Redis 与 Spring Cache：缓存抽象与实战

// 穿透解法一：空值缓存——查不到也写一个短 TTL 的空标记，挡住后续请求public User findById(Long id) {    String key = "user:" + id;    String cached = redis.opsForValue().get(key);    if (cached != null) {        return "NULL".equals(cached) ? null : JSON.parseObject(cached, User.class);    }    User user = userMapper.selectById(id);    if (user == null) {        // 短 TTL，别留太久，否则变成「缓存污染」        redis.opsForValue().set(key, "NULL", Duration.ofMinutes(2));        return null;    }    redis.opsForValue().set(key, JSON.toJSONString(user), Duration.ofMinutes(30));    return user;}

// 击穿解法：互斥锁重建——只让一个线程去查库，其余等待后重读缓存public User findByIdWithLock(Long id) {    String key = "user:" + id;    User user = getFromCache(key);    if (user != null) return user;    String lockKey = "lock:user:" + id;    // SETNX 抢锁：谁抢到谁去重建，抢不到就稍后重试读缓存    Boolean locked = redis.opsForValue().setIfAbsent(lockKey, "1", Duration.ofSeconds(10));    try {        if (Boolean.TRUE.equals(locked)) {            user = userMapper.selectById(id);                     // 只有持锁线程查库            redis.opsForValue().set(key, JSON.toJSONString(user), Duration.ofMinutes(30));        } else {            Thread.sleep(50);                                     // 等一等，再去读缓存            return getFromCache(key);        }    } finally {        if (Boolean.TRUE.equals(locked)) redis.delete(lockKey);    }    return user;}

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

spring:  data:    redis:      host: localhost      port: 6379      password: ${REDIS_PASSWORD:}      database: 0      timeout: 2s      lettuce:        pool:          max-active: 16      # 连接池上限，别设太大，Redis 本身是单线程          max-idle: 8          min-idle: 2

@Autowiredprivate RedisTemplate<String, Object> redisTemplate;public void save(User user) {    redisTemplate.opsForValue().set("user:1", user);   // 存一个 Java 对象}

127.0.0.1:6379> GET "user:1""\xac\xed\x00\x05sr\x00\x04com..User\x..."

@Configurationpublic class RedisConfig {    @Bean    public RedisTemplate<String, Object> redisTemplate(RedisConnectionFactory factory) {        RedisTemplate<String, Object> template = new RedisTemplate<>();        template.setConnectionFactory(factory);        // 泛型化 JSON 序列化器：会写入 @class 信息，反序列化时能还原具体类型        GenericJackson2JsonRedisSerializer jsonSerializer =                new GenericJackson2JsonRedisSerializer();        StringRedisSerializer keySerializer = new StringRedisSerializer();        template.setKeySerializer(keySerializer);        // key 用纯字符串，redis-cli 可读        template.setHashKeySerializer(keySerializer);        template.setValueSerializer(jsonSerializer);     // value 用 JSON        template.setHashValueSerializer(jsonSerializer);        template.afterPropertiesSet();        return template;    }}

@Configuration@EnableCaching        // 打开缓存抽象（代理拦截，与 @Transactional 同源）public class CacheConfig {}

@Servicepublic class UserService {    // key 用 SpEL 拼：user::42    @Cacheable(cacheNames = "user", key = "#id")    public User findById(Long id) {        return userMapper.selectById(id);      // 只有未命中才会执行这一行    }    // 更新后刷新缓存：返回值会被写入 user::#user.id    @CachePut(cacheNames = "user", key = "#user.id")    public User update(User user) {        userMapper.updateById(user);        return user;    }    // 删除：清掉指定 key    @CacheEvict(cacheNames = "user", key = "#id")    public void delete(Long id) {        userMapper.deleteById(id);    }}

package com.example.cache;public record Product(Long id, String name, java.math.BigDecimal price) {}

package com.example.cache;import org.springframework.stereotype.Service;@Servicepublic class ProductService {    private int dbHits = 0;    // key 显式写成 #id：无参方法才不会出现 SimpleKey.EMPTY 覆盖问题    @org.springframework.cache.annotation.Cacheable(cacheNames = "product", key = "#id")    public Product find(Long id) {        System.out.println("[db] SELECT * FROM product WHERE id=" + (++dbHits));        return new Product(id, "蜜蜂公仔", new java.math.BigDecimal("39.00"));    }    @org.springframework.cache.annotation.CachePut(cacheNames = "product", key = "#id")    public Product rename(Long id, String newName) {        System.out.println("[db] UPDATE product SET name=? WHERE id=" + id);        return new Product(id, newName, new java.math.BigDecimal("39.00"));    }    @org.springframework.cache.annotation.CacheEvict(cacheNames = "product", key = "#id")    public void evict(Long id) {        System.out.println("[cache] evict product::" + id);    }    public int dbHits() { return dbHits; }}

package com.example.cache;import org.springframework.boot.SpringApplication;import org.springframework.boot.autoconfigure.SpringBootApplication;import org.springframework.cache.annotation.EnableCaching;import org.springframework.context.ConfigurableApplicationContext;@SpringBootApplication@EnableCaching          // ← 忘这一行，下面四行输出会变成四行 [db]，且一句报错都没有public class CacheDemo {    public static void main(String[] args) {        try (ConfigurableApplicationContext ctx = SpringApplication.run(CacheDemo.class, args)) {            ProductService svc = ctx.getBean(ProductService.class);            svc.find(1L);              // 未命中 → 查库并回填            svc.find(1L);              // 命中 → 方法根本不执行            svc.rename(1L, "限量公仔"); // @CachePut → 执行方法并覆盖缓存            svc.find(1L);              // 命中 → 拿到改名后的对象            System.out.println("db hits = " + svc.dbHits());        }    }}

[db] SELECT * FROM product WHERE id=1[db] UPDATE product SET name=? WHERE id=1db hits = 1