Redis实现消息队列的方式汇总以及代码实现
- 前言
- 开始前准备
- 1、添加依赖
- 2、添加配置的Bean
- 具体实现
- 一、从最简单的开始:List 队列
- 代码实现
- 二、发布订阅模式:Pub/Sub
- 1、使用RedisMessageListenerContainer实现订阅
- 2、还可以使用redisTemplate实现订阅
- 三、、 趋于成熟的队列:Stream
- 具体java代码实现:
- 总结
前言
经常听到很多人讨论,关于「把 Redis 当作队列来用是否合适」的问题。
有些人表示赞成,他们认为 Redis 很轻量,用作队列很方便。也些人则反对,认为 Redis 会「丢」数据,最好还是用「专业」的队列中间件更稳妥。
这篇文章就聊一聊把 Redis 当作队列,究竟是否合适这个问题。我们会从简单到复杂,一步步带你梳理其中的细节,把常用的实现方式展现一遍。
开始前准备
1、添加依赖
<dependency>
<groupId>org.projectlombok</groupId>
<artifactId>lombok</artifactId>
</dependency>
<dependency>
<groupId>org.springframework.boot</groupId>
<artifactId>spring-boot-starter-data-redis</artifactId>
</dependency>
<dependency>
<groupId>org.springframework.boot</groupId>
<artifactId>spring-boot-starter-web</artifactId>
</dependency>
<dependency>
<groupId>cn.hutool</groupId>
<artifactId>hutool-all</artifactId>
<version>5.4.0</version>
</dependency>
2、添加配置的Bean
避免不方便用软件查看存储的数据
/**
* redisTemplate 序列化使用的jdkSerializeable, 存储二进制字节码, 所以自定义序列化类
* @param redisConnectionFactory
* @return
*/
@Bean
public RedisTemplate<String, Object> redisTemplate(RedisConnectionFactory redisConnectionFactory) {
RedisTemplate<String, Object> redisTemplate = new RedisTemplate<>();
redisTemplate.setConnectionFactory(redisConnectionFactory);
<span >// 使用Jackson2JsonRedisSerialize 替换默认序列化</span>
<span >Jackson2JsonRedisSerializer</span> jackson2JsonRedisSerializer <span >=</span> <span >new</span> <span >Jackson2JsonRedisSerializer</span><span >(</span><span >Object</span><span >.</span><span >class</span><span >)</span><span >;</span>
<span >ObjectMapper</span> objectMapper <span >=</span> <span >new</span> <span >ObjectMapper</span><span >(</span><span >)</span><span >;</span>
objectMapper<span >.</span><span >setVisibility</span><span >(</span><span >PropertyAccessor</span><span >.</span><span >ALL</span><span >,</span> <span >JsonAutoDetect<span >.</span>Visibility</span><span >.</span><span >ANY</span><span >)</span><span >;</span>
objectMapper<span >.</span><span >enableDefaultTyping</span><span >(</span><span >ObjectMapper<span >.</span>DefaultTyping</span><span >.</span><span >NON_FINAL</span><span >)</span><span >;</span>
jackson2JsonRedisSerializer<span >.</span><span >setObjectMapper</span><span >(</span>objectMapper<span >)</span><span >;</span>
<span >// 设置value的序列化规则和 key的序列化规则</span>
redisTemplate<span >.</span><span >setKeySerializer</span><span >(</span><span >new</span> <span >StringRedisSerializer</span><span >(</span><span >)</span><span >)</span><span >;</span>
<span >//jackson2JsonRedisSerializer就是JSON序列号规则,</span>
redisTemplate<span >.</span><span >setValueSerializer</span><span >(</span>jackson2JsonRedisSerializer<span >)</span><span >;</span>
redisTemplate<span >.</span><span >afterPropertiesSet</span><span >(</span><span >)</span><span >;</span>
