RabbitMQ(一) - 基本结构、SpringBoot整合RabbitMQ、工作队列、发布订阅、直接、主题交换机模式

RabbitMQ结构

• Publisher ： 生产者

• Queue: 存储消息的容器队列；

• Consumer:消费者

• Connection：消费者与消息服务的TCP连接

• Channel:信道，是TCP里面的虚拟连接。例如：电缆相当于TCP，信道是一条独立光纤束，一条TCP连接上创建多少条信道是没有限制的。TCP一旦打开，就会出AMQP信道。无论是发布消息，接收消息，订阅队列，这些动作都是通过信道完成的。
  Broker: 一台消息服务就是一个Broker;

• Exchange:交换机、负责接收生产者的消息，转发到队列中、交换机和队列通过路由键绑定、可以理解为每个队列都有自己的名称；
  

SpringBoot整合RabbitMQ

• Queue
  • 消息存放于队列中， 若是RabbitMQ挂了，则消息会丢失，因此要开启持久化， 将durable设置为true,
  • 若是没有消费者消费该队列，则该队列会自动删除， 因此需要将autoDelete参数设置为false;

    public Queue(String name) {//  队列名称， 是否持久化，是否独占， 是否自动删除this(name, true, false, false);}

• @RabbitListener

@RabbitListener(bindings=@QueueBinding(value= @Queue(value="${mq.config.queue.info}",autoDelete="true"),exchange=@Exchange(value="${mq.config.exchange}",type=ExchangeTypes.DIRECT),key="${mq.config.queue.info.routing.key}"))

用来标记消费者；exchange表示交换器信息、类型；bindings表示监听器要绑定的队列、以及队列信息；
key：代表交换机和队列通过key绑定的；

• AmqpTemplate / RabbitTempldate:
  生产者通过依赖此工具类发送消息；

先安装RabbitMQ，创建SpringBoot项目，修改配置

# 应用名称
spring.application.name=boolfilter# 应用服务 WEB 访问端口
server.port=8080spring.rabbitmq.host=127.0.0.1
spring.rabbitmq.port=5672
spring.rabbitmq.username=guest
spring.rabbitmq.password=guest

入门级别程序

发送hello world程序；

生产者：

public class Tut1Sender {@Autowiredprivate RabbitTemplate template;@Autowiredprivate Queue queue;@Scheduled(fixedDelay = 1000, initialDelay = 500)public void send() {String message = "Hello World!";this.template.convertAndSend(queue.getName(), message);System.out.println(" [x] Sent '" + message + "'");}
}

消费者：

@RabbitListener(queues = "hello")
public class Tut1Receiver {@RabbitHandlerpublic void receive(String in) {System.out.println(" [x] Received '" + in + "'");}
}

将生产者、消费者注入容器；

@Configuration
@EnableScheduling
public class Tut1Config {@Beanpublic Queue hello() {return new Queue("hello");}@Beanpublic Tut1Receiver receiver() {return new Tut1Receiver();}@Beanpublic Tut1Sender sender() {return new Tut1Sender();}
}

运行结果：

[x] Sent ‘Hello World!’
[x] Received ‘Hello World!’
[x] Sent ‘Hello World!’
[x] Received ‘Hello World!’
[x] Sent ‘Hello World!’
…

工作队列

主要思想是避免 立即执行资源密集型任务，必须等待它要完成。相反，我们将任务安排在以后完成。我们将任务封装为消息并将其发送到队列。正在运行的工作进程 在后台将弹出任务并最终执行工作

生产者：

public class Tut2Sender {@Autowiredprivate RabbitTemplate template;@Autowiredprivate Queue queue;AtomicInteger dots = new AtomicInteger(0);AtomicInteger count = new AtomicInteger(0);@Scheduled(fixedDelay = 1000, initialDelay = 500)public void send() {`在这里插入代码片`StringBuilder builder = new StringBuilder("Hello");if (dots.incrementAndGet() == 4) {dots.set(1);}for (int i = 0; i < dots.get(); i++) {builder.append('.');}builder.append(count.incrementAndGet());String message = builder.toString();template.convertAndSend(queue.getName(), message);System.out.println(" [x] Sent '" + message + "'");}}

