之前寫過一篇文章《如何在springcloud分散式系統中實現分散式鎖？》，由於自己僅僅是閱讀了相關的書籍，和查閱了相關的資料，就認為那樣的是可行的。那篇文章實現的大概思路是用setNx命令和setEx配合使用。 setNx是一個耗時操作，因為它需要查詢這個鍵是否存在，就算redis的百萬的qps，在高併發的場景下，這種操作也是有問題的。關於redis實現分散式鎖，redis官方推薦使用redlock。

一、redlock簡介

在不同程序需要互斥地訪問共享資源時，分散式鎖是一種非常有用的技術手段。實現高效的分散式鎖有三個屬性需要考慮：

• 安全屬性：互斥，不管什麼時候，只有一個客戶端持有鎖
• 效率屬性A:不會死鎖
• 效率屬性B：容錯，只要大多數redis節點能夠正常工作，客戶端端都能獲取和釋放鎖。

Redlock是redis官方提出的實現分散式鎖管理器的演算法。這個演算法會比一般的普通方法更加安全可靠。關於這個演算法的討論可以看下官方文件。

二、怎麼用java使用 redlock

在pom檔案引入redis和redisson依賴：

<!-- redis-->
        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>
 
spring-boot-starter-data-redis</artifactId>
        </dependency>
        <!-- redisson-->
        <dependency>
            <groupId>org.redisson</groupId>
            <artifactId>redisson</artifactId>
            <version>3.3.2</version>
        </dependency
 
>

AquiredLockWorker介面類，，主要是用於獲取鎖後需要處理的邏輯：

/**
 * Created by fangzhipeng on 2017/4/5.
 * 獲取鎖後需要處理的邏輯
 */
public interface AquiredLockWorker<T> {
     T invokeAfterLockAquire() throws Exception;
}

DistributedLocker 獲取鎖管理類：

/**
 * Created by fangzhipeng on 2017/4/5.
 * 獲取鎖管理類
 */
public interface DistributedLocker {

     /**
      * 獲取鎖
      * @param resourceName  鎖的名稱
      * @param worker 獲取鎖後的處理類
      * @param <T>
      * @return 處理完具體的業務邏輯要返回的資料
      * @throws UnableToAquireLockException
      * @throws Exception
      */
     <T> T lock(String resourceName, AquiredLockWorker<T> worker) throws UnableToAquireLockException, Exception;

     <T> T lock(String resourceName, AquiredLockWorker<T> worker, int lockTime) throws UnableToAquireLockException, Exception;

}

UnableToAquireLockException ，不能獲取鎖的異常類：

/**
 * Created by fangzhipeng on 2017/4/5.
 * 異常類
 */
public class UnableToAquireLockException extends RuntimeException {

    public UnableToAquireLockException() {
    }

    public UnableToAquireLockException(String message) {
        super(message);
    }

    public UnableToAquireLockException(String message, Throwable cause) {
        super(message, cause);
    }
}

RedissonConnector 連線類：

/**
 * Created by fangzhipeng on 2017/4/5.
 * 獲取RedissonClient連線類
 */
@Component
public class RedissonConnector {
    RedissonClient redisson;
    @PostConstruct
    public void init(){
        redisson = Redisson.create();
    }

    public RedissonClient getClient(){
        return redisson;
    }

}

RedisLocker 類，實現了DistributedLocker：

import org.redisson.api.RLock;
import org.redisson.api.RedissonClient;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.stereotype.Component;
import java.util.concurrent.TimeUnit;

/**
 * Created by fangzhipeng on 2017/4/5.
 */
@Component
public class RedisLocker  implements DistributedLocker{

    private final static String LOCKER_PREFIX = "lock:";

    @Autowired
    RedissonConnector redissonConnector;
    @Override
    public <T> T lock(String resourceName, AquiredLockWorker<T> worker) throws InterruptedException, UnableToAquireLockException, Exception {

        return lock(resourceName, worker, 100);
    }

    @Override
    public <T> T lock(String resourceName, AquiredLockWorker<T> worker, int lockTime) throws UnableToAquireLockException, Exception {
        RedissonClient redisson= redissonConnector.getClient();
        RLock lock = redisson.getLock(LOCKER_PREFIX + resourceName);
      // Wait for 100 seconds seconds and automatically unlock it after lockTime seconds
        boolean success = lock.tryLock(100, lockTime, TimeUnit.SECONDS);
        if (success) {
            try {
                return worker.invokeAfterLockAquire();
            } finally {
                lock.unlock();
            }
        }
        throw new UnableToAquireLockException();
    }
}

測試類：

  @Autowired
    RedisLocker distributedLocker;
    @RequestMapping(value = "/redlock")
    public String testRedlock() throws Exception{

        CountDownLatch startSignal = new CountDownLatch(1);
        CountDownLatch doneSignal = new CountDownLatch(5);
        for (int i = 0; i < 5; ++i) { // create and start threads
            new Thread(new Worker(startSignal, doneSignal)).start();
        }
        startSignal.countDown(); // let all threads proceed
        doneSignal.await();
        System.out.println("All processors done. Shutdown connection");
        return "redlock";
    }

     class Worker implements Runnable {
        private final CountDownLatch startSignal;
        private final CountDownLatch doneSignal;

        Worker(CountDownLatch startSignal, CountDownLatch doneSignal) {
            this.startSignal = startSignal;
            this.doneSignal = doneSignal;
        }

        public void run() {
            try {
                startSignal.await();
                distributedLocker.lock("test",new AquiredLockWorker<Object>() {

                    @Override
                    public Object invokeAfterLockAquire() {
                        doTask();
                        return null;
                    }

                });
            }catch (Exception e){

            }
        }

        void doTask() {
            System.out.println(Thread.currentThread().getName() + " start");
            Random random = new Random();
            int _int = random.nextInt(200);
            System.out.println(Thread.currentThread().getName() + " sleep " + _int + "millis");
            try {
                Thread.sleep(_int);
            } catch (InterruptedException e) {
                e.printStackTrace();
            }
            System.out.println(Thread.currentThread().getName() + " end");
            doneSignal.countDown();
        }
    }

執行測試類：

Thread-48 start 
Thread-48 sleep 99millis 
Thread-48 end 
Thread-49 start 
Thread-49 sleep 118millis 
Thread-49 end 
Thread-52 start 
Thread-52 sleep 141millis 
Thread-52 end 
Thread-50 start 
Thread-50 sleep 28millis 
Thread-50 end 
Thread-51 start 
Thread-51 sleep 145millis 
Thread-51 end

從執行結果上看，在非同步任務的情況下，確實是獲取鎖之後才能執行執行緒。不管怎麼樣，這是redis官方推薦的一種方案，可靠性比較高。有什麼問題歡迎留言。

三、參考資料