使用Zookeeper来为你的程序加上Leader Election的功能。

阅读更多

ZooKeeper是Hadoop的正式子项目，它是一个针对大型分布式系统的可靠协调系统，提供的功能包括：配置维护、名字服务、分布式同步、组服务等。ZooKeeper的目标就是封装好复杂易出错的关键服务，将简单易用的接口和性能高效、功能稳定的系统提供给用户。

废话不多说，

package com.ericsson.threef.zookeeper;

import org.apache.zookeeper.*;
import org.apache.zookeeper.data.Stat;

import java.net.InetAddress;
import java.net.UnknownHostException;
import java.util.List;
import java.util.concurrent.CountDownLatch;

/**
 * Created by IntelliJ IDEA.
 * User: EDENYIN
 * Date: 1/4/11
 * Time: 3:13 PM
 * To change this template use File | Settings | File Templates.
 */
public class LeaderElection implements Watcher, Runnable {


    private String zookeeperConnectionString;
    private String rootPath;

    private ZooKeeper zk;

    private byte[] hostAddress;


    public LeaderElection(String zookeeperConnectionString, String rootPath) {
        this.zookeeperConnectionString = zookeeperConnectionString;
        this.rootPath = rootPath;
        try {
            hostAddress = InetAddress.getLocalHost().getHostAddress().getBytes();
        } catch (UnknownHostException e) {
            e.printStackTrace();
        }
        buildZK();
    }


    private void buildZK() {
        System.out.println("Build zk client");
        try {
            zk = new ZooKeeper(zookeeperConnectionString, 10000, this);
            Stat s = zk.exists(rootPath, false);
            if (s == null) {
                zk.create(rootPath, new byte[0], ZooDefs.Ids.OPEN_ACL_UNSAFE, CreateMode.PERSISTENT);
                zk.create(rootPath + "/ELECTION", new byte[0], ZooDefs.Ids.OPEN_ACL_UNSAFE, CreateMode.PERSISTENT);
            }
            String value = zk.create(rootPath + "/ELECTION/n_", hostAddress, ZooDefs.Ids.OPEN_ACL_UNSAFE, CreateMode.EPHEMERAL_SEQUENTIAL);
        } catch (Exception e) {
            e.printStackTrace();
            System.err.println("Error connect to zoo keeper");
        }
    }


    public void process(WatchedEvent event) {
        System.out.println(event);
        if (event.getState() == Event.KeeperState.Disconnected || event.getState() == Event.KeeperState.Expired) {
            System.out.println("Zookeeper connection timeout.");
            buildZK();
        }

    }

    public void run() {
            while (true) {
                try {
                    List children = zk.getChildren(rootPath + "/ELECTION", false);
                    String leaderPath = "Not found";
                    int minValue = -1;
                    for (int i=0;i 
 
 
 
附件是完整的maven项目。 
 
 
 
Reference: http://hadoop.apache.org/zookeeper/docs/r3.3.1/recipes.html#sc_leaderElection 
 
 
 
 
 
 
  
 
   
ZooKeeperLeaderEle.zip (33.1 KB)
 
   
下载次数: 81