RabbitMQ：延时队列

RabbitMQ：延时队列

（一）：在队列上设置TTL 

参考文献：RabbitMQ学习之延时队列

 其中，【延时队列delay.1m.queue】配置参数如下需配置：

【生产者配置】：

package com.caox.rabbitmq.demo._15_rabbitmq_delay_query_ttl;

import com.rabbitmq.client.Channel;
import com.rabbitmq.client.Connection;
import com.rabbitmq.client.ConnectionFactory;


/**
 * Created by nazi on 2018/8/1.
 * 延时队列 队列配置延时
 */


public class Producer {

    private static String queue_name = "test.queue";

    public static void main(String[] args) throws Exception {
        ConnectionFactory factory = new ConnectionFactory();
        factory.setHost("127.0.0.1");
        factory.setUsername("caoxia");
        factory.setPassword("caoxia123456");
        Connection connection = factory.newConnection();
        Channel channel = connection.createChannel();
        // 声明队列
        channel.queueDeclare(queue_name, true, false, false, null);
        String message = "hello world!" + System.currentTimeMillis();
        channel.basicPublish("delaysync.exchange", "deal.message", null, message.getBytes());
        System.out.println("sent message: " + message + ",date:" + System.currentTimeMillis());
        // 关闭频道和连接
        channel.close();
        connection.close();
    }
}

【消费者配置】：

package com.caox.rabbitmq.demo._15_rabbitmq_delay_query_ttl;

import com.rabbitmq.client.Channel;
import com.rabbitmq.client.Connection;
import com.rabbitmq.client.ConnectionFactory;
import com.rabbitmq.client.QueueingConsumer;

/**
 * Created by nazi on 2018/8/1.
 * 延时队列 队列配置延时
 */

public class Consumer {

    private static String queue_name = "test.queue";

    public static void main(String[] args) throws Exception {
        ConnectionFactory factory = new ConnectionFactory();
        factory.setHost("127.0.0.1");
        factory.setUsername("caoxia");
        factory.setPassword("caoxia123456");
        Connection connection = factory.newConnection();
        Channel channel = connection.createChannel();
        // 声明队列
        channel.queueDeclare(queue_name, true, false, false, null);
        // 绑定路由
        channel.queueBind(queue_name, "delay.exchange", "deal.message");
        QueueingConsumer consumer = new QueueingConsumer(channel);
        // 指定消费队列
        channel.basicConsume(queue_name, true, consumer);
        while (true) {
            // nextDelivery是一个阻塞方法（内部实现其实是阻塞队列的take方法）
            QueueingConsumer.Delivery delivery = consumer.nextDelivery();
            String message = new String(delivery.getBody());
            System.out.println("received message:" + message + ",date:" + System.currentTimeMillis());
        }
    }

}

（二）：在消息上设置TTL  

 

【生产者配置】：

package com.caox.rabbitmq.demo._14_rabbitmq_delay_news_ttl;

import com.rabbitmq.client.AMQP;
import com.rabbitmq.client.Channel;
import com.rabbitmq.client.Connection;
import com.rabbitmq.client.ConnectionFactory;

import java.util.HashMap;
/**
 * Created by nazi on 2018/8/1.
 * 延时队列 消息配置延时
 */

public class ProducerUpt {

    private static String queue_name = "message_ttl_queue";

    public static void main(String[] args) throws Exception {
        ConnectionFactory factory = new ConnectionFactory();
        factory.setHost("127.0.0.1");
        factory.setUsername("caoxia");
        factory.setPassword("caoxia123456");
        Connection connection = factory.newConnection();
        Channel channel = connection.createChannel();
        HashMap<String, Object> arguments = new HashMap<String, Object>();
        arguments.put("x-dead-letter-exchange", "amq.direct");
        arguments.put("x-dead-letter-routing-key", "message_ttl_routingKey");
        channel.queueDeclare("delay_queue", true, false, false, arguments);

        // 声明队列
        channel.queueDeclare(queue_name, true, false, false, null);
        // 绑定路由
        channel.queueBind(queue_name, "amq.direct", "message_ttl_routingKey");

        String message = "hello world!" + System.currentTimeMillis();
        // 设置延时属性
        AMQP.BasicProperties.Builder builder = new AMQP.BasicProperties.Builder();
        // 持久性 non-persistent (1) or persistent (2)
        AMQP.BasicProperties properties = builder.expiration("120000").deliveryMode(2).build();
        // routingKey =delay_queue 进行转发
        channel.basicPublish("", "delay_queue", properties, message.getBytes());
        System.out.println("sent message: " + message + ",date:" + System.currentTimeMillis());
        // 关闭频道和连接
        channel.close();
        connection.close();
    }
}

【消费者配置】： 

package com.caox.rabbitmq.demo._14_rabbitmq_delay_news_ttl;

import java.util.HashMap;

import com.rabbitmq.client.Channel;
import com.rabbitmq.client.Connection;
import com.rabbitmq.client.ConnectionFactory;
import com.rabbitmq.client.QueueingConsumer;

/**
 * Created by nazi on 2018/8/1.
 * 延时队列 消息配置延时
 */


public class ConsumerUpt {

    private static String queue_name = "message_ttl_queue";

    public static void main(String[] args) throws Exception {
        ConnectionFactory factory = new ConnectionFactory();
        factory.setHost("127.0.0.1");
        factory.setUsername("caoxia");
        factory.setPassword("caoxia123456");
        Connection connection = factory.newConnection();
        Channel channel = connection.createChannel();
        HashMap<String, Object> arguments = new HashMap<String, Object>();
        arguments.put("x-dead-letter-exchange", "amq.direct");
        arguments.put("x-dead-letter-routing-key", "message_ttl_routingKey");
        channel.queueDeclare("delay_queue", true, false, false, arguments);

        // 声明队列
        channel.queueDeclare(queue_name, true, false, false, null);
        // 绑定路由
        channel.queueBind(queue_name, "amq.direct", "message_ttl_routingKey");

        QueueingConsumer consumer = new QueueingConsumer(channel);
        // 指定消费队列
        channel.basicConsume(queue_name, true, consumer);
        while (true) {
            // nextDelivery是一个阻塞方法（内部实现其实是阻塞队列的take方法）
            QueueingConsumer.Delivery delivery = consumer.nextDelivery();
            String message = new String(delivery.getBody());
            System.out.println("received message:" + message + ",date:" + System.currentTimeMillis());
        }
    }

}

【结论】：

虽然 consumer 从来看不到过期的 message ，但是在过期 message 到达 queue 的头部时确实会被真正的丢弃（或者 dead-lettered ）。

【在队列上设置TTL】：当对每一个 queue 设置了 TTL 值时不会产生任何问题，因为过期的 message 总是会出现在 queue 的头部。一旦消息过期，就会从队列中抹去。

【在消息上设置TTL】：当对每一条 message 设置了 TTL 时，过期的 message 可能会排队于未过期 message 的后面，直到这些消息被 consume 到或者过期了。在这种情况下，这些过期的 message 使用的资源将不会被释放，且会在 queue 统计信息中被计算进去（例如，queue 中存在的 message 的数量）。即使消息过期，也不会马上从队列中抹去，因为每条消息是否过期时在即将投递到消费者之前判定的。为什么两者得处理方法不一致？因为第一种方法里，队列中已过期的消息肯定在队列头部，RabbitMQ只要定期从队头开始扫描是否有过期消息即可，而第二种方法里，每条消息的过期时间不同，如果要删除所有过期消息，势必要扫描整个队列，所以不如等到此消息即将被消费时再判定是否过期，如果过期，再进行删除。