1、服务器动态上下线监听案例
1、需求
某分布式系统中,主节点可以有多台,可以动态上下线,任意一台客户端都能实时感知到主节点服务器的上下线。
2、需求分析
3、具体实现
(1)先在集群上创建/servers 节点
(2)在IDEA中创建包名:org.example.zkcase1
(3)服务端向Zookeeper注册代码
package org.example.zkcase1; import org.apache.zookeeper.*; import java.io.IOException; /** * @ClassName DistributeServer * @Description TODO * @Author Zouhuiming * @Date 2023/5/25 9:12 * @Version 1.0 */ public class DistributeServer { private String connectString="hadoop102:2181,hadoop103:2181,hadoop104:2181"; private int sessionTimeout=2000; ZooKeeper zk=null; private String parentNode="/servers"; //创建到zk的客户端连接 public void getConnect() throws IOException { zk=new ZooKeeper(connectString, sessionTimeout, new Watcher() { @Override public void process(WatchedEvent watchedEvent) { } }); } //注册服务器 public void registServer(String hostname) throws InterruptedException, KeeperException { String create = zk.create(parentNode + "/servers", hostname.getBytes(), ZooDefs.Ids.OPEN_ACL_UNSAFE, CreateMode.EPHEMERAL_SEQUENTIAL); System.out.println(hostname+"is online"+create); } //业务功能 public void business(String hostname) throws InterruptedException { System.out.println(hostname+" is working..."); Thread.sleep(Long.MAX_VALUE); } public static void main(String[] args) throws IOException, InterruptedException, KeeperException { //1、获取zk连接 DistributeServer server=new DistributeServer(); server.getConnect(); //2、利用zk连接注册服务器信息 server.registServer(args[0]); //3、启动业务功能 server.business(args[0]); } }
(4)客户端代码
package org.example.zkcase1; import org.apache.zookeeper.KeeperException; import org.apache.zookeeper.WatchedEvent; import org.apache.zookeeper.Watcher; import org.apache.zookeeper.ZooKeeper; import java.io.IOException; import java.util.ArrayList; import java.util.List; /** * @ClassName DistributeClient * @Description TODO * @Author Zouhuiming * @Date 2023/5/25 9:23 * @Version 1.0 */ public class DistributeClient { private String connectionString="hadoop102:2181,hadoop103:2181,hadoop104:2181"; private int sessionTimeout=2000; private ZooKeeper zk; private String parentNode="/servers"; //创建到zk的客户端连接 public void getConnection() throws IOException { zk=new ZooKeeper(connectionString, sessionTimeout, new Watcher() { @Override public void process(WatchedEvent watchedEvent) { //再次启动监听 try { getServerList(); } catch (InterruptedException e) { throw new RuntimeException(e); } catch (KeeperException e) { throw new RuntimeException(e); } } }); } //获取服务器列表信息 public void getServerList() throws InterruptedException, KeeperException { //1、获取服务器子节点信息,并且对父节点进行监听 List<String> children = zk.getChildren(parentNode, true); //2、储存服务器信息列表 ArrayList<String> servers=new ArrayList<>(); //3、遍历所有节点,获取节点中的主机名称信息 for (String child : children) { byte[] data = zk.getData(parentNode + "/" + child, false, null); servers.add(new String(data)); } //4、打印服务器列表信息 System.out.println(servers); } //业务功能 public void business() throws InterruptedException { System.out.println("client is working..."); Thread.sleep(Long.MAX_VALUE); } public static void main(String[] args) throws IOException, InterruptedException, KeeperException { //1、获取zk连接 DistributeClient client=new DistributeClient(); client.getConnection(); //2、获取servers的子节点信息,从中获取服务器信息列表 client.getServerList(); //3、业务进程启动 client.business(); } }
4、测试
(1)在Linux命令行上操作增加减少服务器
(a)启动 DistributeClient 客户端
(b)在 hadoop102 上 zk 的客户端/servers 目录上创建临时带序号节点
(c)观察IDEA控制台变化
(d)执行删除操作
(e)观察IDEA控制台变化
(2)在IDEA上操作增加和减少服务器
(a)启动 DistributeClient 客户端(如果已经启动过,不需要重启)
(b)启动 DistributeServer 服务
①点击 Edit Configurations…
②在弹出的窗口中(Program arguments)输入想启动的主机,例如,hadoop102
③回到 DistributeServer 的 main 方 法 , 右 键 , 在 弹 出 的 窗 口 中 点 击 Run“DistributeServer.main()
④观察 DistributeServer 控制台,提示 hadoop102 is working
⑤观察 DistributeClient 控制台,提示 hadoop102 已经上线
2、ZooKeeper 分布式锁案例
什么加做分布式锁呢?
