Spring boot 整合RabbitMQ（二）：工作队列（Work queues）

现在，我们将发送一些字符串，把这些字符串当作复杂的任务。我们并没有一个真实的复杂任务，类似于图片大小被调整或 pdf 文件被渲染，所以我们通过 sleep () 方法来模拟这种情况。我们在字符串中加上点号（.）来表示任务的复杂程度，一个点（.）将会耗时 1 秒钟。比如 “Hello…” 就会耗时 3 秒钟。了解springcloud架构可以加求求：三五三六二四七二五九

如果您尚未设置项目，请参阅第一个教程中的设置。我们将遵循与第一个教程相同的模式：创建一个包（tut2）并创建 Tut2Config、Tut2Receiver 和 Tut2Sender。

代码整合
首先创建一个新的包（tut2），我们将在这里放置我们的三个类。在配置类 Tut2Config 中，我们设置了两个配置文件 ——tut2 和 work-queues。我们利用 Spring 来将队列 Queue 暴露为一个 bean。我们配置消费者，并定义两个 bean 以对应于上图中的工作进程 receiver1 和 receiver2。
配置类

@Profile({"tut2", "work-queues"})
@Configuration
public class Tut2Config {

    @Bean
    public Queue queue() {
        return new Queue("work-queues");
    }

    /**
     * 定义两个消费者，并且给了他们不同的标识
     */
  @Profile ("receiver")
    private static class ReceiverConfig {
        @Bean
        public Tut2Receiver receiver1() {
            return new Tut2Receiver(1);
        }

        @Bean
        public Tut2Receiver receiver2() {
            return new Tut2Receiver(2);
        }
    }

   　@Profile("sender")
   　 @Bean
　　 public Tut2Sender sender() {
        return new Tut2Sender();
    }

}

生产者
我们简单修改一下生产者的代码，以添加点号（.）的方式来人为的增加该任务的时长，字符串中的每个点号（.）都会增加 1s 的耗时。

public class Tut2Sender {

 @Autowired
 private AmqpTemplate template;

 @Autowired   
 private Queue queue;

    int dots = 0;
    int count = 0;

   @Scheduled(fixedDelay = 1000, initialDelay = 500)
    public void send(){
        StringBuilder builder = new StringBuilder("Hello");
        if (dots++ == 3) {
            dots = 1;
        }
        for (int i = 0; i < dots; i++) {
            builder.append('.');
        }
        builder.append(Integer.toString(++count));
        String message = builder.toString();
        template.convertAndSend(queue.getName(), message);
        System.out.println(" [x] Sent '" + message + "'");
    }

}

消费者
我们的消费者 Tut2Receiver 通过 doWork () 方法模拟了一个耗时的虚假任务，它需要为消息体中每一个点号（.）模拟 1 秒钟的操作。并且我们为消费者增加了一个实例编号，以知道是哪个实例消费了消息和处理的时长。

@RebbitListener(queues = "work-queues")
public class Tut2Receiver {

    private int instance;

    public Tut2Receiver(int instance) {
        this.instance = instance;
    }

    @RabbitHandler
    public 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);
            }
        }
    }
}

运行
maven 编译

mvn clean package -Dmaven.test.skip=true
运行

java -jar target/rabbitmq-tutorial-0.0.1-SNAPSHOT.jar --spring.profiles.active=tut2,sender --tutorial.client.duration=60000
java -jar target/rabbitmq-tutorial-0.0.1-SNAPSHOT.jar --spring.profiles.active=tut2,receiver --tutorial.client.duration=60000
输出

// Sender
Ready … running for 10000ms
[x] Sent ‘Hello.1’
[x] Sent ‘Hello…2’
[x] Sent ‘Hello…3’
[x] Sent ‘Hello.4’
[x] Sent ‘Hello…5’
[x] Sent ‘Hello…6’
[x] Sent ‘Hello.7’
[x] Sent ‘Hello…8’
[x] Sent ‘Hello…9’

