册,那么新建brokerData并且存入brokerAddrTable
registerFirst =true;
brokerData =newBrokerData(clusterName, brokerName,newHashMap<Long,String>());this.brokerAddrTable.put(brokerName, brokerData);}//获取此集群的此brokerName下面的 brokerId到brokerAdder的映射map//相同的clusterName和brokerName下面可能有多个broker节点,比如主从架构Map<Long,String> brokerAddrsMap = brokerData.getBrokerAddrs();//Switch slave to master: first remove <1, IP:PORT> in namesrv, then add <0, IP:PORT>//The same IP:PORT must only have one record in brokerAddrTableIterator<Entry<Long,String>> it = brokerAddrsMap.entrySet().iterator();while(it.hasNext()){Entry<Long,String> item = it.next();//如果broker地址一样,但是brokerId不一样,那么此时的情况应该是: 主从架构下,master节点挂了,slave节点成为master//我需要首先移除此前存在的相同brokerAddr的元素//例如,此前存在<0, IP1:PORT> <1, IP2:PORT> 两个元素。此时主节点挂了,从节点成为主节点,上报的数据会变成<0, IP2:PORT>if(null!= brokerAddr && brokerAddr.equals(item.getValue())&& brokerId != item.getKey()){
it.remove();}}//更新,将当前broekr的Id到地址的关系存入brokerAddrsMapString oldAddr = brokerData.getBrokerAddrs().put(brokerId, brokerAddr);
registerFirst = registerFirst ||(null== oldAddr);/*
* 3 更新或者创建topic的队列配置信息集合topicQueueTable
* 针对Master节点的操作
*///如果当前broker是主broker节点if(null!= topicConfigWrapper
&&MixAll.MASTER_ID == brokerId){//如果当前broker的topic配置信息的数据版本DataVersion发生了变化if(this.isBrokerTopicConfigChanged(brokerAddr, topicConfigWrapper.getDataVersion())|| registerFirst){//获取上报的topic配置信息map集合ConcurrentMap<String,TopicConfig> tcTable =
topicConfigWrapper.getTopicConfigTable();if(tcTable !=null){for(Map.Entry<String,TopicConfig> entry : tcTable.entrySet()){//更新或者新建topic配置信息this.createAndUpdateQueueData(brokerName, entry.getValue());}}}}/*
* 4 存入或者更新中broker状态信息集合brokerLiveTable
* 存入或者更新的信息包括最新的更新时间戳设置为当前时间,brokerLiveTable被nameServer用于执行心跳检测操作
*/BrokerLiveInfo prevBrokerLiveInfo =this.brokerLiveTable.put(brokerAddr,newBrokerLiveInfo(System.currentTimeMillis(),
topicConfigWrapper.getDataVersion(),
channel,
haServerAddr));//如果此前的prevBrokerLiveInfo为null,那么表示新上报broker,打印日志if(null== prevBrokerLiveInfo){
log.info("new broker registered, {} HAServer: {}", brokerAddr, haServerAddr);}/*
* 5 存入或者更新消费过滤信息集合filterServerList
* ClassFilter模式的消费过滤集合的操作
*/if(filterServerList !=null){if(filterServerList.isEmpty()){this.filterServerTable.remove(brokerAddr);}else{this.filterServerTable.put(brokerAddr, filterServerList);}}/*
* 6 对返回结果设置HaServerAddr以及MasterAddr的地址
* 针对slave节点的操作
*/if(MixAll.MASTER_ID != brokerId){//获取mater节点的地址String masterAddr = brokerData.getBrokerAddrs().get(MixAll.MASTER_ID);if(masterAddr !=null){//获取master节点的brokerLiveInfo信息BrokerLiveInfo brokerLiveInfo =this.brokerLiveTable.get(masterAddr);if(brokerLiveInfo !=null){//将master节点的HaServerAddr以及mater节点的地址存入result中返回给slave节点
result.setHaServerAddr(brokerLiveInfo.getHaServerAddr());
result.setMasterAddr(masterAddr);}}}}finally{/*
* 解锁
*/this.lock.writeLock().unlock();}}catch(Exception e){
log.error("registerBroker Exception", e);}return result;}
3 NameServer的心跳检测服务
NameServer在启动的时候,会启动一个定时周期任务。默认每隔10秒执行一次扫描无效的Broker,并清除Broker相关路由信息的任务。
/*
* 5 启动一个定时任务
* 首次启动延迟5秒执行,此后每隔10秒执行一次扫描无效的Broker,并清除Broker相关路由信息的任务
*/this.scheduledExecutorService.scheduleAtFixedRate(newRunnable(){@Overridepublicvoidrun(){//扫描notActive的brokerNamesrvController.this.routeInfoManager.scanNotActiveBroker();}},5,10,TimeUnit.SECONDS);
3.1 scanNotActiveBroker扫描清除不活跃broker
这个方法就是nameserver中每隔10秒执行一次扫描无效的Broker,并清除无效的Broker的连接以及路由信息的任务方法。
方法很简单,就是扫描RouteInfoManager的brokerLiveTable集合中的每一个BrokerLiveInfo。然后判断如果当前时间戳 大于 上次接收心跳的时间戳 + Broker过期时间,那么就剔除该Broker,Broker过期时间默认120秒。