<span >return</span> redisTemplate<span >;</span>
<span >}</span>
具体实现
一、从最简单的开始:List 队列
首先,我们先从最简单的场景开始讲起。
如果你的业务需求足够简单,想把 Redis 当作队列来使用,肯定最先想到的就是使用 List 这个数据类型。
因为 List 底层的实现就是一个「链表」,在头部和尾部操作元素,时间复杂度都是 O(1),这意味着它非常符合消息队列的模型。
如果把 List 当作队列,你可以这么来用。
代码实现
生产者端读取:
@RestController
@RequestMapping("/redis01")
public class RedisTest1 {
<span >@Autowired</span>
<span >private</span> <span >RedisTemplate</span><span ><span ><</span><span >String</span><span >,</span> <span >Object</span><span >></span></span> redisTemplate<span >;</span>
<span >//LPUSH 发布消息</span>
<span >@GetMapping</span><span >(</span><span >"/set"</span><span >)</span>
<span >public</span> <span >void</span> <span >set</span><span >(</span><span >String</span> code<span >)</span><span >{<!-- --></span>
redisTemplate<span >.</span><span >opsForList</span><span >(</span><span >)</span><span >.</span><span >leftPush</span><span >(</span><span >"code"</span><span >,</span>code<span >)</span><span >;</span>
<span >}</span>
<span >// RPOP 拉取消息</span>
<span >@GetMapping</span><span >(</span><span >"/get1"</span><span >)</span>
<span >public</span> <span >String</span> <span >get1</span><span >(</span><span >String</span> key<span >)</span><span >{<!-- --></span>
<span >Object</span> code <span >=</span> redisTemplate<span >.</span><span >opsForList</span><span >(</span><span >)</span><span >.</span><span >rightPop</span><span >(</span>key<span >)</span><span >;</span>
<span >if</span> <span >(</span>code<span >!=</span><span >null</span><span >)</span><span >{<!-- --></span>
<span >return</span> code<span >.</span><span >toString</span><span >(</span><span >)</span><span >;</span>
<span >}</span>
<span >return</span> <span >"redis中没数据!"</span><span >;</span>
<span >}</span>
实现模型:这个模型也非常简单容易理解。
但这里有个小问题,当队列中已经没有消息了,消费者在执行 RPOP 时,会返回 NULL。
一般在编写消费者时,会采用一个死循环,这个实现方式就是不断去队列中拉取数据。
@GetMapping("/get2")
public String get2(String key) throws InterruptedException {
while (true){
Object code = redisTemplate.opsForList().rightPop(key);
System.out.println(code);
// 读取到消息,退出,没读到继续循环
if (code!=null){
return code.toString();
}
}
}
如果此时队列为空,那消费者依旧会频繁拉取消息,这会造成「CPU 空转」,不仅浪费 CPU 资源,还会对 Redis 造成压力。
怎么解决这个问题呢?
也很简单,当队列为空时,我们可以「休眠」一会,再去尝试拉取消息。代码可以修改成这样:
@GetMapping("/get2")
public String get2(String key) throws InterruptedException {
while (true){
Object code = redisTemplate.opsForList().rightPop(key);
System.out.println(code);
// 读取到消息,退出,没读到继续循环
if (code!=null){
return code.toString();
}
Thread.sleep(2000);
}
}
这就解决了 CPU 空转问题。
这个问题虽然解决了,但又带来另外一个问题:当消费者在休眠等待时,有新消息来了,那消费者处理新消息就会存在「延迟」。
假设设置的休眠时间是 2s,那新消息最多存在 2s 的延迟。
要想缩短这个延迟,只能减小休眠的时间。但休眠时间越小,又有可能引发 CPU 空转问题。
鱼和熊掌不可兼得。
那如何做,既能及时处理新消息,还能避免 CPU 空转呢?
Redis 是否存在这样一种机制:如果队列为空,消费者在拉取消息时就「阻塞等待」,一旦有新消息过来,就通知我的消费者立即处理新消息呢?