// Receiver
Ready … running for 10000ms
instance 1 [x] Received ‘Hello.1’
instance 2 [x] Received ‘Hello…2’
instance 1 [x] Done in 1.005s
instance 1 [x] Received ‘Hello…3’
instance 2 [x] Done in 2.007s
instance 2 [x] Received ‘Hello.4’
instance 2 [x] Done in 1.005s
instance 1 [x] Done in 3.01s
instance 1 [x] Received ‘Hello…5’
instance 2 [x] Received ‘Hello…6’
instance 1 [x] Done in 2.006s
instance 1 [x] Received ‘Hello.7’
instance 1 [x] Done in 1.002s
instance 1 [x] Received ‘Hello…9’
instance 2 [x] Done in 3.01s
instance 2 [x] Received ‘Hello…8’
prefetch
从消费者这端的输出可以看出来，instance 1 得到的任务编号始终是奇数（Hello.1，Hello…3，Hello…5，Hello.7），而 instance 2 得到的任务编号始终是偶数。了解springcloud架构可以加求求：三五三六二四七二五九
如果感觉这次的输出只是巧合，可以多试几次或通过 --tutorial.client.duration= 调整时长得到更多的输出，而结果肯定都是一样的。

这里设计的问题就是之前在基础概念里讲到的调度策略的问题了。要实现公平调度（Fair dispatch）就是设置 prefetch 的值，实现方式有两种。

全局设置
在 application.yml 中设置 spring.rabbitmq.listener.simple.prefetch=1 即可，这会影响到本 Spring Boot 应用中所有使用默认 SimpleRabbitListenerContainerFactory 的消费者。

网上很多人说改配置 pring.rabbitmq.listener.prefetc，实测已经无效，应该是版本的问题。我所使用的版本（RabbitMQ：3.7.4，Spring Boot: 2.0.1.RELEASE），除了 spring.rabbitmq.listener.simple.prefetch，还有一个 spring.rabbitmq.listener.direct.prefetch 可以配置。

改了配置后再运行，可以看到 instance 1 可以获取到”Hello…6”、”Hello…12” 了。

Ready … running for 60000ms
instance 1 [x] Received ‘Hello.1’
instance 2 [x] Received ‘Hello…2’
instance 1 [x] Done in 1.004s
instance 1 [x] Received ‘Hello…3’
instance 2 [x] Done in 2.008s
instance 2 [x] Received ‘Hello.4’
instance 2 [x] Done in 1.004s
instance 2 [x] Received ‘Hello…5’
instance 1 [x] Done in 3.012s
instance 1 [x] Received ‘Hello…6’
instance 2 [x] Done in 2.007s
instance 2 [x] Received ‘Hello.7’
instance 2 [x] Done in 1.004s
instance 2 [x] Received ‘Hello…8’
instance 1 [x] Done in 3.011s
instance 1 [x] Received ‘Hello…9’
instance 2 [x] Done in 2.007s
instance 2 [x] Received ‘Hello.10’
instance 2 [x] Done in 1.006s
instance 2 [x] Received ‘Hello…11’
instance 1 [x] Done in 3.01s
instance 1 [x] Received ‘Hello…12’
特定消费者
上边是改了全局的消费者，如果只针对特定的消费者的话，又怎么处理呢？
我们可以通过自定义 RabbitListenerContainerFactory 来实现。

@Bean
public RabbitListenerContainerFactory<SimpleMessageListenerContainer> prefetchOneRabbitListenerContainerFactory(ConnectionFactory rabbitConnectionFactory) {
    SimpleRabbitListenerContainerFactory factory = new SimpleRabbitListenerContainerFactory();
    factory.setConnectionFactory(rabbitConnectionFactory);
    factory.setPrefetchCount(1);
    return factory;
}

然后在特定的消费者上指定 containerFactory

@RebbitListener(queues = "hello", containerFactory = "prefetchTenRabbitListenerContainerFactory")
public void receive(String in) {
    System.out.println(" [x] Received '" + in + "'")
}