Redis实现的分布式锁和分布式限流
随着现在分布式越来越普遍,分布式锁也十分常用,我的上一篇文章解释了使用zookeeper实现分布式锁(传送门),本次咱们说一下如何用Redis实现分布式锁和分布限流。
Redis有个事务锁,就是如下的命令,这个命令的含义是将一个value设置到一个key中,如果不存在将会赋值并且设置超时时间为30秒,如何这个key已经存在了,则不进行设置。
SET key value NX PX 30000
这个事务锁很好的解决了两个单独的命令,一个设置set key value nx,即该key不存在的话将对其进行设置,另一个是expire key seconds,设置该key的超时时间。我们可以想一下,如果这两个命令用程序单独使用会存在什么问题:
1. 如果一个set key的命令设置了key,然后程序异常了,expire时间没有设置,那么这个key会一直锁住。
2. 如果一个set key时出现了异常,但是直接执行了expire,过了一会儿之后另一个进行set key,还没怎么执行代码,结果key过期了,别的线程也进入了锁。
还有很多出问题的可能点,这里我们就不讨论了,下面咱们来看看如何实现吧。本文使用的Spring Boot 2.x + Spring data redis + Swagger +lombok + AOP + lua脚本。在实现的过程中遇到了很多问题,都一一解决实现了。依赖的POM文件如下:
<?xml version="1.0" encoding="UTF-8"?> <project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd"> <modelVersion>4.0.0</modelVersion> <parent> <groupId>org.springframework.boot</groupId> <artifactId>spring-boot-starter-parent</artifactId> <version>2.1.2.RELEASE</version> <relativePath/> <!-- lookup parent from repository --> </parent> <groupId>com.hqs</groupId> <artifactId>distributedlock</artifactId> <version>0.0.1-SNAPSHOT</version> <name>distributedlock</name> <description>Demo project for Spring Boot</description> <properties> <java.version>1.8</java.version> </properties> <dependencies> <dependency> <groupId>org.springframework.boot</groupId> <artifactId>spring-boot-starter-aop</artifactId> </dependency> <dependency> <groupId>org.springframework.boot</groupId> <artifactId>spring-boot-starter-web</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-devtools</artifactId> <scope>runtime</scope> </dependency> <dependency> <groupId>org.projectlombok</groupId> <artifactId>lombok</artifactId> <optional>true</optional> </dependency> <dependency> <groupId>org.springframework.boot</groupId> <artifactId>spring-boot-starter-test</artifactId> <scope>test</scope> </dependency> <dependency> <groupId>io.springfox</groupId> <artifactId>springfox-swagger-ui</artifactId> <version>2.9.2</version> </dependency> <dependency> <groupId>io.springfox</groupId> <artifactId>springfox-swagger2</artifactId> <version>2.9.2</version> <scope>compile</scope> </dependency> <dependency> <groupId>redis.clients</groupId> <artifactId>jedis</artifactId> <version>2.9.0</version> </dependency> </dependencies> <build> <plugins> <plugin> <groupId>org.springframework.boot</groupId> <artifactId>spring-boot-maven-plugin</artifactId> </plugin> </plugins> </build> </project>
使用了两个lua脚本,一个用于执行lock,另一个执行unlock。咱们简单看一下,lock脚本就是采用Redis事务执行的set nx px命令,其实还有set nx ex命令,这个ex命令是采用秒的方式进行设置过期时间,这个px是采用毫秒的方式设置过期时间。value需要使用一个唯一的值,这个值在解锁的时候需要判断是否一致,如果一致的话就进行解锁。这个也是官方推荐的方法。另外在lock的地方我设置了一个result,用于输出测试时的结果,这样就可以结合程序去进行debug了。
local expire = tonumber(ARGV[2]) local ret = redis.call('set', KEYS[1], ARGV[1], 'NX', 'PX', expire) local strret = tostring(ret) --用于查看结果,我本机获取锁成功后程序返回随机结果"table: 0x7fb4b3700fe0",否则返回"false" redis.call('set', 'result', strret) if strret == 'false' then return false else return true end
redis.call('del', 'result') if redis.call('get', KEYS[1]) == ARGV[1] then return redis.call('del', KEYS[1]) else return 0 end
来看下代码,主要写了两个方法,一个是用与锁另外一个是用于结解锁。这块需要注意的是使用RedisTemplate<String, String>,这块意味着key和value一定都是String的,我在使用的过程中就出现了一些错误。首先初始化两个脚本到程序中,然后调用执行脚本。
