Zookeeper简介及使用

⼀、Zookeeper简介

1、zookeeper简介 动物管理员

Apache ZooKeeper致⼒于开发和维护开源服务器，实现⾼度可靠的分布式协调。2、什么是ZooKeeper？

ZooKeeper是⼀种集中式服务，⽤于维护配置信息，命名，提供分布式同步和提供组服务。所有这 些类型的服务都以分布式应⽤程序的某种形式使⽤。每次实施它们都需要做很多⼯作来修复不可避 免的错误和竞争条件。由于难以实现这些类型的服务，应⽤程序最初通常会吝啬它们，这使得它们 在变化的情况下变得脆弱并且难以管理。即使正确完成，这些服务的不同实现也会在部署应⽤程序 时导致管理复杂性。3、zookeeper功能 （1）存储数据 （2）监听

4、zookeeper⼯作机制

基于观察者模式设计的分布式服务管理框架5、zookeeper的存储结构 ⽬录树结构

6、zookeeper应⽤场景 （1）集群统⼀配置管理 （2）集群统⼀命名服务 （3）集群统⼀管理

（4）服务器的动态上下线感知 （5）负载均衡

7、安装zookeeper单机版

（1）下载安装包 zookeeper-3.4.10.tar.gz （2）上传安装包到Linux alt+p（SecureCRT中） （3）解压

tar -zxvf zookeeper-3.4.10.tar.gz （4）修改配置⽂件

进⼊zookeeper安装⽬录下的conf⽂件夹下 mv zoo_sample.cfg zoo.cfg vi zoo.cfg

dataDir=/root/hd/zookeeper-3.4.10/zkData （5）启动zk

bin/zkServer.sh start （6）查看状态

bin/zkServer.sh status （7）启动客户端 bin/zkCli.sh

8、完全分布式安装

（1）下载安装包 zookeeper-3.4.10.tar.gz （2）上传安装包到Linux alt+p（SecureCRT中） （3）解压

tar -zxvf zookeeper-3.4.10.tar.gz （4）修改配置⽂件

进⼊zookeeper安装⽬录下的conf⽂件夹下 mv zoo_sample.cfg zoo.cfg vi zoo.cfg

dataDir=/root/hd/zookeeper-3.4.10/zkData ###############cluster############### server.1=hd09-1:2888:3888 server.2=hd09-2:2888:3888 server.3=hd09-3:2888:3888

其中server.后⾯的数字1、2、3分别是机器hd09-1、hd09-2、hd09-3中zookeeper-3.4.10/zkData/myid⽂件中的值 （5）添加⽂件myid

$cd zookeeper-3.4.10/zkData $touch myid

（6）添加内容在myid为1 $ vi myid 1

（7）发送zookeeper⽂件到其它机器 $ scp -r zookeeper-3.4.10 hd09-2:$PWD $ scp -r zookeeper-3.4.10 hd09-3:$PWD （8）修改myid依次为2 3 （9）修改环境变量

vi /etc/profile

export ZOOKEEPER_HOME=/root/hd/zookeeper3.4.10 export PATH=$PATH:$ZOOKEEPER_HOME/bin （10）⽣效环境变量 source /etc/profile （11）启动zookeeper zkServer.sh start （12）查看状态 zkServer.sh status （13）关闭zookeeper zkServer.sh stop

9、客户端的命令⾏操作（1）启动客户端 zkCli.sh（2）查看帮助 help

（3）查看当前znode所包含的内容 ls /

（4）创建节点 create /hello 18（5）创建短暂znode create -e /haha tom（6）创建带序号znode create -s /bigdata tom（7）创建短暂带序号

create -e -s /bigdata tom（8）查看此节点的详细信息 ls2 /

（9）获得节点值监听 get /hello watch（10）监听路径 ls / watch

（11）修改znode数据 set /hello iiiii（12）删除节点 delete /hello（13）递归删除 rmr /delireba

（14）查看节点状态信息 stat /

⼆、Zookeeper⼯作机制1、Zookeeper⼯作机制

2、Zookeeper⽬录结构

3、Zookeeper选举机制

三、Zookeeper简单客户端APIpackage com.css.zk;import java.io.IOException;import java.util.List;