幸运的是,Redis 确实提供了「阻塞式」拉取消息的命令:BRPOP / BLPOP,这里的 B 指的是阻塞(Block)。
在java中也已经封装好了,调用pop方法时,直接设置一个过期时间就行
@GetMapping("/get3")
public String get3(String key) throws InterruptedException {
Object code = redisTemplate.opsForList().rightPop(key,0, TimeUnit.SECONDS);
if (code==null){
return "数据读取超时!";
}
return code.toString();
}
使用 BRPOP 这种阻塞式方式拉取消息时,还支持传入一个「超时时间」,如果设置为 0,则表示不设置超时,直到有新消息才返回,否则会在指定的超时时间后返回 NULL。
这个方案不错,既兼顾了效率,还避免了 CPU 空转问题,一举两得。
注意:如果设置的超时时间太长,这个连接太久没有活跃过,可能会被 Redis Server 判定为无效连接,之后 Redis Server
会强制把这个客户端踢下线。所以,采用这种方案,客户端要有重连机制。
解决了消息处理不及时的问题,你可以再思考一下,这种队列模型,有什么缺点?
我们一起来分析一下:
- 不支持重复消费:消费者拉取消息后,这条消息就从 List 中删除了,无法被其它消费者再次消费,即不支持多个消费者消费同一批数据
- 消息丢失:消费者拉取到消息后,如果发生异常宕机,那这条消息就丢失了
第一个问题是功能上的,使用 List 做消息队列,它仅仅支持最简单的,一组生产者对应一组消费者,不能满足多组生产者和消费者的业务场景。
第二个问题就比较棘手了,因为从 List 中 POP 一条消息出来后,这条消息就会立即从链表中删除了。也就是说,无论消费者是否处理成功,这条消息都没办法再次消费了。
这也意味着,如果消费者在处理消息时异常宕机,那这条消息就相当于丢失了。
针对这 2 个问题怎么解决呢?我们一个个来看。
二、发布订阅模式:Pub/Sub
从名字就能看出来,这个模块是 Redis 专门是针对「发布/订阅」这种队列模型设计的。
它正好可以解决前面提到的第一个问题:重复消费。
即多组生产者、消费者的场景,我们来看它是如何做的。
Redis 提供了 PUBLISH / SUBSCRIBE 命令,来完成发布、订阅的操作。
依赖继续用前面的就行
1、使用RedisMessageListenerContainer实现订阅
- 通过实现MessageListener接口来处理接收到的消息。这允许您在Spring应用程序中以更高级的方式处理消息,例如使用依赖注入和其他Spring功能。它还支持基于注解的消息监听器,使消息处理更加简洁和灵活。
- 该方式是Spring Data Redis库提供的方法,用于在Spring应用程序中使用Redis的发布订阅功能。它需要创建一个MessageListenerContainer对象,并通过调用addMessageListener方法来添加消息监听器。
- 添加监控器,用于监听通道
为了便于更加直观的对比测试,我添加了两个
/**
* @author zhengfuping
* @version 1.0
* @description: TODO 配置监控器
* @date 2023/7/28 17:10
*/
@Component
public class RedisMessaeListener1 implements MessageListener {
@Override
public void onMessage(Message message, byte[] pattern) {
<span >String</span> channel <span >=</span> <span >new</span> <span >String</span><span >(</span>message<span >.</span><span >getChannel</span><span >(</span><span >)</span><span >)</span><span >;</span>
<span >String</span> body <span >=</span> <span >new</span> <span >String</span><span >(</span>message<span >.</span><span >getBody</span><span >(</span><span >)</span><span >)</span><span >;</span>
<span >System</span><span >.</span>out<span >.</span><span >println</span><span >(</span><span >"监听器1号:消息: "</span> <span >+</span> body <span >+</span> <span >" 通道QQ: "</span> <span >+</span> channel<span >)</span><span >;</span>
<span >}</span>
}
/########################/
/**
• @author zhengfuping
• @version 1.0
• @description: TODO 配置监控器1
• @date 2023/8/2 11:24
*/
@Component
public class RedisMessaeListener2 implements MessageListener {
@Override
public void onMessage(Message message, byte[] pattern) {
String channel = new String(message.getChannel());
String body = new String(message.getBody());
System.out.println("监听器2号:消息: “ + body + ” 通道QQ: " + channel);
}
}
- 配置订阅可以单和多个
用于绑定主题(通道)和监听器,该发送是
/**
* @author zhengfuping
* @version 1.0
* @description: TODO 使用RedisMessageListenerContainer直接注入到bean进行监听
* @date 2023/7/28 15:43
*/
@Configuration
public class RedisPubSubExample {
<span >@Autowired</span>
<span >private</span> <span >RedisMessaeListener1</span> redisMessaeListener1<span >;</span>
<span >@Autowired</span>
<span >private</span> <span >RedisMessaeListener2</span> redisMessaeListener2<span >;</span>