package com.hqs.distributedlock.lock; import lombok.extern.slf4j.Slf4j; import org.springframework.beans.factory.annotation.Autowired; import org.springframework.data.redis.core.RedisTemplate; import org.springframework.data.redis.core.script.RedisScript; import org.springframework.stereotype.Component; import java.util.Collections; @Slf4j @Component public class DistributedLock { //注意RedisTemplate用的String,String,后续所有用到的key和value都是String的 @Autowired private RedisTemplate<String, String> redisTemplate; @Autowired RedisScript<Boolean> lockScript; @Autowired RedisScript<Long> unlockScript; public Boolean distributedLock(String key, String uuid, String secondsToLock) { Boolean locked = false; try { String millSeconds = String.valueOf(Integer.parseInt(secondsToLock) * 1000); locked =redisTemplate.execute(lockScript, Collections.singletonList(key), uuid, millSeconds); log.info("distributedLock.key{}: - uuid:{}: - timeToLock:{} - locked:{} - millSeconds:{}", key, uuid, secondsToLock, locked, millSeconds); } catch (Exception e) { log.error("error", e); } return locked; } public void distributedUnlock(String key, String uuid) { Long unlocked = redisTemplate.execute(unlockScript, Collections.singletonList(key), uuid); log.info("distributedLock.key{}: - uuid:{}: - unlocked:{}", key, uuid, unlocked); } }
还有一个就是脚本定义的地方需要注意,返回的结果集一定是Long, Boolean,List, 一个反序列化的值。这块要注意。
package com.hqs.distributedlock.config; import com.sun.org.apache.xpath.internal.operations.Bool; import lombok.extern.slf4j.Slf4j; import org.springframework.beans.factory.annotation.Qualifier; import org.springframework.context.annotation.Bean; import org.springframework.context.annotation.Configuration; import org.springframework.core.io.ClassPathResource; import org.springframework.data.redis.core.script.DefaultRedisScript; import org.springframework.data.redis.core.script.RedisScript; import org.springframework.scripting.ScriptSource; import org.springframework.scripting.support.ResourceScriptSource; @Configuration @Slf4j public class BeanConfiguration { /** * The script resultType should be one of * Long, Boolean, List, or a deserialized value type. It can also be null if the script returns * a throw-away status (specifically, OK). * @return */ @Bean public RedisScript<Long> limitScript() { RedisScript redisScript = null; try { ScriptSource scriptSource = new ResourceScriptSource(new ClassPathResource("/scripts/limit.lua")); // log.info("script:{}", scriptSource.getScriptAsString()); redisScript = RedisScript.of(scriptSource.getScriptAsString(), Long.class); } catch (Exception e) { log.error("error", e); } return redisScript; } @Bean public RedisScript<Boolean> lockScript() { RedisScript<Boolean> redisScript = null; try { ScriptSource scriptSource = new ResourceScriptSource(new ClassPathResource("/scripts/lock.lua")); redisScript = RedisScript.of(scriptSource.getScriptAsString(), Boolean.class); } catch (Exception e) { log.error("error" , e); } return redisScript; } @Bean public RedisScript<Long> unlockScript() { RedisScript<Long> redisScript = null; try { ScriptSource scriptSource = new ResourceScriptSource(new ClassPathResource("/scripts/unlock.lua")); redisScript = RedisScript.of(scriptSource.getScriptAsString(), Long.class); } catch (Exception e) { log.error("error" , e); } return redisScript; } @Bean public RedisScript<Long> limitAnother() { DefaultRedisScript<Long> redisScript = new DefaultRedisScript<>(); redisScript.setScriptSource(new ResourceScriptSource(new ClassPathResource("/scripts/limit.lua"))); redisScript.setResultType(Long.class); return redisScript; } }
