> For the complete documentation index, see [llms.txt](https://db-woms.gitbook.io/db/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://db-woms.gitbook.io/db/redis/redis-lua-readme/redis-lua-script-distributed-lock.md). # redis实现分布式锁 ## 1. 环境 ```groovy implementation 'org.springframework.boot:spring-boot-starter-data-redis' implementation 'org.springframework.boot:spring-boot-starter-web' compileOnly 'org.projectlombok:lombok' ``` ``` spring.redis.host=127.0.0.1 spring.redis.port=6379 ``` 测试redis是否链接成功 我在redis里面有1000块钱 ``` 127.0.0.1:6379> keys * 1) "money:woms" 127.0.0.1:6379> get money:woms "1000" ``` ```java import org.springframework.beans.factory.annotation.Autowired; import org.springframework.data.redis.core.StringRedisTemplate; import org.springframework.web.bind.annotation.GetMapping; import org.springframework.web.bind.annotation.RestController; @RestController public class TestController { @Autowired private StringRedisTemplate stringRedisTemplate; @GetMapping("name") public Object name() { String womsMoney = stringRedisTemplate.opsForValue().get("money:woms"); return womsMoney; } } ``` ```bash $ curl localhost:8080/name -s|jq 1000 ``` ## 为什么要用分布式锁 当多个线程同时对value进行操作的话，就会有问题 模拟10个并发并发请求，每个请求的目的是扣除我的100快，这样1000块钱被10个请求执行，应该最后是0元，但是结果呢？ 刚开始，我有1000 ``` 127.0.0.1:6379> set money:woms 1000 OK ``` 10个线程同时消费，每个线程消费100 ```java import java.util.concurrent.TimeUnit; import org.springframework.beans.factory.annotation.Autowired; import org.springframework.data.redis.core.StringRedisTemplate; import org.springframework.web.bind.annotation.GetMapping; import org.springframework.web.bind.annotation.RestController; @RestController public class TestController { @Autowired private StringRedisTemplate stringRedisTemplate; @GetMapping("name") public Object name() throws Exception { for(int i=0; i<10; i++) { Thread thread = new Thread(() -> { Integer womsMoneyInteger = Integer.parseInt(stringRedisTemplate.opsForValue().get("money:woms")); stringRedisTemplate.opsForValue().set("money:woms", String.valueOf(womsMoneyInteger - 100)); }); thread.start(); } TimeUnit.SECONDS.sleep(2); String now_money = stringRedisTemplate.opsForValue().get("money:woms"); return now_money; } } ``` 最后看结果，发现我还有900 ``` 127.0.0.1:6379> get money:woms "900" ``` 为什么会出现这种情况呢？就是第一个用户get的时候，get的是1000，第二个用户get的时候，还是1000，第一个用户-100，剩900，第二个用户在之前的查询的1000的基础上-100，相当于应该-200，这里只-100 ## redis实现分布式锁 ```java import java.util.concurrent.TimeUnit; import org.springframework.data.redis.core.StringRedisTemplate; import lombok.extern.slf4j.Slf4j; @Slf4j public class RedisLock { public boolean addLock(StringRedisTemplate redisTemplate, String key, String value, long timeout, int expire) { long nano = System.nanoTime();// 开始时间 while ((System.nanoTime() - nano) < timeout) { if (redisTemplate.opsForValue().setIfAbsent(key, value, expire, TimeUnit.SECONDS)) { log.info(Thread.currentThread().getName()); return true; } } return false; } public void unlock(StringRedisTemplate stringRedisTemplate, String key, String value) { String keyValue = stringRedisTemplate.opsForValue().get(key); if (value.equals(keyValue)) { log.info(Thread.currentThread().getName()); stringRedisTemplate.delete(key); } } } ``` ```java import java.util.UUID; import java.util.concurrent.TimeUnit; import org.springframework.beans.factory.annotation.Autowired; import org.springframework.data.redis.core.StringRedisTemplate; import org.springframework.web.bind.annotation.GetMapping; import org.springframework.web.bind.annotation.RestController; import lombok.extern.slf4j.Slf4j; @RestController @Slf4j public class TestController { @Autowired private StringRedisTemplate stringRedisTemplate; @GetMapping("name") public Object name() throws Exception { for (int i = 0; i < 10; i++) { Thread thread = new Thread(() -> { String uuid = UUID.randomUUID().toString(); RedisLock redisLock = new RedisLock(); if (redisLock.addLock(stringRedisTemplate, "money:woms:lock", uuid, 1000 * 1000 * 1000L, 180)) { log.info(Thread.currentThread().getName()); Integer womsMoneyInteger = Integer.parseInt(stringRedisTemplate.opsForValue().get("money:woms")); stringRedisTemplate.opsForValue().set("money:woms", String.valueOf(womsMoneyInteger - 100)); } redisLock.unlock(stringRedisTemplate, "money:woms:lock", uuid); }); thread.start(); } TimeUnit.SECONDS.sleep(10); String now_money = stringRedisTemplate.opsForValue().get("money:woms"); return now_money; } } ``` 其原理就是使用另一个key作为当前操作的锁，给另一个key设置值，如果设置成功，可以对目标key的值进行操作，操作完成以后把作为锁的key删除。 其他用户由于给目标key设置值的时候，锁key值已近存在，就会在超时时间内一直设置值，如果超时时间内设置成功，则可以进行操作，设置不成功，就不进行操作，其目的就一个就是在同一个时间只能有一个用户对值进行修改。 如果使用lua脚本怎么实现呢？ ```java import java.util.Arrays; import java.util.concurrent.TimeUnit; import javax.annotation.PostConstruct; import org.springframework.beans.factory.annotation.Autowired; import org.springframework.core.io.ClassPathResource; import org.springframework.data.redis.core.StringRedisTemplate; import org.springframework.data.redis.core.script.DefaultRedisScript; import org.springframework.scripting.support.ResourceScriptSource; import org.springframework.web.bind.annotation.GetMapping; import org.springframework.web.bind.annotation.RestController; import lombok.extern.slf4j.Slf4j; @RestController @Slf4j public class TestController { @Autowired private StringRedisTemplate stringRedisTemplate; private DefaultRedisScript getRedisScript; @PostConstruct public void init() { getRedisScript = new DefaultRedisScript(); getRedisScript.setResultType(String.class); getRedisScript.setScriptSource(new ResourceScriptSource(new ClassPathResource("222.lua"))); } @GetMapping("name") public Object name() throws Exception { for (int i = 0; i < 10; i++) { Thread thread = new Thread(() -> { log.info(Thread.currentThread().getName()); stringRedisTemplate.execute(getRedisScript, Arrays.asList("money:woms")); }); thread.start(); } TimeUnit.SECONDS.sleep(10); String now_money = stringRedisTemplate.opsForValue().get("money:woms"); return now_money; } } ``` ```lua local money = redis.call("get",KEYS[1]) redis.call("set", KEYS[1], money - 100) ``` lua本身就是原子性，可以不用任何锁，直接操作