分布式处理中，总会存在多个服务节点同时工作，并且节点数量会随着网络规模的变化而动态增减，服务节点也有可能发生宕机与恢复。面对着动态增减的服务节点，我们如何保证客户请求被服务器正确处理呢。我们可以通过zookeeper临时节点创建与自动删除来掌握服务节点的动态增减。

ignite分布式缓存支持使用zookeeper发现ignite节点的增减，这正是zookeeper管理服务节点的一个典型应用场景。我们来看看关键代码

// 关键方法，创建包含自增长id名称的目录，这个方法支持了分布式锁的实现        // 四个参数：        // 1、目录名称 2、目录文本信息         // 3、文件夹权限，Ids.OPEN_ACL_UNSAFE表示所有权限         // 4、目录类型，CreateMode.EPHEMERAL_SEQUENTIAL表示创建临时目录，session断开连接则目录自动删除        String createdPath = zk.create(                "/" + clusterNode + "/" + serverNode,                 address.getBytes("utf-8"),                 Ids.OPEN_ACL_UNSAFE,                 CreateMode.EPHEMERAL_SEQUENTIAL);

采用CreateMode.EPHEMERAL_SEQUENTIAL模式创建临时节点，可以支持服务节点的实时管理。没错，这个模式和《》中创建有序节点支持分布式共享锁是一致的。EPHEMERAL_SEQUENTIAL表示创建有序的临时目录节点，zookeeper客户端创建临时节点后，只要session断开，该临时节点会自动删除。

所以，服务器在zookeeper上创建一个临时目录节点，通过节点事件监听我们可以知道服务器已经加入到服务网络中，监听到临时目录节点删除事件，我们可以知道该节点对应的服务器已经脱离服务网络。下面我们看看具体代码

1、 服务器启动后在zookeeper创建临时目录

package com.coshaho.learn.zookeeper;import org.apache.zookeeper.CreateMode;import org.apache.zookeeper.WatchedEvent;import org.apache.zookeeper.Watcher;import org.apache.zookeeper.ZooDefs.Ids;import org.apache.zookeeper.ZooKeeper;/** *  * 服务节点启动后注册到zookeeper * @author coshaho * */public class AppServer extends Thread{    private String clusterNode = "Locks";    private String serverNode = "mylock";    private String serverName;    private long sleepTime;    public void run()    {        try         {            connectZookeeper(serverName);        }         catch (Exception e)         {            e.printStackTrace();        }    }    public void connectZookeeper(String address) throws Exception     {        ZooKeeper zk = new ZooKeeper("192.168.1.104:12181", 5000, new Watcher()         {            public void process(WatchedEvent event)             {}        });        // 关键方法，创建包含自增长id名称的目录，这个方法支持了分布式锁的实现        // 四个参数：        // 1、目录名称 2、目录文本信息         // 3、文件夹权限，Ids.OPEN_ACL_UNSAFE表示所有权限         // 4、目录类型，CreateMode.EPHEMERAL_SEQUENTIAL表示创建临时目录，session断开连接则目录自动删除        String createdPath = zk.create(                "/" + clusterNode + "/" + serverNode,                 address.getBytes("utf-8"),                 Ids.OPEN_ACL_UNSAFE,                 CreateMode.EPHEMERAL_SEQUENTIAL);        System.out.println("create: " + createdPath);        Thread.sleep(sleepTime);    }        public AppServer(String serverName, long sleepTime)    {        this.serverName = serverName;        this.sleepTime = sleepTime;    }}

2、 节点管理服务监听zookeeper临时目录节点创建删除事件

package com.coshaho.learn.zookeeper;import java.util.ArrayList;import java.util.List;import org.apache.zookeeper.WatchedEvent;import org.apache.zookeeper.Watcher;import org.apache.zookeeper.Watcher.Event.EventType;import org.apache.zookeeper.ZooKeeper;/** *  * 客户端注册监听server节点变化 * @author coshaho * */public class AppMaster {    private String clusterNode = "Locks";    private ZooKeeper zk;    private volatile List
     
       serverList;    public void connectZookeeper() throws Exception     {        // 注册全局默认watcher        zk = new ZooKeeper("192.168.1.104:12181", 5000, new Watcher()         {            public void process(WatchedEvent event)             {                if (event.getType() == EventType.NodeChildrenChanged                         && ("/" + clusterNode).equals(event.getPath()))                 {                    try                     {                        updateServerList();                    }                     catch (Exception e)                     {                        e.printStackTrace();                    }                }            }        });        updateServerList();    }    private void updateServerList() throws Exception     {        List
      
        newServerList = new ArrayList
       
        ();        // watcher注册后，只能监听事件一次，参数true表示继续使用默认watcher监听事件        List
        
          subList = zk.getChildren("/" + clusterNode, true);        for (String subNode : subList)        {            // 获取节点数据            byte[] data = zk.getData("/" + clusterNode + "/" + subNode, false, null);            newServerList.add(new String(data, "utf-8"));        }        serverList = newServerList;        System.out.println("server list updated: " + serverList);    }    public static void main(String[] args) throws Exception     {        AppMaster ac = new AppMaster();        ac.connectZookeeper();        Thread.sleep(Long.MAX_VALUE);    }}
        
       
      
     

3、 启动两个服务器

package com.coshaho.learn.zookeeper;public class Server1 {    public static void main(String[] args) throws Exception     {        AppServer server1 = new AppServer("Server1", 5000);        server1.start();    }}package com.coshaho.learn.zookeeper;public class Server2 {    public static void main(String[] args) throws Exception     {        AppServer server1 = new AppServer("Server2", 10000);        server1.start();    }}

4、 运行结果

log4j:WARN No appenders could be found for logger (org.apache.zookeeper.ZooKeeper).log4j:WARN Please initialize the log4j system properly.log4j:WARN See http://logging.apache.org/log4j/1.2/faq.html#noconfig for more info.server list updated: []server list updated: [Server1]server list updated: [Server2, Server1]server list updated: [Server2]server list updated: []