所谓的剔除该Broker,就是两件事:
- 调用RemotingUtil#closeChannel关闭和当前broker的通道,即关闭与此broker的socket连接。
- 调用RouteInfoManager#onChannelDestroy清除该broker的无效的路由信息。
/**
* RouteInfoManager的方法
* <p>
* 扫描brokerLiveTable,清除无效的broker连接和路由信息
*/publicvoidscanNotActiveBroker(){//根据brokerLiveTable进行路由信息的校验和移除Iterator<Entry<String,BrokerLiveInfo>> it =this.brokerLiveTable.entrySet().iterator();while(it.hasNext()){Entry<String,BrokerLiveInfo> next = it.next();//获取上次接收心跳的时间戳long last = next.getValue().getLastUpdateTimestamp();//如果当前时间戳 大于 上次接收心跳的时间戳 + Broker过期时间,那么就剔除该Broker,Broker过期时间默认120秒//即如果某个Broker 120秒内没有上报心跳包,那么任务该broker失效了,可能是宕机了,于是移除相关路由信息if((last + BROKER_CHANNEL_EXPIRED_TIME)<System.currentTimeMillis()){/*
* 1 关闭和当前broker的通道,即关闭与此broker的socket连接
*/RemotingUtil.closeChannel(next.getValue().getChannel());//从brokerLiveTable中移除此项
it.remove();//记录日志
log.warn("The broker channel expired, {} {}ms", next.getKey(), BROKER_CHANNEL_EXPIRED_TIME);/*
* 2 清除无效的路由信息
*/this.onChannelDestroy(next.getKey(), next.getValue().getChannel());}}}
3.1.1 onChannelDestroy清除路由信息
该方法用于在断开连接之后请求与该broker相关的无效路由信息。
删除信息的时候同样需要先加写锁,然后从brokerLiveTable、filterServerTable、brokerAddrTable、clusterAddrTable、topicQueueTable这五个路由表中删除数据,可以看作是registerBroker注册broker信息方法的逆向操作,比较简单。
/**
* 清除无效路由信息
*
* @param remoteAddr 无效的brokerAddr
* @param channel 通道
*/publicvoidonChannelDestroy(String remoteAddr,Channel channel){String brokerAddrFound =null;if(channel !=null){try{try{//加读锁在brokerLiveTable中查找channel相等的brokerAddrthis.lock.readLock().lockInterruptibly();Iterator<Entry<String,BrokerLiveInfo>> itBrokerLiveTable =this.brokerLiveTable.entrySet().iterator();while(itBrokerLiveTable.hasNext()){Entry<String,BrokerLiveInfo> entry = itBrokerLiveTable.next();if(entry.getValue().getChannel()== channel){
brokerAddrFound = entry.getKey();break;}}}finally{//释放读锁this.lock.readLock().unlock();}}catch(Exception e){
log.error("onChannelDestroy Exception", e);}}if(null== brokerAddrFound){
brokerAddrFound = remoteAddr;}else{
log.info("the broker's channel destroyed, {}, clean it's data structure at once", brokerAddrFound);}/*
* 请求无效的broker路由信息
*/if(brokerAddrFound !=null&& brokerAddrFound.length()>0){try{try{//加写锁this.lock.writeLock().lockInterruptibly();/*
* 1 从brokerLiveTable中移除数据
*/this.brokerLiveTable.remove(brokerAddrFound);/*
* 2 从filterServerTable中移除数据
*/this.filterServerTable.remove(brokerAddrFound);String brokerNameFound =null;boolean removeBrokerName =false;/*
* 3 从brokerAddrTable中删除数据
*/Iterator<Entry<String,BrokerData>> itBrokerAddrTable =this.brokerAddrTable.entrySet().iterator();while(itBrokerAddrTable.hasNext()&&(null== brokerNameFound)){BrokerData brokerData = itBrokerAddrTable.next().getValue();//遍历brokerData里面的BrokerAddrsIterator<Entry<Long,String>> it = brokerData.getBrokerAddrs().entrySet().iterator();while(it.hasNext()){Entry<Long,String> entry = it.next();Long brokerId = entry.getKey();String brokerAddr = entry.getValue();//将BrokerAddrs中对应的brokerAddr删除if(brokerAddr.equals(brokerAddrFound)){
brokerNameFound = brokerData.getBrokerName();
it.remove();
log.info("remove brokerAddr[{}, {}] from brokerAddrTable, because channel destroyed",
brokerId, brokerAddr);break;}}//如果BrokerAddrs为空了,那么直接移除整个brokerAddrTable的项目if(brokerData.getBrokerAddrs().isEmpty()){
removeBrokerName =true;
itBrokerAddrTable.remove();
log.info("remove brokerName[{}] from brokerAddrTable, because channel destroyed",
brokerData.getBrokerName());}}/*
* 4 从clusterAddrTable中删除数据
