RocketMQ
一、MQ概述
Message Queue,是一种提供消息队列服务的中间件。提供了消息生产、存储、消费全过程API的软件系统。
MQ的作用
- 限流削峰:当用户发送超量请求时,将请求暂存,以便后期慢慢处理。如果不使用MQ暂存直接请求到业务系统中容易引起系统崩溃。
- 异步解耦:若上游系统和下游系统为同步调用,会大大降低系统的吞吐量和并发量。MQ层实现两个系统之间的异步调用
- 数据收集:分布式系统会产生海量数据流,如业务日志、监控数据、用户行为。针对这些数据流采集汇总,进行大数据分析。
主流应用的MQ产品
- Kafka:Scala/Java语言开发。特点是高吞吐量,但会丢数据,常用与大数据领域的实时计算、日志采集等场景。不遵循任何MQ协议,使用自研协议。
- RocketMQ:Java语言开发。经过数年阿里双十一考验,性能与稳定性非常高,功能全面。不遵循任何MQ协议,使用自研协议。开源版不如云上版(阿里商业版)
MQ常见协议
-
JMS:Java Messaging Service。Java平台上有关MOM(Message Orientated Middleware)的技术规范。他便于Java应用程序的消息交换,提供标准的接口简化开发。ActiveMQ时典型实现
-
STOMP:Streaming Text Orientated Message Protocol。是一种MOM的简单文本协议。STOMP提供一个可互操作的连接格式,允许 客户端与任意STOMP消息代理进行交互。ActiveMQ时典型实现
-
AMQP:Advanced Message Queuing Protocol。一个提供统一消息服务的应用层标准,是应用层协议的一个开放标准。RabbitMQ是典型实现
-
MQTT:Message Queueing Telemetry Transport。IBM开发的一个即时通讯协议(二进制协议),主要用于服务器和低功耗IoT设备之间的通信
二、基本概念
主题(Topic):表示一类消息的集合(可以理解为消息的类型),每个消息只能属于一个主题,是RocketMQ进行消息订阅的基本单位。一个生产者可以同时发送多种Topic消息,而一个消费者只能接收一种Topic消息
标签(Tag):用于快速过滤消息
三、Linux部署RocketMQ服务
1、在官网下载编译好的二进制压缩包,版本5.0.0即可,上传到Linux中
2、进行解压
3、配置环境变量ROCKETMQ_HOME和NAMESRV_ADDR
4、配置bin目录下的runserver.sh,根据实际情况修改JVM的内存参数
5、配置bin目录下的runbroker.sh,根据实际情况修改JVM的内存参数
6、执行nohup命令后台运行RocketMQ服务(nameserver必须先启动,broker需要再nameserver上注册)
# 启动nameserver
nohup bin/mqnamesrv &
# 启动broker
nohup bin/mqbroker -c [confFile] & # -c可指定加载的配置文件,默认为conf/broker.conf
# 查看日志rocketmq是否成功启动
tail nohup.out
# 查看进程
jps
# 停止broker
sh bin/mqshutdown broker
# 停止namesrv
sh bin/mqshutdown namesrv
7、执行命令测试(rocketmq提供的测试样例,生产者会发送一千条消息)
bin/tools.sh org.apache.rocketmq.example.quickstart.Producer
8、执行命令测试(rocketmq提供的测试样例,消费者会接受一千条消息)
bin/tools.sh org.apache.rocketmq.example.quickstart.Consumer
四、RocketMQ API
生产者同步发送消息
public void test_SyncProducer() throws MQClientException {
DefaultMQProducer producer = new DefaultMQProducer("producer_group_name");
//设置注册服务的ip地址的端口
producer.setNamesrvAddr(RocketMQConstant.NAME_SRV_ADDR);
//启动生产者
producer.start();
for(int i=0; i<3; i++){
try {
// 封装消息,设置topic,tag(用于消息快速过滤),消息数据
Message message = new Message(
"TopicTest",
"TagA",
"ID04287777",
("Hello, RocketMQ " + i).getBytes(RemotingHelper.DEFAULT_CHARSET));
//同步发送消息并获取发送结果,producer从broker获取发送结果
SendResult sendResult = producer.send(message);
System.out.println(sendResult);
Thread.sleep(1500);
} catch (Exception e) {
throw new RuntimeException(e);
}
}
producer.shutdown();
}
生产者异步发送消息
public void test_AsyncProducer() throws Exception{
DefaultMQProducer producer = new DefaultMQProducer(RocketMQConstant.PRODUCER_GROUP_NAME);
producer.setNamesrvAddr(RocketMQConstant.NAME_SRV_ADDR);
producer.start();
producer.setRetryTimesWhenSendAsyncFailed(0);
int messageCount = 10;
final CountDownLatch countDownLatch = new CountDownLatch(messageCount);
for(int i=0; i<messageCount; i++){
final int index = i;
// 封装消息,设置topic,tag(用于消息快速过滤),消息数据
Message message = new Message(
"TopicTest",
"TagA",
"ID04287777",
("Hello, RocketMQ " + i).getBytes(RemotingHelper.DEFAULT_CHARSET));
// 异步发送消息,若broker有响应会调用SendCallback中的方法
producer.send(message, new SendCallback() {
public void onSuccess(SendResult sendResult) {
countDownLatch.countDown();
System.out.println(" Send Message "+ index +" OK: "+sendResult);
}
public void onException(Throwable throwable) {
countDownLatch.countDown();
System.out.println(" Send Message "+ index +" Exception: "+throwable);
}
});
//单向发送
producer.sendOneway(message);