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.Ids;import org.apache.zookeeper.ZooKeeper;import org.apache.zookeeper.data.Stat;import org.junit.Before;import org.junit.Test;public class ZkClient {

private String connectString = \"192.168.146.132:2181,192.168.146.133:2181,192.168.146.134:2181\"; private int sessionTimeout = 3000;

ZooKeeper zkCli = null;

// 初始化客户端 @Before

public void init() throws IOException {

zkCli = new ZooKeeper(connectString, sessionTimeout, new Watcher() { // 回调监听 @Override

public void process(WatchedEvent event) {

// System.out.println(event.getPath() + \"\\" + event.getState() + \"\\" + event.getType()); try {

List children = zkCli.getChildren(\"/\ for (String c : children) { // System.out.println(c); }

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

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

// 创建⼦节点 @Test

public void createZnode() throws KeeperException, InterruptedException {

String path = zkCli.create(\"/hello\ System.out.println(path); }

// 获取⼦节点 @Test

public void getChild() throws KeeperException, InterruptedException { List children = zkCli.getChildren(\"/\ for (String c : children) { System.out.println(c); }

Thread.sleep(Long.MAX_VALUE); }

// 删除节点 @Test

public void rmChildData() throws KeeperException, InterruptedException { // byte[] data = zkCli.getData(\"/bbq\ // System.out.println(new String(data)); zkCli.delete(\"/hello\ }

// 修改数据 @Test

public void setData() throws KeeperException, InterruptedException { zkCli.setData(\"/hello\ }

// 判断节点是否存在 @Test

public void testExist() throws KeeperException, InterruptedException { Stat exists = zkCli.exists(\"/hello\

System.out.println(exists == null ? \"not exists\" : \"exists\"); }}

四、Zoopeeper监听API1、监听单节点内容package com.css.zk;import java.io.IOException;

import org.apache.zookeeper.KeeperException;

import org.apache.zookeeper.WatchedEvent;import org.apache.zookeeper.Watcher;import org.apache.zookeeper.ZooKeeper;

// 监听单节点内容

public class WatchDemo {

public static void main(String[] args) throws IOException, KeeperException, InterruptedException {

ZooKeeper zkCli = new ZooKeeper(\"192.168.146.132:2181,192.168.146.133:2181,192.168.146.134:2181\ new Watcher() { // 监听回调 @Override

public void process(WatchedEvent event) { } });

byte[] data = zkCli.getData(\"/hello\ // 监听的具体内容 @Override

public void process(WatchedEvent event) {

System.out.println(\"监听路径为：\" + event.getPath()); System.out.println(\"监听的类型为：\" + event.getType()); System.out.println(\"监听被修改了\"); } }, null);

System.out.println(new String(data)); Thread.sleep(Long.MAX_VALUE); }}

2、监听⽬录package com.css.zk;import java.io.IOException;import java.util.List;

import org.apache.zookeeper.KeeperException;import org.apache.zookeeper.WatchedEvent;import org.apache.zookeeper.Watcher;import org.apache.zookeeper.ZooKeeper;// 监听⽬录

public class WatchDemo1 { static List children = null;

public static void main(String[] args) throws IOException, KeeperException, InterruptedException {

ZooKeeper zkCli = new ZooKeeper(\"192.168.146.132:2181,192.168.146.133:2181,192.168.146.134:2181\ new Watcher() { // 监听回调 @Override

public void process(WatchedEvent event) { System.out.println(\"正在监听中......\"); } });

// 监听⽬录

children = zkCli.getChildren(\"/\ @Override

public void process(WatchedEvent event) {

System.out.println(\"监听路径为：\" + event.getPath()); System.out.println(\"监听的类型为：\" + event.getType()); System.out.println(\"监听被修改了\"); for (String c : children) { System.out.println(c); } } });

Thread.sleep(Long.MAX_VALUE);