比如说"进程 1"在使用该资源的时候,会先去获得锁,"进程 1"获得锁以后会对该资源保持独占,这样其他进程就无法访问该资源,"进程 1"用完该资源以后就将锁释放掉,让其他进程来获得锁,那么通过这个锁机制,我们就能保证了分布式系统中多个进程能够有序的访问该临界资源。那么我们把这个分布式环境下的这个锁叫作分布式锁。
2.1 原生 Zookeeper 实现分布式锁案例
1、分布式锁实现
package org.example.lock; import org.apache.zookeeper.*; import org.apache.zookeeper.data.Stat; import java.io.IOException; import java.util.Collections; import java.util.List; import java.util.concurrent.CountDownLatch; /** * @ClassName DistributedLock * @Description TODO * @Author Zouhuiming * @Date 2023/5/25 14:51 * @Version 1.0 */ public class DistributedLock { //Zookeeper server列表 private String connectString="hadoop102:2181,hadoop103:2181,hadoop104:2181"; //超时时间 private int sessionTimeout=2000; private ZooKeeper zk; private String rootNode="locks"; private String subNode="seq-"; //当前client等待的子节点 private String waitPath; //Zookeeper连接 private CountDownLatch connectLatch=new CountDownLatch(1); //Zookeeper节点等待 private CountDownLatch waitLatch=new CountDownLatch(1); //当前client创建的子节点 private String currentNode; //和zk服务建立连接,并创建根节点 public DistributedLock() throws IOException, InterruptedException, KeeperException { zk=new ZooKeeper(connectString, sessionTimeout, new Watcher() { @Override public void process(WatchedEvent watchedEvent) { //连接建立时,打开latch,唤醒wait在该latch上的线程 if (watchedEvent.getState()==Event.KeeperState.SyncConnected){ try { connectLatch.countDown(); } catch (Exception e){ e.printStackTrace(); } } //发生了waitPath的删除事件 if (watchedEvent.getType()==Event.EventType.NodeDeleted&&watchedEvent.getPath().equals(waitPath)){ waitLatch.countDown(); } } }); //等待连接建立 connectLatch.await(); //获取根节点状态 Stat stat = zk.exists("/" + rootNode, false); if (stat==null){ System.out.println("根节点不存在"); zk.create("/"+rootNode,new byte[0], ZooDefs.Ids.OPEN_ACL_UNSAFE,CreateMode.PERSISTENT); } } //加锁方法 public void zkLock(){ try { //在根节点下创建临时顺序节点,返回值为创建的节点路径 currentNode=zk.create("/"+rootNode+"/"+subNode,null,ZooDefs.Ids.OPEN_ACL_UNSAFE,CreateMode.EPHEMERAL_SEQUENTIAL); //wait一会儿,让结果更加清晰一些 Thread.sleep(10); //注意,没有必要监听"/locks"的子节点的变化情况 List<String> childrenNodes = zk.getChildren("/" + rootNode, false); //列表中只有一个子节点,那肯定就是currentNode,说明client获得锁 if (childrenNodes.size()==1){ return; } else { //对根节点下的所有临时顺序节点从小到大排序 Collections.sort(childrenNodes); //当前节点名称 String thisNode = currentNode.substring(("/" + rootNode + "/").length()); int index = childrenNodes.indexOf(thisNode); if (index==-1){ System.out.println("数据异常"); } else if (index==0) { //index==0,说明thisNode在列表中最小,当前client获得锁 return; }else { //获得排名比current前一位的节点 this.waitPath="/"+rootNode+"/"+childrenNodes.get(index-1); //在waitPath上注册监听器,当waitPath被删除时,Zookeeper会回调监听器的process方法 zk.getData(waitPath,true,new Stat()); //进入等待锁状态 waitLatch.await(); return; } } } catch (KeeperException e) { throw new RuntimeException(e); } catch (InterruptedException e) { throw new RuntimeException(e); } } //解锁方法 public void zkUnlock(){ try { zk.delete(this.currentNode,-1); } catch (InterruptedException e) { throw new RuntimeException(e); } catch (KeeperException e) { throw new RuntimeException(e); } } }
2、分布式锁测试
package org.example.lock; import org.apache.zookeeper.KeeperException; import java.io.IOException; /** * @ClassName DistributedLockTest * @Description TODO * @Author Zouhuiming * @Date 2023/5/25 15:27 * @Version 1.0 */ public class DistributedLockTest { public static void main(String[] args) throws IOException, InterruptedException, KeeperException { //创建分布式锁 final DistributedLock lock1=new DistributedLock(); final DistributedLock lock2=new DistributedLock(); new Thread(new Runnable() { @Override public void run() { //获取锁对象 try { lock1.zkLock(); System.out.println("线程1获取锁"); Thread.sleep(5*1000); lock1.zkUnlock(); System.out.println("线程1释放锁"); } catch (InterruptedException e) { e.printStackTrace(); } } }).start(); new Thread(new Runnable() { @Override public void run() { //获取锁对象 try { lock2.zkLock(); System.out.println("线程2获取锁"); Thread.sleep(5*1000); lock2.zkUnlock(); System.out.println("线程2释放锁"); } catch (InterruptedException e) { e.printStackTrace(); } } }).start(); } }
