springboot rabbitmq

Posted haohao1234

tags:

篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了springboot rabbitmq相关的知识,希望对你有一定的参考价值。

一.消息发送端

pom.xml

    <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-amqp</artifactId>
        </dependency>
application.yml
spring:
  datasource:
    type: com.alibaba.druid.pool.DruidDataSource
    username: root
    password: 123456
    url: jdbc:mysql://192.168.99.100:3306/migu_cms?useUnicode=true&characterEncoding=utf8&serverTimezone=GMT&allowMultiQueries=true
  jackson:
    date-format: yyyy-MM-dd
    time-zone: GMT+8
  rabbitmq:
    username: guest
    password: guest
    host: 192.168.99.100
    publisher-returns: true
    publisher-confirm-type: correlated #必须配置这个才会确认回调
rabbitTemplate.setConfirmCallback 在消息确认投递到brocker后回调将redis中保存的msg删除
@Configuration
public class RabbitConfig {
    public static final Logger logger = LoggerFactory.getLogger(RabbitConfig.class);
    @Autowired
    CachingConnectionFactory cachingConnectionFactory;
    @Autowired
    StringRedisTemplate redisTemplate;
    @Bean
    RabbitTemplate rabbitTemplate(){
        RabbitTemplate rabbitTemplate = new RabbitTemplate(cachingConnectionFactory);// 从 producer 到 rabbitmq broker cluster 则会返回一个 confirmCallback
        //1 发送数据并返回(不确认rabbitmq服务器已成功接收)
        //2 异步的接收从rabbitmq返回的ack确认信息
        //3 收到ack后调用confirmCallback函数
        rabbitTemplate.setConfirmCallback((correlationData, ack, cause)->{
            String msgId = correlationData.getId();
            if(ack){
                redisTemplate.opsForHash().delete(Const.MAIL_CLIENT_LOG, msgId);
                logger.info(msgId+":消息发送成功");
            } else {
                logger.error(msgId+":消息发送失败");
            }
        });
        // 没有投递到QUEUE
        rabbitTemplate.setReturnCallback((msg, repCode, repText, exchange, routingkey)->{
            logger.error("消息发送失败");
            // ??
            rabbitTemplate.send(msg);
        });
        return rabbitTemplate;
    }

    @Bean
    Queue mailQueue(){
        return new Queue(Const.MAIL_QUEUE_NAME, true);
    }
    @Bean
    DirectExchange mailExchange(){
        return new DirectExchange(Const.MAIL_EXCHANGE_NAME, true, false);
    }
    @Bean
    Binding mailBinding(){
        return BindingBuilder.bind(mailQueue()).to(mailExchange()).with(Const.MAIL_ROUTING_KEY_NAME);
    }
}

redis中保存的message,带时间可以判断超过时间后重试发送,确保发送到brocker

public class MessageWithTime<T> {
    private long time;
    private T message;
}

执行完业务发送消息

// 新增banner
        bannerMapper.insertSelective(banner);
        // 发送rabbitmq
        String msgId = UUID.randomUUID().toString();
        rabbitTemplate.convertAndSend(Const.MAIL_EXCHANGE_NAME, Const.MAIL_ROUTING_KEY_NAME, banner, new CorrelationData(msgId));
        // redis缓存保存作为可靠性
        MessageWithTime<Banner> messageWithTime = new MessageWithTime<>(System.currentTimeMillis(), banner);
        redisTemplate.opsForHash().put(Const.MAIL_CLIENT_LOG, msgId, new Gson().toJson(messageWithTime));

定时从redis中获取未确认的消息,进行重发

@Component
public class RetryTask {
    public static final Logger logger = LoggerFactory.getLogger(RetryTask.class);
    @Autowired
    StringRedisTemplate redisTemplate;
    @Autowired
    RabbitTemplate rabbitTemplate;

    @PostConstruct
    private void startRetry(){
        new Thread(()->{
            while(true){
                try {
                    Thread.sleep(Const.RETRY_TIME_INTERVAL);