}}

五、Zookeeper动态上下线感知系统

需求

某分布式系统当中，主节点有多台，可以进⾏动态上下线，当有任何⼀台机器发⽣了动态的上下线 任何⼀台客户端都能感知到。思路

（1）创建客户端与服务端（2）启动client端 监听（3）启动server端 注册

（4）当server端 发⽣上下线（5）client端都能感知到

1、服务端

package com.css.serverclient;import java.io.IOException;

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.Ids;import org.apache.zookeeper.ZooKeeper;

public class ZkServer {

public static void main(String[] args) throws IOException, KeeperException, InterruptedException { // 1.连接zkServer

ZkServer zkServer = new ZkServer(); zkServer.getConnect();

// 2.注册节点信息 服务器ip添加到zk中 zkServer.regist(args[0]); // 3.业务逻辑处理

zkServer.build(args[0]); }

private String connectString = \"192.168.146.132:2181,192.168.146.133:2181,192.168.146.134:2181\"; private int sessionTimeout = 3000; ZooKeeper zkCli = null; // 定义⽗节点

private String parentNode = \"/servers\";

// 1.连接zkServer

public void getConnect() throws IOException {

zkCli = new ZooKeeper(connectString, sessionTimeout, new Watcher() { @Override

public void process(WatchedEvent event) { } }); }

// 2.注册信息

public void regist(String hostname) throws KeeperException, InterruptedException {

String node = zkCli.create(parentNode + \"/server\ CreateMode.EPHEMERAL_SEQUENTIAL); System.out.println(node); }

// 3.构造服务器

public void build(String hostname) throws InterruptedException { System.out.println(hostname + \":服务器上线了！\"); Thread.sleep(Long.MAX_VALUE); }

}

2、客户端

package com.css.serverclient;import java.io.IOException;import java.util.ArrayList;import java.util.List;

import org.apache.zookeeper.KeeperException;import org.apache.zookeeper.WatchedEvent;import org.apache.zookeeper.Watcher;import org.apache.zookeeper.ZooKeeper;

// 客户端

public class ZkClient {

public static void main(String[] args) throws IOException, InterruptedException, KeeperException { // 1.获取连接

ZkClient zkClient = new ZkClient(); zkClient.getConnect(); // 2.监听服务的节点信息 zkClient.getServers(); // 3.业务逻辑（⼀直监听） zkClient.getWatch(); }

// 3.业务逻辑

public void getWatch() throws InterruptedException { Thread.sleep(Long.MAX_VALUE); }

// 2.监听服务的节点信息

public void getServers() throws KeeperException, InterruptedException { List children = zkCli.getChildren(\"/servers\ ArrayList serverList = new ArrayList(); // 获取每个节点的数据 for (String c : children) {

byte[] data = zkCli.getData(\"/servers/\" + c, true, null); serverList.add(new String(data)); }

// 打印服务器列表

System.out.println(serverList); }

private String connectString = \"192.168.146.132:2181,192.168.146.133:2181,192.168.146.134:2181\"; private int sessionTimeout = 3000; ZooKeeper zkCli = null;

// 1.连接集群

public void getConnect() throws IOException {

zkCli = new ZooKeeper(connectString, sessionTimeout, new Watcher() { @Override

public void process(WatchedEvent event) { List children = null; try {

// 监听⽗节点

children = zkCli.getChildren(\"/servers\

// 创建集合存储服务器列表

ArrayList serverList = new ArrayList(); // 获取每个节点的数据 for (String c : children) {

byte[] data = zkCli.getData(\"/servers/\" + c, true, null); serverList.add(new String(data));

}

// 打印服务器列表

System.out.println(serverList); } catch (KeeperException e) { e.printStackTrace();

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

3、只启动客户端时，控制台输出为：

此时在集群的某⼀台机器执⾏以下命令：

此时客户端控制台输出：

4、同时启动客户端和服务端 服务端执⾏

此时服务端输出：

客户端输出：

这样就实现了简单的动态上下线感知。