System.out.println("Message "+index+" send done");
}
//在100条消息发送完后关闭
countDownLatch.await(5, TimeUnit.SECONDS);
producer.shutdown();
}
生产者单向发送消息
public void test_OneWayProducer() throws Exception{
DefaultMQProducer producer = new DefaultMQProducer(RocketMQConstant.PRODUCER_GROUP_NAME);
producer.setNamesrvAddr(RocketMQConstant.NAME_SRV_ADDR);
producer.start();
producer.setRetryTimesWhenSendAsyncFailed(0);
int messageCount = 10;
final CountDownLatch countDownLatch = new CountDownLatch(messageCount);
for(int i=0; i<messageCount; i++){
final int index = i;
// 封装消息,设置topic,tag(用于消息快速过滤),消息数据
Message message = new Message(
"TopicTest",
"TagA",
"ID04287777",
("Hello, RocketMQ " + i).getBytes(RemotingHelper.DEFAULT_CHARSET));
//单向发送
producer.sendOneway(message);
System.out.println("Message "+index+" send done");
}
//在100条消息发送完后关闭
countDownLatch.await(5, TimeUnit.SECONDS);
producer.shutdown();
}
消费者推模式
public static void test_PushConsumer() throws Exception{
DefaultMQPushConsumer consumer = new DefaultMQPushConsumer("consumer_group_name");
consumer.setNamesrvAddr(RocketMQConstant.NAME_SRV_ADDR);
consumer.setConsumeFromWhere(ConsumeFromWhere.CONSUME_FROM_LAST_OFFSET);
//消费者订阅的消息topic和tag(subExpression,*表示任意)
consumer.subscribe("TopicTest", "*");
consumer.registerMessageListener(new MessageListenerConcurrently() {
public ConsumeConcurrentlyStatus consumeMessage(List<MessageExt> list, ConsumeConcurrentlyContext consumeConcurrentlyContext) {
System.out.println("Receive New Message : "+list);
return ConsumeConcurrentlyStatus.CONSUME_SUCCESS;
}
});
consumer.start();
System.out.println("Consumer Start...");
}
消费者拉模式
不同于推模式消费者,拉模式下需要手动管理消息队列MessageQueue和偏移量offset的映射关系。但是最新的LitePullConsumer底层源码已经实现对mq和offset的管理,比较方便。
//拉模式消费者
public static void test_LitePullConsumer() throws Exception{
DefaultLitePullConsumer litePullConsumer = new DefaultLitePullConsumer(RocketMQConstant.CONSUMER_GROUP_NAME);
litePullConsumer.setNamesrvAddr(RocketMQConstant.NAME_SRV_ADDR);
litePullConsumer.setConsumeFromWhere(ConsumeFromWhere.CONSUME_FROM_LAST_OFFSET);
litePullConsumer.subscribe("TopicTest", "*");
litePullConsumer.start();
try {
while(true){
List<MessageExt> messageExts = litePullConsumer.poll();
System.out.printf("%s%n", messageExts);
}
}finally {
litePullConsumer.shutdown();
}
}
RocketMQ传递对象,对象所属类需要实现序列化接口,并且将对象转换为字节数组存入消息体中。
顺序消息
保证消息的局部有序(其中几条消息的有序,不一定是全部消息都要有序),以防止受到网络传输的影响。
实现原理
生产者将一组有序的消息一次发到同一个MessageQueue中(依靠队列的特点保证局部有序性)。消费者消费完一个MessageQueue的消息后才会去消费下一个MessageQueue的消息。
public class OrderProducer {
public static void main(String[] args) {
DefaultMQProducer producer = new DefaultMQProducer(WanfengConstant.PRODUCER_GROUP_NAME);
try {
producer.setNamesrvAddr(WanfengConstant.NAMESRV_ADDR);
producer.start();
for(int i=0; i<5; i++){
//用于指定顺序的id
int orderId = i;
for(int j=0; j<5; j++){
Message message = new Message(
WanfengConstant.ORDER_TOPIC,
"order_"+orderId,
"KEY"+orderId,
("order_"+orderId+" step "+j).getBytes(RemotingHelper.DEFAULT_CHARSET)
);
//实现消息队列选择器对象,使同一个orderId的消息发送到同一个消息队列
SendResult sendResult = producer.send(
message,
new MessageQueueSelector() {
@Override
public MessageQueue select(List<MessageQueue> mqs, Message msg, Object arg) {
Integer id = (Integer) arg;
int index = id % mqs.size();
return mqs.get(index);
}
},
orderId
);
System.out.printf("%s%n", sendResult);
}
}
}catch(Exception e){
e.printStackTrace();