*/if(brokerNameFound !=null&& removeBrokerName){Iterator<Entry<String,Set<String>>> it =this.clusterAddrTable.entrySet().iterator();while(it.hasNext()){Entry<String,Set<String>> entry = it.next();String clusterName = entry.getKey();Set<String> brokerNames = entry.getValue();//将brokerNames中对应的brokerName删除boolean removed = brokerNames.remove(brokerNameFound);if(removed){
log.info("remove brokerName[{}], clusterName[{}] from clusterAddrTable, because channel destroyed",
brokerNameFound, clusterName);//如果brokerNames为空了,那么直接移除整个clusterAddrTable的项目if(brokerNames.isEmpty()){
log.info("remove the clusterName[{}] from clusterAddrTable, because channel destroyed and no broker in this cluster",
clusterName);
it.remove();}break;}}}/*
* 5 从topicQueueTable中删除数据
*/if(removeBrokerName){Iterator<Entry<String,List<QueueData>>> itTopicQueueTable =this.topicQueueTable.entrySet().iterator();while(itTopicQueueTable.hasNext()){Entry<String,List<QueueData>> entry = itTopicQueueTable.next();String topic = entry.getKey();List<QueueData> queueDataList = entry.getValue();Iterator<QueueData> itQueueData = queueDataList.iterator();while(itQueueData.hasNext()){QueueData queueData = itQueueData.next();//将queueDataList中与brokerName一致的queueData删除if(queueData.getBrokerName().equals(brokerNameFound)){
itQueueData.remove();
log.info("remove topic[{} {}], from topicQueueTable, because channel destroyed",
topic, queueData);}}//如果queueDataList为空了,那么直接移除整个topicQueueTable的项目if(queueDataList.isEmpty()){
itTopicQueueTable.remove();
log.info("remove topic[{}] all queue, from topicQueueTable, because channel destroyed",
topic);}}}}finally{//释放锁this.lock.writeLock().unlock();}}catch(Exception e){
log.error("onChannelDestroy Exception", e);}}}
如果broker和nameserver之间的长连接异常关闭,那么此前绑定的BrokerHousekeepingService就发挥了作用,BrokerHousekeepingService继承了ChannelEventListener,当触发连接异常事件时,BrokerHousekeepingService内部的方法同样会调用RouteInfoManager#onChannelDestroy清除路由信息。
publicclassBrokerHousekeepingServiceimplementsChannelEventListener{privatestaticfinalInternalLogger log =InternalLoggerFactory.getLogger(LoggerName.NAMESRV_LOGGER_NAME);privatefinalNamesrvController namesrvController;publicBrokerHousekeepingService(NamesrvController namesrvController){this.namesrvController = namesrvController;}//连接事件,不处理@OverridepublicvoidonChannelConnect(String remoteAddr,Channel channel){}//连接关闭事件@OverridepublicvoidonChannelClose(String remoteAddr,Channel channel){this.namesrvController.getRouteInfoManager().onChannelDestroy(remoteAddr, channel);}//连接异常事件@OverridepublicvoidonChannelException(String remoteAddr,Channel channel){this.namesrvController.getRouteInfoManager().onChannelDestroy(remoteAddr, channel);}//连接闲置事件@OverridepublicvoidonChannelIdle(String remoteAddr,Channel channel){this.namesrvController.getRouteInfoManager().onChannelDestroy(remoteAddr, channel);}}
4 总结
本那次我们学习了Broker和NameServer之间的心跳服务的源码。包括Broker的心跳上报、NameServer的心跳处理、NameServer的心跳检测三部分的源码。从这些源码可以知道:
- broker的信息会向nameServer集群中的每一个节点都上报数据,即心跳包,上报的数据包括broker的基本信息,例如brokerAddr、brokerId、brokerName、clusterName等,以及该broker的topic配置信息,比如topicName名字、perm权限、读写队列数量等等属性,当前上报的数据的时间戳版本Dataversion,以及消费过滤信息集合filterServerList。
- nameServer收到心跳包之后会解析数据并存储在RouteInfoManager的5个map属性中topicQueueTable、brokerAddrTable、clusterAddrTable、brokerLiveTable、filterServerTable。
- 每个nameserver之间不会互相通信,数据不会同步,另外,Nameserver的所有路由数据都存储在内存中,不存在持久化操作,所以nameserver非常的轻量级。
- nameServer没有数据同步、持久化等机制,这可能会造成数据的不一致,但是能够保证服务的高可用,而对于RocketMQ这样的组件来说,可以牺牲一时的数据不一致,但是不能容忍服务的不可,即nameServer保证了CAP中的AP。
版权归原作者 刘Java 所有, 如有侵权,请联系我们删除。