怎么保证RabbitMq消息不丢失

生产者

  • 做消息的持久化
    • 消息持久化不能保证完全不丢失消息,可以存在存储磁盘的时候还没有存储完成,但是服务宕机了也会导致消息丢失,通过发布确定保证消息不丢失
  • 消息确定机制
import com.rabbitmq.client.*;
public class PersistentProducer {
    private final static String QUEUE_NAME = "persistent_queue";

    public static void main(String[] argv) throws Exception {
        ConnectionFactory factory = new ConnectionFactory();
        factory.setHost("localhost");
        try (Connection connection = factory.newConnection();
             Channel channel = connection.createChannel()) {

            // 声明一个持久化队列
            channel.queueDeclare(QUEUE_NAME, true, false, false, null);

            // 启用生产者确认
            channel.confirmSelect();

            String message = "Persistent message with producer confirm!";
            channel.basicPublish("", QUEUE_NAME, 
                MessageProperties.PERSISTENT_TEXT_PLAIN, 
                message.getBytes());

            // 检查消息是否成功发送
            if (channel.waitForConfirms()) {
                System.out.println("Message sent successfully!");
            } else {
                System.out.println("Message failed to send!");
            }
        }
    }
}
  • channel.queueDeclare(QUEUE_NAME, true, false, false, null):队列是持久化的,确保 RabbitMQ 重启后队列不会丢失。
  • MessageProperties.PERSISTENT_TEXT_PLAIN:确保消息持久化存储。
  • channel.confirmSelect():启用生产者确认,确保消息成功送达 RabbitMQ。
  • channel.waitForConfirms():等待确认,如果生产者发送消息时发生失败,会捕获错误。

交换机

  • 选择合适的交换机类型:常见的交换机类型包括 direct、fanout、topic 和 headers,选择正确的类型来确保消息路由正确。
  • 使用死信队列(Dead Letter Exchange, DLX):如果消息因某些原因无法被消费,可以将消息转发到死信队列进行进一步处理。
import com.rabbitmq.client.*;

public class DirectExchangeProducer {
    private final static String EXCHANGE_NAME = "direct_logs";
    private final static String QUEUE_NAME = "persistent_queue";

    public static void main(String[] argv) throws Exception {
        ConnectionFactory factory = new ConnectionFactory();
        factory.setHost("localhost");
        try (Connection connection = factory.newConnection();
             Channel channel = connection.createChannel()) {

            // 声明交换机和队列
            channel.exchangeDeclare(EXCHANGE_NAME, "direct");
            channel.queueDeclare(QUEUE_NAME, true, false, false, null);
            channel.queueBind(QUEUE_NAME, EXCHANGE_NAME, "info");

            String message = "Hello, Direct Exchange!";
            channel.basicPublish(EXCHANGE_NAME, "info", 
                MessageProperties.PERSISTENT_TEXT_PLAIN, 
                message.getBytes());

            System.out.println("Sent: " + message);
        }
    }
}

public class DeadLetterConsumer {
    private final static String DLX_QUEUE = "dlx_queue";

    public static void main(String[] argv) throws Exception {
        ConnectionFactory factory = new ConnectionFactory();
        factory.setHost("localhost");
        try (Connection connection = factory.newConnection();
             Channel channel = connection.createChannel()) {

            // 声明死信队列
            channel.queueDeclare(DLX_QUEUE, true, false, false, null);

            // 创建消费者回调
            DeliverCallback deliverCallback = (consumerTag, delivery) -> {
                String message = new String(delivery.getBody(), "UTF-8");
                System.out.println("Dead Letter Queue Received: " + message);
            };

            // 设置死信队列消费者
            channel.basicConsume(DLX_QUEUE, true, deliverCallback, consumerTag -> {});
        }
    }
}

消费者

  • 进行手动应答
  • 消息重试
import com.rabbitmq.client.*;

public class AckConsumer {
    private final static String QUEUE_NAME = "persistent_queue";

    public static void main(String[] argv) throws Exception {
        ConnectionFactory factory = new ConnectionFactory();
        factory.setHost("localhost");
        try (Connection connection = factory.newConnection();
             Channel channel = connection.createChannel()) {

            // 声明一个持久化队列
            channel.queueDeclare(QUEUE_NAME, true, false, false, null);

            // 创建一个消费者回调
            DeliverCallback deliverCallback = (consumerTag, delivery) -> {
                String message = new String(delivery.getBody(), "UTF-8");
                System.out.println("Received: " + message);

                try {
                    // 模拟消息处理
                    if (message.contains("error")) {
                        throw new Exception("Error while processing message");
                    }

                    // 手动确认消息
                    channel.basicAck(delivery.getEnvelope().getDeliveryTag(), false);
                    System.out.println("Message processed and acknowledged");
                } catch (Exception e) {
                    // 如果消息处理失败,可以将消息重新放回队列
                    System.out.println("Error processing message, requeueing: " + e.getMessage());
                    channel.basicNack(delivery.getEnvelope().getDeliveryTag(), false, true);
                }
            };

            // 设置手动确认
            channel.basicConsume(QUEUE_NAME, false, deliverCallback, consumerTag -> {});
        }
    }
}
  • channel.basicAck(delivery.getEnvelope().getDeliveryTag(), false):消息处理成功后确认消息。
  • channel.basicNack(delivery.getEnvelope().getDeliveryTag(), false, true):如果消费失败,重新将消息投递到队列中,供其他消费者处理。
posted @   pengrq  阅读(44)  评论(0编辑  收藏  举报
相关博文:
阅读排行:
· DeepSeek “源神”启动!「GitHub 热点速览」
· 我与微信审核的“相爱相杀”看个人小程序副业
· 微软正式发布.NET 10 Preview 1:开启下一代开发框架新篇章
· C# 集成 DeepSeek 模型实现 AI 私有化(本地部署与 API 调用教程)
· spring官宣接入deepseek,真的太香了~
点击右上角即可分享
微信分享提示