                    Map<Object, Object> map = redisTemplate.opsForHash().entries(Const.MAIL_CLIENT_LOG);

                    for(Map.Entry entry : map.entrySet()){
                        MessageWithTime<Banner> messageWithTime = (MessageWithTime<Banner>) entry.getValue();
                        String msgId = (String)entry.getKey();
                        // 超过三次重试
                        if(messageWithTime.getTime() + 3*Const.RETRY_TIME_INTERVAL < System.currentTimeMillis()){
                            logger.error("发送消息失败超过15秒 " + new Gson().toJson(messageWithTime));
                            redisTemplate.opsForHash().delete(Const.MAIL_CLIENT_LOG, entry.getKey());
                        } else {
                            // 重试
                            rabbitTemplate.convertAndSend(Const.MAIL_EXCHANGE_NAME, Const.MAIL_ROUTING_KEY_NAME, messageWithTime.getMessage(), new CorrelationData((String)entry.getKey()));
                            logger.info("重试发送消息:"+ msgId);
                        }
                    }
                } catch (Exception e) {
                    e.printStackTrace();
                }
            }
        }).start();
    }
}

二。消费端

消息的事务性

(1) 处理成功,从队列中删除消息
(2) 处理失败(网络问题,程序问题,服务挂了),将消息重新放回队列
为了做到这点,我们使用rabbitmq的手动ack模式

pom.xml

    <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-amqp</artifactId>
        </dependency>

application.properties

spring.rabbitmq.username=guest
spring.rabbitmq.password=guest
spring.rabbitmq.host=192.168.99.100
spring.rabbitmq.port=5672
# 开启手动确认 spring.rabbitmq.listener.simple.acknowledge-mode=manual spring.rabbitmq.listener.simple.prefetch=100
@Component
public class MailReceiver {
    public static final Logger logger = LoggerFactory.getLogger(MailReceiver.class);
    @Autowired
    JavaMailSender javaMailSender;
    @Autowired
    MailProperties mailProperties;
    @Autowired
    TemplateEngine templateEngine;
    @Autowired
    StringRedisTemplate redisTemplate;

    @RabbitListener(queues = Const.MAIL_QUEUE_NAME)
    public void handler(Message message, Channel channel){
        Long tag = null;
        try {
            Banner banner = (Banner) message.getPayload();
            MessageHeaders header = message.getHeaders();
            tag = (Long)header.get(AmqpHeaders.DELIVERY_TAG);
            String msgId = (String)header.get("spring_returned_message_correlation");
        // 防止重复消费
if(redisTemplate.opsForHash().hasKey(Const.MAIL_SERVER_LOG, msgId)){ logger.info("消息已经被消费:"+msgId); channel.basicAck(tag, false); return; } // 发送邮件 MimeMessage msg = javaMailSender.createMimeMessage(); MimeMessageHelper helper = new MimeMessageHelper(msg); helper.setTo("1234@qq.com"); helper.setFrom("1234@qq.com"); helper.setSubject("创建Banner"); helper.setSentDate(new Date()); Context context = new Context(); context.setVariable("bannerId", banner.getId()); context.setVariable("bannerName", banner.getBannerName()); context.setVariable("createTime", banner.getCreateTime()); String mail = templateEngine.process("mail", context); helper.setText(mail, true); javaMailSender.send(msg); // 成功后发送ack确认,会将队列中该消息删除 channel.basicAck(tag, false); redisTemplate.opsForHash().put(Const.MAIL_SERVER_LOG, msgId, "java"); logger.info("邮件发送成功"); } catch (Exception e) { e.printStackTrace(); // 消费失败后发送nack使信息重新投递 // channel.basicNack否认消息,消息重回队列给下一个消费者消费 try { channel.basicNack(tag, false, true); } catch (IOException ex) { ex.printStackTrace(); } logger.error("邮件发送失败"); } } }

 

以上是关于springboot rabbitmq的主要内容,如果未能解决你的问题,请参考以下文章

RabbitMQ---延迟队列,整合springboot

SpringBoot整合RabbitMQ(源代码)

RabbitMq一文彻底弄懂RabbitMq的四种交换机原理及springboot实战应用

springboot~rabbitmq的队列初始化和绑定

springboot rabbitmq整合

Springboot整合Rabbitmq(史上最详细)