消费者：

@RabbitListener(queues = "hello")
public class Tut2Receiver {private final int instance;public Tut2Receiver(int i) {this.instance = i;}@RabbitHandlerpublic void receive(String in) throws InterruptedException {StopWatch watch = new StopWatch();watch.start();System.out.println("instance " + this.instance +" [x] Received '" + in + "'");doWork(in);watch.stop();System.out.println("instance " + this.instance +" [x] Done in " + watch.getTotalTimeSeconds() + "s");}private void doWork(String in) throws InterruptedException {for (char ch : in.toCharArray()) {if (ch == '.') {Thread.sleep(1000);}}}
}

队列、生产者、消费者注入容器：

@Configuration
public class Tut2Config {@Beanpublic Queue hello() {return new Queue("hello");}private static class ReceiverConfig {@Beanpublic Tut2Receiver receiver1() {return new Tut2Receiver(1);}@Beanpublic Tut2Receiver receiver2() {return new Tut2Receiver(2);}}@Beanpublic Tut2Sender sender() {return new Tut2Sender();}
}

运行结果：

[x] Sent ‘Hello.1’
instance 1 [x] Received ‘Hello.1’
[x] Sent ‘Hello…2’
instance 2 [x] Received ‘Hello…2’
instance 1 [x] Done in 1.0062309s
[x] Sent ‘Hello…3’
instance 1 [x] Received ‘Hello…3’
instance 2 [x] Done in 2.0085791s
[x] Sent ‘Hello.4’
instance 2 [x] Received ‘Hello.4’
…

消息确认

• SpringBoot整合RabbitMQ代码中，若消费者消费出现异常，则会重新进入队列， 一般生产环境中，是要有重试机制的；
• 若是要关闭重试机制、则设置defaultRequeueRejected=false， 或者抛出AmqpRejectAndDontRequeueException异常，这样框架会帮我们自动提交确认channel.basicAck()
• 重试机制也会存在问题、若是消费者服务关闭、则消息会不断重新入队、导致RabbitMQ内存最终爆满宕机；
• 消息的ACK确认机制默认是打开的；如果忘记了ACK，后果很严重，当Consumer退出时，消息会一直重新分发，然后RabbitMq会占用越来越多的内存，由于RabbitMq会长时间运行，出现“内存泄露”是致命的

异常处理方案：

• 使用try-catch捕捉
• 使用重试机制、超过一定次数、则丢弃消息或放入死信队列；

spring.rabbitmq.listener.retry.max-attempts=5 //重试超过5次，消息丢弃；

公平调度与循环调度

• 默认情况下，RabbitMQ 会将每条消息发送给下一个消费者。平均而言，每个消费者将获得相同数量的 消息。这种分发消息的方式称为轮询。 在这种模式下，调度不一定完全按照我们想要的方式工作。 若是存在两台机器，一台性能好、一台性能差， 而RabbitMQ对此一无所知，仍然会调度 消息均匀。发生这种情况是因为 RabbitMQ 只是在消息时调度消息 进入队列。它不看未确认的数量 面向消费者的消息。它只是盲目地发送每 n 条消息 给第 n 个消费者，这就导致了一台机器特别忙碌、一台机器空闲；

• “公平调度”是Spring AMQP的默认配置。Consumer可以向服务器声明一个prefetchCount, 表示轮到自己时、自己可处理多少消息；这样RabbitMQ转发消息给消费者时、会先看Consumer正在处理的消息数量是否达到了prefetchCount, 若已达到该值，则发给其他的Consumer;

发布/订阅

特点：一条消息同时会被所有消费者消息；X是交换机(Exchange)；交换机和队列进行绑定(Binding)
交换机负责接收生产者发送的消息，再转发消息到队列中；实现了生产者与队列的解耦；