producer.shutdown();
}
}
}
public class OrderConsumer {
public static void main(String[] args) {
DefaultMQPushConsumer consumer = new DefaultMQPushConsumer(WanfengConstant.CONSUMER_GROUP_NAME);
consumer.setNamesrvAddr(WanfengConstant.NAMESRV_ADDR);
consumer.setConsumeFromWhere(ConsumeFromWhere.CONSUME_FROM_LAST_OFFSET);
try {
consumer.subscribe(WanfengConstant.ORDER_TOPIC, "*");
//实现顺序消息监听者接口
consumer.registerMessageListener(new MessageListenerOrderly() {
@Override
public ConsumeOrderlyStatus consumeMessage(List<MessageExt> msgs, ConsumeOrderlyContext context) {
context.setAutoCommit(true);
for(MessageExt messageExt : msgs){
System.out.println("Receive Message: " + new String(messageExt.getBody()));
}
return ConsumeOrderlyStatus.SUCCESS;
}
});
consumer.start();
System.out.println("Consumer Start...");
} catch (Exception e) {
e.printStackTrace();
consumer.shutdown();
}
}
}
广播消息
生产者发送的消息推送给所有group的消费者
实现原理:将消费者设置MessageModel为广播模式。
public class BroadcastConsumer {
public static void main(String[] args) {
DefaultMQPushConsumer consumer = new DefaultMQPushConsumer(WanfengConstant.CONSUMER_GROUP_NAME);
consumer.setConsumeFromWhere(ConsumeFromWhere.CONSUME_FROM_LAST_OFFSET);
//设定消息模式为广播
consumer.setMessageModel(MessageModel.BROADCASTING);
try {
consumer.subscribe(WanfengConstant.ARCHIVE_TOPIC, "*");
consumer.registerMessageListener(new MessageListenerConcurrently() {
@Override
public ConsumeConcurrentlyStatus consumeMessage(List<MessageExt> msgs, ConsumeConcurrentlyContext context) {
msgs.forEach(messageExt -> {
Archive archive = (Archive) WanfengObjectUtil.bytesToObject(messageExt.getBody());
System.out.println("Receive Message : "+archive.getId());
});
return ConsumeConcurrentlyStatus.CONSUME_SUCCESS;
}
});
consumer.start();
System.out.println("Broadcast Consumer Start...");
}catch (Exception e){
e.printStackTrace();
consumer.shutdown();
}
}
}
若指定MessageModel为CLUSTERING,则生产者发送的消息会随机指定消费者消费。
延迟消息
顾名思义就是消息发送到broker时延迟指定的时间后再发送给消费者。常用于定时发送
过滤消息
过滤消息通过tag实现,在消费者端指定过滤的tag即可。
//消费者订阅tag1或tag2的消息
consumer.subscribe("TopicTest", "tag1 || tag2");
在RocketMQ中,消费者指定过滤条件后,将其上推到Broker中,在Broker中进行tag过滤,以减少网络IO,但同时也增加了Broker的繁忙。
事务消息
public class TransactionProducer {
public static void main(String[] args) {
TransactionMQProducer producer = new TransactionMQProducer(WanfengConstant.PRODUCER_GROUP_NAME);
TransactionListener transactionListener = new TransactionListener() {
@Override
public LocalTransactionState executeLocalTransaction(Message msg, Object arg) {
System.out.println("[WANFENG-INFO] TransactionProducer.executeLocalTransaction(): 执行成功...");
String tags = msg.getTags();
if (StringUtils.contains(tags, "TagA")) {
//消息提交(发送出去)
return LocalTransactionState.COMMIT_MESSAGE;
} else if (StringUtils.contains(tags, "TagB")) {
//消息回滚(丢掉消息)
return LocalTransactionState.ROLLBACK_MESSAGE;
} else {
return LocalTransactionState.UNKNOW;
}
}
@Override
public LocalTransactionState checkLocalTransaction(MessageExt msg) {
System.out.println("[WANFENG-INFO] TransactionProducer.checkLocalTransaction(): 执行成功...");
String tags = msg.getTags();
if (StringUtils.contains(tags, "TagC")) {
return LocalTransactionState.COMMIT_MESSAGE;
} else {
return LocalTransactionState.UNKNOW;
}
}
};
ExecutorService executorService = new ThreadPoolExecutor(
2,
5,
100, TimeUnit.SECONDS,
new ArrayBlockingQueue<>(3)
);