<span >/**
* 订阅三个频道
* @author zhengfuping
* @date 2023/8/2 11:19
* @param redisConnectionFactory redis线程工厂
* @return RedisMessageListenerContainer
*/</span>
// @Bean
public RedisMessageListenerContainer subscribeToChannel(RedisConnectionFactory redisConnectionFactory){
RedisMessageListenerContainer listenerContainer = new RedisMessageListenerContainer();
listenerContainer<span >.</span><span >setConnectionFactory</span><span >(</span>redisConnectionFactory<span >)</span><span >;</span>
<span >List</span><span ><span ><</span><span >Topic</span><span >></span></span> list <span >=</span> <span >new</span> <span >ArrayList</span><span ><span ><</span><span >></span></span><span >(</span><span >)</span><span >;</span>
list<span >.</span><span >add</span><span >(</span><span >new</span> <span >PatternTopic</span><span >(</span><span >"TEST01"</span><span >)</span><span >)</span><span >;</span>
list<span >.</span><span >add</span><span >(</span><span >new</span> <span >PatternTopic</span><span >(</span><span >"TEST02"</span><span >)</span><span >)</span><span >;</span>
list<span >.</span><span >add</span><span >(</span><span >new</span> <span >PatternTopic</span><span >(</span><span >"TEST03"</span><span >)</span><span >)</span><span >;</span>
<span >/*
* redisMessaeListener 消息监听器
* list 订阅的主题(可以单个和多个)
*/</span>
listenerContainer<span >.</span><span >addMessageListener</span><span >(</span>redisMessaeListener1<span >,</span>list<span >)</span><span >;</span>
listenerContainer<span >.</span><span >addMessageListener</span><span >(</span>redisMessaeListener2<span >,</span><span >new</span> <span >PatternTopic</span><span >(</span><span >"TEST01"</span><span >)</span><span >)</span><span >;</span>
<span >return</span> listenerContainer<span >;</span>
<span >}</span>
}
- 向指定频道发送消息
/**
* PUBLISH 发送消息到指定频道
* @author zhengfuping
* @date 2023/8/2 11:14
* @param channel 通道
* @param name 数据
* @param age
* @return Object
*/
@GetMapping("/pub")
public Object pub(String channel,String name,Integer age) {
User user = new User(name, age);
redisTemplate.convertAndSend(channel,user);
return user;
}
2、还可以使用redisTemplate实现订阅
- 该方式是Redis客户端的方法,用于在独立的Redis客户端中直接使用发布订阅功能。它需要创建一个Redis连接对象,并通过调用subscribe方法来订阅一个或多个频道。
- 如果您只是在独立的Redis客户端中使用发布订阅功能,并且不需要使用Spring的其他功能,则可以选择connection.subscribe
/**
* 自行添加订阅
*/
@GetMapping("/sub")
public void sub(String channel) {
RedisConnection connection = redisTemplate.getConnectionFactory().getConnection();
<span >/*
* MessageListener:监听器,直接使用内部类实现绑定监听可以把数据传递出去
* channel 订阅频道
*/</span>
connection<span >.</span><span >subscribe</span><span >(</span><span >(</span>message<span >,</span> pattern<span >)</span> <span >-></span> <span >{<!-- --></span>
<span >String</span> channel1 <span >=</span> <span >new</span> <span >String</span><span >(</span>message<span >.</span><span >getChannel</span><span >(</span><span >)</span><span >)</span><span >;</span>
<span >String</span> body <span >=</span> <span >new</span> <span >String</span><span >(</span>message<span >.</span><span >getBody</span><span >(</span><span >)</span><span >)</span><span >;</span>
