基于redis的分布式锁的分析与实践

SET resource_name my_random_value NX PX 30000
 手动删除锁（Lua脚本）：

if redis.call("get",KEYS[1]) == ARGV[1] then
    return redis.call("del",KEYS[1])
else
    return 0
end

package com.caiya.cms.web.component;

import com.caiya.cache.CacheException;
import com.caiya.cache.redis.JedisCache;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

import java.util.Objects;
import java.util.concurrent.TimeUnit;

/**
 * redis实现分布式锁
 * 可实现特性:
 * 1、使多线程无序排队获取和释放锁;
 * 2、丢弃未成功获得锁的线程处理;
 * 3、只释放线程本身加持的锁;
 * 4、避免死锁
 *
 * @author wangnan
 * @since 1.0
 */
public final class RedisLock {

    private static final Logger logger = LoggerFactory.getLogger(RedisLock.class);

    /**
     * 尝试加锁(仅一次)
     *
     * @param lockKey       锁key
     * @param lockValue     锁value
     * @param expireSeconds 锁超时时间(秒)
     * @return 是否加锁成功
     * @throws CacheException
     */
    public static boolean tryLock(String lockKey, String lockValue, long expireSeconds) throws CacheException {
        JedisCache jedisCache = JedisCacheFactory.getInstance().getJedisCache();
        try {
            String response = jedisCache.set(lockKey, lockValue, "nx", "ex", expireSeconds);
            return Objects.equals(response, "OK");
        } finally {
            jedisCache.close();
        }
    }

    /**
     * 加锁(指定最大尝试次数范围内)
     *
     * @param lockKey       锁key
     * @param lockValue     锁value
     * @param expireSeconds 锁超时时间(秒)
     * @param tryTimes      最大尝试次数
     * @param sleepMillis   每两次尝试之间休眠时间(毫秒)
     * @return 是否加锁成功
     * @throws CacheException
     */
    public static boolean lock(String lockKey, String lockValue, long expireSeconds, int tryTimes, long sleepMillis) throws CacheException {
        boolean result;
        int count = 0;
        do {
            count++;
            result = tryLock(lockKey, lockValue, expireSeconds);
            try {
                TimeUnit.MILLISECONDS.sleep(sleepMillis);
            } catch (InterruptedException e) {
                logger.error(e.getMessage(), e);
            }
        } while (!result && count <= tryTimes);
        return result;
    }

    /**
     * 释放锁
     *
     * @param lockKey   锁key
     * @param lockValue 锁value
     */
    public static void unlock(String lockKey, String lockValue) {
        JedisCache jedisCache = JedisCacheFactory.getInstance().getJedisCache();
        try {
            String luaScript = "if redis.call('get',KEYS[1]) == ARGV[1] then return redis.call('del',KEYS[1]) else return 0 end";
            Object result = jedisCache.eval(luaScript, 1, lockKey, lockValue);
//            Objects.equals(result, 1L);
        } catch (Exception e) {
            logger.error(e.getMessage(), e);
        } finally {
            jedisCache.close();
        }
//        return false;
    }


    private RedisLock() {
    }

}

        ...
        String lockKey = Constant.DEFAULT_CACHE_NAME + ":addItemApply:" + applyPriceDTO.getItemId() + "_" + applyPriceDTO.getSupplierId();// 跟业务相关的唯一拼接键
        String lockValue = Constant.DEFAULT_CACHE_NAME + ":" + System.getProperty("JvmId") + ":" + Thread.currentThread().getName() + ":" + System.currentTimeMillis();// 生成集群环境中的唯一值
        boolean locked = RedisLock.tryLock(lockKey, lockValue, 100);// 只尝试一次，在本次处理过程中直接拒绝其他线程的请求
        if (!locked) {
            throw new IllegalAccessException("您的操作太频繁了,休息一下再来吧~");
        }
        try {
            // 开始处理核心业务逻辑
            Item item = itemService.queryItemByItemId(applyPriceDTO.getItemId());
            ...
            ...
        } finally {
            RedisLock.unlock(lockKey, lockValue);// 在finally块中释放锁
        }

        ...
        String lockKey = Constant.DEFAULT_CACHE_NAME + ":addItemApply:" + applyPriceDTO.getItemId() + "_" + applyPriceDTO.getSupplierId();
        String lockValue = Constant.DEFAULT_CACHE_NAME + ":机器编号:" + Thread.currentThread().getName() + ":" + System.currentTimeMillis();
        boolean locked = RedisLock.lock(lockKey, lockValue, 100, 20, 100);// 非公平锁，无序竞争（这里需要合理根据业务处理情况设置最大尝试次数和每次休眠时间）
        if (!locked) {
            throw new IllegalAccessException("系统太忙,本次操作失败");// 一般来说，不会走到这一步；如果真的有这种情况，并且在合理设置锁尝试次数和等待响应时间之后仍然处理不过来，可能需要考虑优化程序响应时间或者用消息队列排队执行了
        }

        try {
            // 开始处理核心业务逻辑
            Item item = itemService.queryItemByItemId(applyPriceDTO.getItemId());
            ...
            ...
        } finally {
            RedisLock.unlock(lockKey, lockValue);
        }
        ...