(2)观察控制台变化
2.2 Curator框架实现分布式锁案例
1、原生的 Java API 开发存在的问题
(1)会话连接是异步的,需要自己去处理。比如使用 CountDownLatch
(2)Watch 需要重复注册,不然就不能生效
(3)开发的复杂性还是比较高的
(4)不支持多节点删除和创建。需要自己去递归
2、Curator 是一个专门解决分布式锁的框架,解决了原生 JavaAPI 开发分布式遇到的问题。
3、Curator 案例实操
(1)添加依赖
<dependency> <groupId>org.apache.curator</groupId> <artifactId>curator-framework</artifactId> <version>4.3.0</version> </dependency> <dependency> <groupId>org.apache.curator</groupId> <artifactId>curator-recipes</artifactId> <version>4.3.0</version> </dependency> <dependency> <groupId>org.apache.curator</groupId> <artifactId>curator-client</artifactId> <version>4.3.0</version> </dependency>
(2)代码实现
package org.example.zkCase2; import org.apache.curator.framework.CuratorFramework; import org.apache.curator.framework.CuratorFrameworkFactory; import org.apache.curator.framework.recipes.locks.InterProcessLock; import org.apache.curator.framework.recipes.locks.InterProcessMutex; import org.apache.curator.retry.ExponentialBackoffRetry; /** * @ClassName CuratorLockTest * @Description TODO * @Author Zouhuiming * @Date 2023/5/25 15:52 * @Version 1.0 */ public class CuratorLockTest { private String rootNode="/locks"; //zookeeper server列表 private String connectString="hadoop102:2181,hadoop103:2181,hadoop104:2181"; //connection超时时间 private int connectionTimeOut=2000; //session超时时间 private int sessionTimeOut=2000; public static void main(String[] args) { new CuratorLockTest().test(); } //测试 private void test(){ //创建分布式锁1 final InterProcessLock lock1=new InterProcessMutex(getCuratorFramework(),rootNode); //创建分布式锁2 final InterProcessLock lock2=new InterProcessMutex(getCuratorFramework(),rootNode); new Thread(new Runnable() { @Override public void run() { //获取锁对象 try { lock1.acquire(); System.out.println("线程1获取锁"); //测试锁重入 lock1.acquire(); System.out.println("线程1再次获取锁"); lock1.release(); System.out.println("线程1释放锁"); lock1.release(); System.out.println("线程1再次释放锁"); } catch (Exception e) { throw new RuntimeException(e); } } }).start(); new Thread(new Runnable() { @Override public void run() { //获取锁对象 try { lock2.acquire(); System.out.println("线程2获取锁"); //测试锁重入 lock2.acquire(); System.out.println("线程2再次获取锁"); lock2.release(); System.out.println("线程2释放锁"); lock2.release(); System.out.println("线程2再次释放锁"); } catch (Exception e) { throw new RuntimeException(e); } } }).start(); } //分布式锁初始化 private CuratorFramework getCuratorFramework() { //重试策略,初始时间3秒,重试3次 ExponentialBackoffRetry policy = new ExponentialBackoffRetry(3000, 3); //通过工厂创建Curator CuratorFramework client= CuratorFrameworkFactory.builder() .connectString(connectString) .connectionTimeoutMs(connectionTimeOut) .sessionTimeoutMs(sessionTimeOut) .retryPolicy(policy) .build(); //开启连接 client.start(); System.out.println("zookeeper 初始化完成..."); return client; } }
(3)观察控制台变化
3、企业面试真题
3.1 选举机制
半数机制·,超过半数的投票通过,即通过。
1、第一次启动选举机制:
投票过半数时,服务器id大的胜出
2、第二次启动选举规则:
(1)EPOCH大的直接胜出
(2)EPOCH相同,事务id大的胜出
(3)事务id相同,服务器id大的胜出
3.2 生产集群安装多少zk合适
安装奇数台。
生产经验:
- 10台服务器:3台zk
- 20台服务器:5台zk
- 100台服务器:11台zk
- 200台服务器:11台zk
服务器台数多:
(1)好处:提高可靠性
(2)坏处:提高通信延时
3.3 常用命令
ls、get、create、delete