zookeeper 分布式锁

利用zookeeper实现

当不少进程须要访问共享资源时，咱们能够经过zk来实现分布式锁。主要步骤是： 
1.创建一个节点，假如名为：lock 。节点类型为持久节点（PERSISTENT） 
2.每当进程须要访问共享资源时，会调用分布式锁的lock()或tryLock()方法得到锁，这个时候会在第一步建立的lock节点下创建相应的顺序子节点，节点类型为临时顺序节点（EPHEMERAL_SEQUENTIAL），经过组成特定的名字name+lock+顺序号。 
3.在创建子节点后，对lock下面的全部以name开头的子节点进行排序，判断刚刚创建的子节点顺序号是不是最小的节点，假如是最小节点，则得到该锁对资源进行访问。 
4.假如不是该节点，就得到该节点的上一顺序节点，并给该节点是否存在注册监听事件。同时在这里阻塞。等待监听事件的发生，得到锁控制权。 
5.当调用完共享资源后，调用unlock（）方法，关闭zk，进而能够引起监听事件，释放该锁。 
实现的分布式锁是严格的按照顺序访问的并发锁。

 

代码实现

import java.io.IOException;
import java.util.Collections;
import java.util.List;
import java.util.concurrent.CountDownLatch;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.TimeUnit;

import org.apache.zookeeper.CreateMode;
import org.apache.zookeeper.KeeperException;
import org.apache.zookeeper.WatchedEvent;
import org.apache.zookeeper.Watcher;
import org.apache.zookeeper.ZooDefs;
import org.apache.zookeeper.ZooKeeper;
import org.apache.zookeeper.data.Stat;

public class DistributedLock implements Watcher {

    ZooKeeper zk = null; // zookeeper原生api去实现一个分布式锁

    private String root = "/locks";

    private String myZonode; // 表示当前获取到的锁名称-也就是节点名称

    private String waitNode; // 表示当前等待的节点

    private CountDownLatch latch;

    private static final int SESSION_TIMEOUT = 10000; // 超时时间

    /**
     * 构造函数初始化
     * 
     * @param config
     *            表示zookeeper链接串
     */
    public DistributedLock(String config) {
        try {
            zk = new ZooKeeper(config, SESSION_TIMEOUT, this);
            Stat stat = zk.exists(root, false); // 判断是否是已经存在locks节点，不须要监听root节点
            if (stat == null) { // 若是不存在，则建立根节点
                zk.create(root, new byte[0], ZooDefs.Ids.OPEN_ACL_UNSAFE,
                        CreateMode.PERSISTENT);
            }
        } catch (IOException e) {
            e.printStackTrace();
        } catch (InterruptedException e) {
            e.printStackTrace();
        } catch (KeeperException e) {
            e.printStackTrace();
        }
    }

    public void process(WatchedEvent event) {
        if (this.latch != null) { // 若是计数器不为空话话，释放计数器锁
            this.latch.countDown();
        }
    }

    /**
     * 获取锁的方法
     */
    public boolean lock(String name) {
        if (tryLock(name)) {
            return true;
        }
        try {
            return waitLock(waitNode, SESSION_TIMEOUT);
        } catch (KeeperException e) {
            e.printStackTrace();
        } catch (InterruptedException e) {
            e.printStackTrace();
        }
        return false;
    }

    /**
     * 释放锁操做的方法
     */
    public void unlock() {
        try {
            zk.delete(myZonode, -1);
            myZonode = null;
            zk.close();
        } catch (InterruptedException e) {
            e.printStackTrace();
        } catch (KeeperException e) {
            e.printStackTrace();
        }

    }

    private boolean tryLock(String name) {
        String splitStr = name; // lock_0000000001
        try {
            // 建立一个有序的临时节点，赋值给myznode
            myZonode = zk.create(root + "/" + splitStr, new byte[0],
                    ZooDefs.Ids.OPEN_ACL_UNSAFE,
                    CreateMode.EPHEMERAL_SEQUENTIAL);
            List<String> subNodes = zk.getChildren(root, false);
            Collections.sort(subNodes); // 讲全部的子节点排序
            if (myZonode.equals(root + "/" + subNodes.get(0))) {
                // 当前客户端建立的临时有序节点是locks下节点中的最小的节点，表示当前的客户端可以获取到锁
                return true;
            }
            // 不然的话,监听比本身小的节点 locks/lock_0000000003
            String subMyZnode = myZonode
                    .substring((myZonode.lastIndexOf("/") + 1));
            waitNode = subNodes.get(Collections.binarySearch(subNodes,
                    subMyZnode) - 1);// 获取比当前节点小的节点
        } catch (KeeperException e) {
            e.printStackTrace();
        } catch (InterruptedException e) {
            e.printStackTrace();
        }
        return false;
    }

    private boolean waitLock(String lower, long waitTime)
            throws KeeperException, InterruptedException {
        Stat stat = zk.exists(root + "/" + lower, true); // 获取节点状态，并添加监听
        if (stat != null) {
            this.latch = new CountDownLatch(1); // 实例化计数器，让当前的线程等待
            this.latch.await(waitTime, TimeUnit.MILLISECONDS);
            this.latch = null;
        }
        return true;
    }

    public static int count =10;
    
    public static void main(String[] args) {
        ExecutorService executorService = Executors.newCachedThreadPool();

        for (int i = 0; i < 10; i++) {
            Runnable runnable = new Runnable() {

                public void run() {
                    try {
                        DistributedLock distributeLockDemo = new DistributedLock(
                                "127.0.0.1:2181");
                        boolean lock = distributeLockDemo.lock("test_");
                        if (lock) {
                            System.out.println(count--);
                            distributeLockDemo.unlock();
                        }

                    } catch (Exception e) {
                        e.printStackTrace();
                    }

                }
            };

            executorService.execute(runnable);
        }
        executorService.shutdown();

    }

}