好了,这块就写好了,然后写好controller类准备测试。
@PostMapping("/distributedLock") @ResponseBody public String distributedLock(String key, String uuid, String secondsToLock, String userId) throws Exception{ // String uuid = UUID.randomUUID().toString(); Boolean locked = false; try { locked = lock.distributedLock(key, uuid, secondsToLock); if(locked) { log.info("userId:{} is locked - uuid:{}", userId, uuid); log.info("do business logic"); TimeUnit.MICROSECONDS.sleep(3000); } else { log.info("userId:{} is not locked - uuid:{}", userId, uuid); } } catch (Exception e) { log.error("error", e); } finally { if(locked) { lock.distributedUnlock(key, uuid); } } return "ok"; }
我也写了一个测试类,用于测试和输出结果, 使用100个线程,然后锁的时间设置10秒,controller里边需要休眠3秒模拟业务执行。
@Test public void distrubtedLock() { String url = "http://localhost:8080/distributedLock"; String uuid = "abcdefg"; // log.info("uuid:{}", uuid); String key = "redisLock"; String secondsToLive = "10"; for(int i = 0; i < 100; i++) { final int userId = i; new Thread(() -> { MultiValueMap<String, String> params = new LinkedMultiValueMap<>(); params.add("uuid", uuid); params.add("key", key); params.add("secondsToLock", secondsToLive); params.add("userId", String.valueOf(userId)); String result = testRestTemplate.postForObject(url, params, String.class); System.out.println("-------------" + result); } ).start(); } }
获取锁的地方就会执行do business logic, 然后会有部分线程获取到锁并执行业务,执行完业务的就会释放锁。
分布式锁就实现好了,接下来实现分布式限流。先看一下limit的lua脚本,需要给脚本传两个值,一个值是限流的key,一个值是限流的数量。获取当前key,然后判断其值是否为nil,如果为nil的话需要赋值为0,然后进行加1并且和limit进行比对,如果大于limt即返回0,说明限流了,如果小于limit则需要使用Redis的INCRBY key 1,就是将key进行加1命令。并且设置超时时间,超时时间是秒,并且如果有需要的话这个秒也是可以用参数进行设置。
--lua 下标从 1 开始 -- 限流 key local key = KEYS[1] -- 限流大小 local limit = tonumber(ARGV[1]) -- 获取当前流量大小 local curentLimit = tonumber(redis.call('get', key) or "0") if curentLimit + 1 > limit then -- 达到限流大小 返回 return 0; else -- 没有达到阈值 value + 1 redis.call("INCRBY", key, 1) -- EXPIRE后边的单位是秒 redis.call("EXPIRE", key, 10) return curentLimit + 1 end
执行limit的脚本和执行lock的脚本类似。
package com.hqs.distributedlock.limit; import lombok.extern.slf4j.Slf4j; import org.springframework.beans.factory.annotation.Autowired; import org.springframework.data.redis.core.RedisTemplate; import org.springframework.data.redis.core.script.RedisScript; import org.springframework.stereotype.Component; import java.util.Collections; /** * @author huangqingshi * @Date 2019-01-17 */ @Slf4j @Component public class DistributedLimit { //注意RedisTemplate用的String,String,后续所有用到的key和value都是String的 @Autowired private RedisTemplate<String, String> redisTemplate; @Autowired RedisScript<Long> limitScript; public Boolean distributedLimit(String key, String limit) { Long id = 0L; try { id = redisTemplate.execute(limitScript, Collections.singletonList(key), limit); log.info("id:{}", id); } catch (Exception e) { log.error("error", e); } if(id == 0L) { return false; } else { return true; } } }
接下来咱们写一个限流注解,并且设置注解的key和限流的大小:
package com.hqs.distributedlock.annotation; import java.lang.annotation.ElementType; import java.lang.annotation.Retention; import java.lang.annotation.RetentionPolicy; import java.lang.annotation.Target; /** * 自定义limit注解 * @author huangqingshi * @Date 2019-01-17 */ @Target(ElementType.METHOD) @Retention(RetentionPolicy.RUNTIME) public @interface DistriLimitAnno { public String limitKey() default "limit"; public int limit() default 1; }
然后对注解进行切面,在切面中判断是否超过limit,如果超过limit的时候就需要抛出异常exceeded limit,否则正常执行。
package com.hqs.distributedlock.aspect; import com.hqs.distributedlock.annotation.DistriLimitAnno; import com.hqs.distributedlock.limit.DistributedLimit; import lombok.extern.slf4j.Slf4j; import org.aspectj.lang.JoinPoint; import org.aspectj.lang.annotation.Aspect; import org.aspectj.lang.annotation.Before; import org.aspectj.lang.annotation.Pointcut; import org.aspectj.lang.reflect.MethodSignature; import org.springframework.beans.factory.annotation.Autowired; import org.springframework.context.annotation.EnableAspectJAutoProxy; import org.springframework.stereotype.Component; import java.lang.reflect.Method; /** * @author huangqingshi * @Date 2019-01-17 */ @Slf4j @Aspect @Component @EnableAspectJAutoProxy(proxyTargetClass = true) public class LimitAspect { @Autowired DistributedLimit distributedLimit; @Pointcut("@annotation(com.hqs.distributedlock.annotation.DistriLimitAnno)") public void limit() {}; @Before("limit()") public void beforeLimit(JoinPoint joinPoint) throws Exception { MethodSignature signature = (MethodSignature) joinPoint.getSignature(); Method method = signature.getMethod(); DistriLimitAnno distriLimitAnno = method.getAnnotation(DistriLimitAnno.class); String key = distriLimitAnno.limitKey(); int limit = distriLimitAnno.limit(); Boolean exceededLimit = distributedLimit.distributedLimit(key, String.valueOf(limit)); if(!exceededLimit) { throw new RuntimeException("exceeded limit"); } } }
因为有抛出异常,这里我弄了一个统一的controller错误处理,如果controller出现Exception的时候都需要走这块异常。如果是正常的RunTimeException的时候获取一下,否则将异常获取一下并且输出。
package com.hqs.distributedlock.util; import lombok.extern.slf4j.Slf4j; import org.springframework.http.HttpStatus; import org.springframework.web.bind.annotation.ControllerAdvice; import org.springframework.web.bind.annotation.ExceptionHandler; import org.springframework.web.bind.annotation.ResponseBody; import org.springframework.web.bind.annotation.ResponseStatus; import org.springframework.web.context.request.NativeWebRequest; import javax.servlet.http.HttpServletRequest; import java.util.HashMap; import java.util.Map; /** * @author huangqingshi * @Date 2019-01-17 * 统一的controller错误处理 */ @Slf4j @ControllerAdvice public class UnifiedErrorHandler { private static Map<String, String> res = new HashMap<>(2); @ExceptionHandler(value = Exception.class) @ResponseStatus(HttpStatus.OK) @ResponseBody public Object processException(HttpServletRequest req, Exception e) { res.put("url", req.getRequestURL().toString()); if(e instanceof RuntimeException) { res.put("mess", e.getMessage()); } else { res.put("mess", "sorry error happens"); } return res; } }
好了,接下来将注解写到自定义的controller上,limit的大小为10,也就是10秒钟内限制10次访问。
@PostMapping("/distributedLimit") @ResponseBody @DistriLimitAnno(limitKey="limit", limit = 10) public String distributedLimit(String userId) { log.info(userId); return "ok"; }
也是来一段Test方法来跑,老方式100个线程开始跑,只有10次,其他的都是limit。没有问题。