ZooKeeper实践：集群管理

Posted 2020-09-18 猿祖

前言：

       随着业务的扩大，用户的增多，访问量的增加，单机模式已经不能支撑，从而出现了从单机模式->垂直应用模式->集群模式，集群模式诞生了，伴随着一堆问题也油然而生，Master怎么选举，机器故障及时移除集群，添加机器了如何及时的感应到，Zookeeper不仅能维护当前的集群服务状态，还能及时的选出master，它们的实现方式都是在Zookeeper上面注册一个EPHEMERAL目录节点，在创建目录的父目录上面调用getChildren(String path,boolean watch)设置watch为true，注册watcher事件，由于是临时节点，当服务器出现问题造成session丢失，相对应的目录节点随之删除，Children节点发生变化，就会调用前面注册的watcher，使集群内的每个节点都得到新的集群信息。反之，新增节点也是这样。

注：master选举，不同之处是注册一个EPHEMERAL_SEQUENTIAL 目录节点，我们可以选择最小编号为master，实现了动态选择master，避免master出现单点故障

一、架构体系

     

二：模拟代码

      监控类： 

package com.zk.config.manager;

import java.io.IOException;
import java.util.Collections;
import java.util.List;
import java.util.concurrent.CountDownLatch;
import org.apache.zookeeper.CreateMode;
import org.apache.zookeeper.KeeperException;
import org.apache.zookeeper.WatchedEvent;
import org.apache.zookeeper.Watcher;
import org.apache.zookeeper.ZooKeeper;
import org.apache.zookeeper.Watcher.Event.EventType;
import org.apache.zookeeper.Watcher.Event.KeeperState;
import org.apache.zookeeper.ZooDefs.Ids;

public class AppRegister {

	private CountDownLatch connectedSemaphore = new CountDownLatch(1);
	private ZooKeeper zooKeeper;
	private Object lock = new Object();
	private String ip;

	private String rootConfig;

	public AppRegister(String ip, String root) {
		this.ip = ip;
		this.rootConfig = root;
		this.zooKeeper = connectZookeeper();

	}

	public ZooKeeper connectZookeeper() {
		synchronized (lock) {
			if (zooKeeper == null) {
				try {
					zooKeeper = new ZooKeeper("192.168.1.1:2181", 5000, new Watcher() {
						public void process(WatchedEvent event) {
							if (KeeperState.SyncConnected == event.getState()) {
								if (EventType.None == event.getType() && null == event.getPath()) {
									try {
										connectedSemaphore.countDown();
										zooKeeper.getChildren(rootConfig, true);
										zooKeeper.create(rootConfig + "/" + ip, ip.getBytes(), Ids.OPEN_ACL_UNSAFE,
												CreateMode.EPHEMERAL_SEQUENTIAL);
									} catch (KeeperException e) {
										e.printStackTrace();
									} catch (InterruptedException e) {
										e.printStackTrace();
									}
								} else if (EventType.NodeChildrenChanged == event.getType()) {
									try {
										List<String> childrenList = zooKeeper.getChildren(rootConfig, true);
										System.out.println("节点变化了：" + childrenList);
										Collections.sort(childrenList);
										System.out.println("我是master了：" + childrenList.get(0));
									} catch (KeeperException e) {
										e.printStackTrace();
									} catch (InterruptedException e) {
										e.printStackTrace();
									}

								}

							}
						}
					}

					);
					connectedSemaphore.await();
				} catch (IOException e) {
					e.printStackTrace();
				} catch (InterruptedException e) {
					e.printStackTrace();
				}

			}
		}
		return zooKeeper;
	}
}

 模拟程序：

package com.zk.config.manager;

public class AppMonitor {
	private final static String rootConfig = "/AppCluster";
	public static void main(String[] args) throws InterruptedException {
		
		/**
		 * 新增机器,选举master
		 */
		addHost();
		
		Thread.sleep(Integer.MAX_VALUE);
	}
	
	private static void  addHost()
	{
		new Thread(new Runnable() {
			public void run() {
				new AppRegister("192.168.1.1", rootConfig);
			}
		}).start();

		new Thread(new Runnable() {
			public void run() {
				try {

					Thread.sleep(2000);
					new AppRegister("192.168.1.2", rootConfig);
				} catch (InterruptedException e) {
					// TODO Auto-generated catch block
					e.printStackTrace();
				}
			}
		}).start();
	
	}

}

 运行结果：

节点变化了：[192.168.1.10000000029]
我是master了：192.168.1.10000000029
节点变化了：[192.168.1.20000000030, 192.168.1.10000000029]
我是master了：192.168.1.10000000029
节点变化了：[192.168.1.20000000030, 192.168.1.10000000029]
我是master了：192.168.1.10000000029