<span >System</span><span >.</span>out<span >.</span><span >println</span><span >(</span><span >"subscribe方式监听:消息: "</span> <span >+</span> body <span >+</span> <span >" 通道QQ: "</span> <span >+</span> channel1<span >)</span><span >;</span>
<span >}</span><span >,</span> channel<span >.</span><span >getBytes</span><span >(</span><span >)</span><span >)</span><span >;</span>
// connection.close();
}
发送消息
@GetMapping("/pub")
public Object pub(String channel,String name,Integer age) {
User user = new User(name, age);
redisTemplate.convertAndSend(channel,user);
return user;
}
最终监听到的结果
三、、 趋于成熟的队列:Stream
我们来看 Stream 是如何解决上面这些问题的。
我们依旧从简单到复杂,依次来看 Stream 在做消息队列时,是如何处理的?
首先,Stream 通过 XADD 和 XREAD 完成最简单的生产、消费模型:
- 生产者发布 2 条消息:
// *表示让Redis自动生成消息ID
127.0.0.1:6379> XADD queue * name zhangsan
"1618469123380-0"
127.0.0.1:6379> XADD queue * name lisi
"1618469127777-0"
- 消费者拉取消息:
// 从开头读取5条消息,0-0表示从开头读取
127.0.0.1:6379> XREAD COUNT 5 STREAMS queue 0-0
1) 1) "queue"
2) 1) 1) "1618469123380-0"
2) 1) "name"
2) "zhangsan"
2) 1) "1618469127777-0"
2) 1) "name"
2) "lisi"
流程图
具体java代码实现:
- 先配置监听消息类
@Slf4j
@Component
public class ListenerMessage implements StreamListener<String, MapRecord<String, String, String>> {
<span >@Override</span>
<span >public</span> <span >void</span> <span >onMessage</span><span >(</span><span >MapRecord</span><span ><span ><</span><span >String</span><span >,</span> <span >String</span><span >,</span> <span >String</span><span >></span></span> entries<span >)</span> <span >{<!-- --></span>
log<span >.</span><span >info</span><span >(</span><span >"接受到来自redis的消息"</span><span >)</span><span >;</span>
<span >System</span><span >.</span>out<span >.</span><span >println</span><span >(</span><span >"message id "</span><span >+</span>entries<span >.</span><span >getId</span><span >(</span><span >)</span><span >)</span><span >;</span>
<span >System</span><span >.</span>out<span >.</span><span >println</span><span >(</span><span >"stream "</span><span >+</span>entries<span >.</span><span >getStream</span><span >(</span><span >)</span><span >)</span><span >;</span>
<span >System</span><span >.</span>out<span >.</span><span >println</span><span >(</span><span >"body "</span><span >+</span>entries<span >.</span><span >getValue</span><span >(</span><span >)</span><span >)</span><span >;</span>
<span >}</span>
}
- 添加工具类,实现初始化
@Component
@Slf4j
public class RedisStreamUtil {
@Autowired
private RedisTemplate<String,Object> redisTemplate;
<span >/**
* @author zhengfuping 添加数据
* @param streamKey
* @param map
* @return RecordId
*/</span>
<span >public</span> <span >RecordId</span> <span >addStream</span><span >(</span><span >String</span> streamKey<span >,</span><span >Map</span><span ><span ><</span><span >String</span><span >,</span> <span >Object</span><span >></span></span> map<span >)</span><span >{<!-- --></span>
<span >RecordId</span> recordId <span >=</span> redisTemplate<span >.</span><span >opsForStream</span><span >(</span><span >)</span><span >.</span><span >add</span><span >(</span>streamKey<span >,</span> map<span >)</span><span >;</span>