RabbitMQ 中消息传递模型的核心思想是生产者 从不将任何消息直接发送到队列

示例1 ： 广播匿名队列

发送者：

public class Tut3Sender {@Autowiredprivate RabbitTemplate template;@Autowiredprivate FanoutExchange fanout;AtomicInteger dots = new AtomicInteger(0);AtomicInteger count = new AtomicInteger(0);@Scheduled(fixedDelay = 1000, initialDelay = 500)public void send() {StringBuilder builder = new StringBuilder("Hello");if (dots.getAndIncrement() == 3) {dots.set(1);}for (int i = 0; i < dots.get(); i++) {builder.append('.');}builder.append(count.incrementAndGet());String message = builder.toString();template.convertAndSend(fanout.getName(), "", message);System.out.println(" [x] Sent '" + message + "'");}}

消费者：

public class Tut3Receiver {@RabbitListener(queues = "#{autoDeleteQueue1.name}")public void receive1(String in) throws InterruptedException {receive(in, 1);}@RabbitListener(queues = "#{autoDeleteQueue2.name}")public void receive2(String in) throws InterruptedException {receive(in, 2);}public void receive(String in, int receiver) throws InterruptedException {StopWatch watch = new StopWatch();watch.start();System.out.println("instance " + receiver + " [x] Received '" + in + "'");doWork(in);watch.stop();System.out.println("instance " + receiver + " [x] Done in "+ watch.getTotalTimeSeconds() + "s");}private void doWork(String in) throws InterruptedException {for (char ch : in.toCharArray()) {if (ch == '.') {Thread.sleep(1000);}}}}

交换机、匿名队列、绑定，生产者、消费者注入容器；

public class Tut3Config {@Beanpublic FanoutExchange fanout() {return new FanoutExchange("tut.fanout");}private static class ReceiverConfig {@Beanpublic Queue autoDeleteQueue1() {return new AnonymousQueue();}@Beanpublic Queue autoDeleteQueue2() {return new AnonymousQueue();}@Beanpublic Binding binding1(FanoutExchange fanout,Queue autoDeleteQueue1) {return BindingBuilder.bind(autoDeleteQueue1).to(fanout);}@Beanpublic Binding binding2(FanoutExchange fanout,Queue autoDeleteQueue2) {return BindingBuilder.bind(autoDeleteQueue2).to(fanout);}@Beanpublic Tut3Receiver receiver() {return new Tut3Receiver();}}@Beanpublic Tut3Sender sender() {return new Tut3Sender();}
}

运行结果：

instance 1 [x] Received 'Hello.1'
instance 2 [x] Received 'Hello.1'
instance 2 [x] Done in 1.0057994s
instance 1 [x] Done in 1.0058073s
....

模拟Spring容器发布ContextRefreshedEvent事件

通常情况下，业务开发中，经常会监听该事件做扩展，例如初始化数据， 打印日志等等；
生产者：

public class AppContextSender {@AutowiredRabbitTemplate rabbitTemplate;@Scheduled(fixedDelay = 1000, initialDelay = 500)public void publishContextRefreshEvent() {rabbitTemplate.convertAndSend("contextRefreshedExchange", "", "publish refreshed event");}
}

消费者：

@RabbitListener(queues = {"initQueue"})
public class InitContextRefreshedConsumer {@RabbitHandlerpublic void consum(String in) {System.out.println("init :"+in);}
}@RabbitListener(queues = "logQueue")
public class LogContextRefreshedConsumer {@RabbitHandlerpublic void consum(String in) {System.out.println("log : "+in);}
}