producer.setExecutorService(executorService);
producer.setTransactionListener(transactionListener);
try {
producer.start();
} catch (Exception e) {
e.printStackTrace();
}
String[] tags = new String[]{"TagA", "TagB", "TagC"};
CountDownLatch countDownLatch = new CountDownLatch(9);
for (int i = 0; i < 9; i++) {
try {
Message message = new Message("TopicTest", tags[i % tags.length], "Key" + i, ("Hello RocketMQ " + i).getBytes(RemotingHelper.DEFAULT_CHARSET));
SendResult sendResult = producer.sendMessageInTransaction(message, null);
System.out.println(sendResult);
Thread.sleep(1000);
countDownLatch.countDown();
} catch (Exception e) {
e.printStackTrace();
}
}
try {
countDownLatch.await();
} catch (InterruptedException e) {
throw new RuntimeException(e);
} finally {
try {
Thread.sleep(100000);
} catch (InterruptedException e) {
throw new RuntimeException(e);
}
producer.shutdown();
}
}
}
ACL权限控制
ACL对用户对Topic资源的访问权限进行控制
在pom依赖中引入acl的依赖包
<dependency>
<groupId>org.apache.rocketmq</groupId>
<artifactId>rocketmq-acl</artifactId>
<version>5.0.0</version>
</dependency>
在服务端的conf/broker.conf文件,添加配置,开启acl
aclEnable=true
在服务端的conf/plain_acl.yml文件,配置具体权限规则(热加载,不需要重启mq)
accounts:
- accessKey: RocketMQ #用户名
secretKey: 12345678 #密码
whiteRemoteAddress: #访问地址白名单
admin: false #是否为管理员(管理员可以访问所有Topic)
defaultTopicPerm: DENY #默认Topic访问权限
defaultGroupPerm: SUB #默认组权限
topicPerms: #Topic对应的权限,若这里找不到则采用defaultTopicPerm
- topicA=DENY
- topicB=PUB|SUB
- topicC=SUB
groupPerms:
# the group should convert to retry topic
- groupA=DENY
- groupB=PUB|SUB
- groupC=SUB
在创建生产者对象时需加入RPCHook(acl的用户信息)
public class AclProducer {
private static final String ACL_ACCESS_KEY = "RocketMQ";
private static final String ACL_SECRET_KEY = "12345678";
/**
* 通过用户名和密码获取RPCHook
* @return
*/
public static RPCHook getAclRPCHook(){
return new AclClientRPCHook(new SessionCredentials(ACL_ACCESS_KEY, ACL_SECRET_KEY));
}
public static void main(String[] args) throws MQClientException, InterruptedException {
//创建生产者时加入用户信息,即RPCHook
DefaultMQProducer producer = new DefaultMQProducer(WanfengConstant.PRODUCER_GROUP_NAME, getAclRPCHook());
producer.setNamesrvAddr(WanfengConstant.NAMESRV_ADDR);
producer.start();
for (int i = 0; i < 20; i++) {
try {
Message message = new Message(
"TopicTest",
WanfengConstant.TAGS_NAME,
("Hello RocketMQ " + i).getBytes(RemotingHelper.DEFAULT_CHARSET) /*消息体转换成二进制数组*/
);
SendResult sendResult = producer.send(message);
System.out.printf("%s%n", sendResult);
} catch (Exception e) {
e.printStackTrace();
Thread.sleep(1000);
}
}
}
}
消息轨迹
Producer,Consumer,Broker处理消息的相关信息
消息轨迹的实现原理是MQ把消息轨迹都往RMQ_SYS_TRACE_TOPIC的Topic中放
在Broker端配置文件开启消息轨迹
traceTopicEnable=true
创建生产者时指定enableMsgTrace参数为true,开启消息轨迹。也可以指定customizedTraceTopic参数来自定义消息轨迹的Topic。
public class TraceProducer {
public static void main(String[] args) throws MQClientException {
//指定enableMsgTrace参数为true,开启消息轨迹
DefaultMQProducer producer = new DefaultMQProducer(WanfengConstant.PRODUCER_GROUP_NAME, true);
producer.setNamesrvAddr(WanfengConstant.NAMESRV_ADDR);
producer.start();
for (int i = 0; i < 20; i++) {
try {
Message message = new Message(
"TopicTest",
WanfengConstant.TAGS_NAME,
("Hello RocketMQ " + i).getBytes(RemotingHelper.DEFAULT_CHARSET) /*消息体转换成二进制数组*/
);
SendResult sendResult = producer.send(message);
System.out.printf("%s%n", sendResult);
} catch (Exception e) {
e.printStackTrace();
}
}
}
}