<span >return</span> recordId<span >;</span>
<span >}</span>
<span >/**
* 用来创建绑定流和组
*/</span>
<span >public</span> <span >void</span> <span >addGroup</span><span >(</span><span >String</span> key<span >,</span> <span >String</span> groupName<span >)</span><span >{<!-- --></span>
redisTemplate<span >.</span><span >opsForStream</span><span >(</span><span >)</span><span >.</span><span >createGroup</span><span >(</span>key<span >,</span>groupName<span >)</span><span >;</span>
<span >}</span>
<span >/**
* 用来判断key是否存在
*/</span>
<span >public</span> <span >boolean</span> <span >hasKey</span><span >(</span><span >String</span> key<span >)</span><span >{<!-- --></span>
<span >if</span><span >(</span>key<span >==</span><span >null</span><span >)</span><span >{<!-- --></span>
<span >return</span> <span >false</span><span >;</span>
<span >}</span><span >else</span><span >{<!-- --></span>
<span >return</span> redisTemplate<span >.</span><span >hasKey</span><span >(</span>key<span >)</span><span >;</span>
<span >}</span>
<span >}</span>
<span >/**
* 用来删除掉消费了的消息
*/</span>
<span >public</span> <span >void</span> <span >delField</span><span >(</span><span >String</span> key<span >,</span><span >String</span> recordIds<span >)</span><span >{<!-- --></span>
redisTemplate<span >.</span><span >opsForStream</span><span >(</span><span >)</span><span >.</span><span >delete</span><span >(</span>key<span >,</span>recordIds<span >)</span><span >;</span>
<span >}</span>
<span >/**
* 用来初始化 实现绑定
*/</span>
<span >public</span> <span >void</span> <span >initStream</span><span >(</span><span >String</span> key<span >,</span> <span >String</span> group<span >)</span><span >{<!-- --></span>
<span >//判断key是否存在,如果不存在则创建</span>
<span >boolean</span> hasKey <span >=</span> <span >hasKey</span><span >(</span>key<span >)</span><span >;</span>
<span >if</span><span >(</span><span >!</span>hasKey<span >)</span><span >{<!-- --></span>
<span >Map</span><span ><span ><</span><span >String</span><span >,</span><span >Object</span><span >></span></span> map <span >=</span> <span >new</span> <span >HashMap</span><span ><span ><</span><span >></span></span><span >(</span><span >)</span><span >;</span>
map<span >.</span><span >put</span><span >(</span><span >"field"</span><span >,</span><span >"value"</span><span >)</span><span >;</span>
<span >RecordId</span> recordId <span >=</span> <span >addStream</span><span >(</span>key<span >,</span> map<span >)</span><span >;</span>
<span >addGroup</span><span >(</span>key<span >,</span>group<span >)</span><span >;</span> <span >//把Stream和gropu绑定</span>
<span >delField</span><span >(</span>key<span >,</span>recordId<span >.</span><span >getValue</span><span >(</span><span >)</span><span >)</span><span >;</span>
log<span >.</span><span >info</span><span >(</span><span >"stream:{}-group:{} initialize success"</span><span >,</span>key<span >,</span>group<span >)</span><span >;</span>
<span >}</span>
<span >}</span>
}
- 添加配置类,配置Stream
/**
* @author zhengfuping
* @version 1.0
* @description: TODO 添加配置类,配置Stream
*/
@Configuration
@Slf4j
public class RedisStreamConfig {