交换机、队列、绑定、生产者、消费者注入容器：

@Configuration
public class ContextRefreshedConfig {@Beanpublic FanoutExchange contextRefreshedExchange(){return new FanoutExchange("contextRefreshedExchange");}@Beanpublic AppContextSender appContextSender() {return new AppContextSender();}public static class ConsumerConfig {@Beanpublic Queue initQueue() {return new Queue("initQueue");}@Beanpublic Queue logQueue() {return new Queue("logQueue");}@Beanpublic Binding initBinding(Queue initQueue, FanoutExchange contextRefreshedExchange) {return BindingBuilder.bind(initQueue).to(contextRefreshedExchange);}@Beanpublic Binding logBinding(Queue logQueue, FanoutExchange contextRefreshedExchange) {return BindingBuilder.bind(logQueue).to(contextRefreshedExchange);}@Beanpublic InitContextRefreshedConsumer initContextRefreshedConsumer() {return new InitContextRefreshedConsumer();}@Beanpublic LogContextRefreshedConsumer logContextRefreshedConsumer() {return new LogContextRefreshedConsumer();}}}

log : publish refreshed event
init :publish refreshed event
log : publish refreshed event
init :publish refreshed event
…

Direct直接模式

• 交换器绑定多个队列，每个绑定关系有自己的路由键；
• 之前业务开发中、有一个交换机、绑定了两个队列，一个队列用来发送邮件，一个队列用来发送短信， 像广播模式下，如果只想发邮件，则没法t做到，使用direct模式和工作模式则可以做到， 最后使用了direct

生产者：

public class BaseServiceSender {@Autowiredprivate RabbitTemplate template;@Autowiredprivate DirectExchange messageExchange;AtomicInteger index = new AtomicInteger(0);AtomicInteger count = new AtomicInteger(0);private final String[] keys = {"sms", "mail"};@Scheduled(fixedDelay = 1000, initialDelay = 500)public void send() {//短信String sms = "{userName: xxx; phone:xxx}";template.convertAndSend(messageExchange.getName(), "sms", sms);//邮件String mail = "{userName: xxx; mail:xxx}";template.convertAndSend(messageExchange.getName(), "mail", mail);}
}

消费者：

@RabbitListener(queues = "mailQueue")
public class MailConsumer {@RabbitHandlerpublic void consum(String in) {System.out.println("send mail : " + in);}
}@RabbitListener(queues = "smsQueue")
public class SmsConsumer {@RabbitHandlerpublic void consum(String in) {System.out.println("send sms : " + in);}
}

交换机、队列，绑定、消费者，生产者注入容器：

@Configuration
public class DirectConfig {@Beanpublic DirectExchange messageExchange() {return new DirectExchange("messageExchange");}@Beanpublic BaseServiceSender baseServiceSender() {return new BaseServiceSender();}public static class ConsumerGroup {@Beanpublic MailConsumer mailConsumer() {return new MailConsumer();}@Beanpublic SmsConsumer smsConsumer() {return new SmsConsumer();}@Beanpublic Queue mailQueue() {return new Queue("mailQueue");}@Beanpublic Queue smsQueue() {return new Queue("smsQueue");}@Beanpublic Binding smsBinding(DirectExchange messageExchange, Queue smsQueue){return BindingBuilder.bind(smsQueue).to(messageExchange).with("sms");}@Beanpublic Binding mailBinding(DirectExchange messageExchange, Queue mailQueue){return BindingBuilder.bind(mailQueue).to(messageExchange).with("mail");}}
}

运行结果

send mail : {userName: xxx; mail:xxx}
send sms : {userName: xxx; phone:xxx}
send sms : {userName: xxx; phone:xxx}
send mail : {userName: xxx; mail:xxx}
…

Topic主题模式

• 发送到主题交换的消息不能有任意routing_key
• • 它必须是单词列表，由点分隔。这 单词可以是任何东西，一些有效的路由密钥示例： “stock.usd.nyse”， “nyse.vmw”， “quick.orange.rabbit”。可以有 路由密钥中随心所欲地包含多个单词，最多可达 255 个 字节。
• 绑定密钥也必须采用相同的形式。主题交换背后的逻辑类似于直接交换 - 发送的消息带有 特定的路由键将被传递到所有队列 绑定匹配的绑定键
• *（星号）可以代替一个词。
• #（哈希）可以替换零个或多个单词。