<span >@Autowired</span>
<span >private</span> <span >RedisStreamUtil</span> redisStream<span >;</span>
<span >@Autowired</span>
<span >private</span> <span >ListenerMessage</span> listenerMessage<span >;</span>
<span >@Bean</span>
<span >public</span> <span >Subscription</span> <span >subscription</span><span >(</span><span >RedisConnectionFactory</span> factory<span >)</span><span >{<!-- --></span>
// 代码中的var是使用了Lombok的可变局部变量。主要是为了方便
// StreamMessageListenerContainer: 消息侦听容器,不能在外部实现。创建后,StreamMessageListenerContainer可以订阅Redis流并使用传入的消息
var options = StreamMessageListenerContainer
.StreamMessageListenerContainerOptions
.builder()
.pollTimeout(Duration.ofSeconds(1))
.build();
redisStream.initStream(“mystream”,“mygroup”); //调用初始化
var listenerContainer = StreamMessageListenerContainer.create(factory,options);
<span >/*
* 注意这里接受到消息后会被自动的确认,如果不想自动确认请使用其他的创建订阅方式
* 消费组 consumer group ,它不能为null (Consumer类型)
* stream offset ,stream的偏移量(StreamOffset 类型)
* listener 不能为null (StreamListener<K,V> 类型)
*/</span>
<span >var</span> subscription <span >=</span> listenerContainer<span >.</span><span >receiveAutoAck</span><span >(</span><span >Consumer</span><span >.</span><span >from</span><span >(</span><span >"mygroup"</span><span >,</span><span >"huhailong"</span><span >)</span><span >,</span>
<span >StreamOffset</span><span >.</span><span >create</span><span >(</span><span >"mystream"</span><span >,</span> <span >ReadOffset</span><span >.</span><span >lastConsumed</span><span >(</span><span >)</span><span >)</span><span >,</span>listenerMessage<span >)</span><span >;</span>
listenerContainer<span >.</span><span >start</span><span >(</span><span >)</span><span >;</span>
<span >return</span> subscription<span >;</span>
<span >}</span>
}
- 调用测试
/**
* @author zhengfuping
* @version 1.0
* @description: TODO
* @date 2023/8/2 16:06
*/
@RestController
@RequestMapping("/redisStream")
public class RedisStreamTest {
<span >@Autowired</span>
<span >private</span> <span >RedisStreamUtil</span> redisStream<span >;</span>
<span >@GetMapping</span><span >(</span><span >"add"</span><span >)</span>
<span >public</span> <span >void</span> <span >add</span><span >(</span><span >String</span> key<span >,</span><span >String</span> data<span >)</span><span >{<!-- --></span>
<span >Map</span><span ><span ><</span><span >String</span><span >,</span> <span >Object</span><span >></span></span> map <span >=</span> <span >new</span> <span >HashMap</span><span ><span ><</span><span >></span></span><span >(</span><span >)</span><span >;</span>
map<span >.</span><span >put</span><span >(</span>key<span >,</span>data<span >)</span><span >;</span>
// 添加数据到mystream流中
RecordId recordId = redisStream.addStream(“mystream”, map);
// 删除流中消费了的指定key的数据
redisStream.delField(“mystream”,recordId.getValue());
}
}
Stream的好处在于可以写入到 RDB 和 AOF 做持久化。
Stream是新增加的数据类型,它与其它数据类型一样,每个写操作,也都会写入到 RDB 和 AOF 中。
我们只需要配置好持久化策略,这样的话,就算 Redis 宕机重启,Stream 中的数据也可以从 RDB 或 AOF 中恢复回来。
总结
好了,总结一下。这篇文章我们从「Redis 能否用作队列」这个角度出发,介绍了 List、Pub/Sub、Stream 在做队列的使用方式,以及它们各自的优劣。
之后又把 Redis 和专业的消息队列中间件做对比,发现 Redis 的不足之处。
最后,我们得出 Redis 做队列的合适场景。