若是消息指定的路由键为"xxx.orange.xxx"， 则会匹配到Q1, 若是"lazy.xxx.xx"则是Q2;

生产者：

public class Tut5Sender {@Autowiredprivate RabbitTemplate template;@Autowiredprivate TopicExchange topic;AtomicInteger index = new AtomicInteger(0);AtomicInteger count = new AtomicInteger(0);private final String[] keys = {"quick.orange.rabbit", "lazy.orange.elephant", "quick.orange.fox","lazy.brown.fox", "lazy.pink.rabbit", "quick.brown.fox"};@Scheduled(fixedDelay = 1000, initialDelay = 500)public void send() {StringBuilder builder = new StringBuilder("Hello to ");if (this.index.incrementAndGet() == keys.length) {this.index.set(0);}String key = keys[this.index.get()];builder.append(key).append(' ');builder.append(this.count.incrementAndGet());String message = builder.toString();template.convertAndSend(topic.getName(), key, message);System.out.println(" [x] Sent '" + message + "'");}}

消费者：

public class Tut5Receiver {@RabbitListener(queues = "#{autoDeleteQueue1.name}")public void receive1(String in) throws InterruptedException {receive(in, 1);}@RabbitListener(queues = "#{autoDeleteQueue2.name}")public void receive2(String in) throws InterruptedException {receive(in, 2);}public void receive(String in, int receiver) throwsInterruptedException {StopWatch watch = new StopWatch();watch.start();System.out.println("instance " + receiver + " [x] Received '"+ in + "'");doWork(in);watch.stop();System.out.println("instance " + receiver + " [x] Done in "+ watch.getTotalTimeSeconds() + "s");}private void doWork(String in) throws InterruptedException {for (char ch : in.toCharArray()) {if (ch == '.') {Thread.sleep(1000);}}}
}

交换器，队列，绑定、生产者，消费者注入容器：

@Configuration
public class Tut5Config {@Beanpublic TopicExchange topic() {return new TopicExchange("tut.topic");}private static class ReceiverConfig {@Beanpublic Tut5Receiver receiver() {return new Tut5Receiver();}@Beanpublic Queue autoDeleteQueue1() {return new AnonymousQueue();}@Beanpublic Queue autoDeleteQueue2() {return new AnonymousQueue();}@Beanpublic Binding binding1a(TopicExchange topic,Queue autoDeleteQueue1) {return BindingBuilder.bind(autoDeleteQueue1).to(topic).with("*.orange.*");}@Beanpublic Binding binding1b(TopicExchange topic,Queue autoDeleteQueue1) {return BindingBuilder.bind(autoDeleteQueue1).to(topic).with("*.*.rabbit");}@Beanpublic Binding binding2a(TopicExchange topic,Queue autoDeleteQueue2) {return BindingBuilder.bind(autoDeleteQueue2).to(topic).with("lazy.#");}}@Beanpublic Tut5Sender sender() {return new Tut5Sender();}}

运行结果：

[x] Sent ‘Hello to lazy.orange.elephant 1’
instance 2 [x] Received ‘Hello to lazy.orange.elephant 1’
instance 1 [x] Received ‘Hello to lazy.orange.elephant 1’
[x] Sent ‘Hello to quick.orange.fox 2’
[x] Sent ‘Hello to lazy.brown.fox 3’
instance 1 [x] Done in 2.0110456s
…

RPC远程过程调用

RabbitMQ也实现了RPC的功能，但是业务开发中，根本没有使用场景，RPC要么使用Dubbo, 要么使用OpenFeign, 使用RabbitMQ做RPC的信息，目前都没有看到；

总结

• 就目前来说、工作队列、发布订阅两个模式，业务开发中会使用到，其他的消息场景很少见。
• 底层是基于RabbitMQ-client做的封装出RabbitTempldate使用；除非远古项目，否则不推荐使用RabbitMQ-Client原生API写，太费时间了。我写了一会就放弃了