【ZooKeeper】4. 服务器动态上下线监听案例

4.1 需求分析

服务器动态上下线

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-fgjrvpYH-1629616974036)(屏幕截图 2021-08-18 164527.png)]

本质上：

1. 服务器与客户端1、2…针对于Zookeeper集群来说，都是客户端
2. 服务器跟Zookeeper集群打交道的是创建节点：create( )操作
3. 客户端1、2…跟Zookeeper集群打交道的是监听：-w 、get

4.2 服务器向集群注册

先在集群上创建/servers 节点：

[zk: localhost:2181(CONNECTED) 10] create /servers "servers"
Created /servers
[zk: localhost:2181(CONNECTED) 11] ls /
[sanguo, servers, zookeeper]

DistributeServer类：

public class DistributeServer {
    private String connectString = "hadoop102:2181,hadoop103:2181,hadoop104:2181";
    private int sessionTimeout = 2000;
    private ZooKeeper zooKeeper;


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

        DistributeServer server = new DistributeServer();
        //1.获取zk连接
        server.getConnect();
        //2.注册服务器到zk集群（将服务器名称作为参数传入）
        server.Regist(args[0]);
        //3.启动业务逻辑（sleep）
        server.business();
    }

    private void getConnect() throws IOException {
        zooKeeper = new ZooKeeper(connectString, sessionTimeout, new Watcher() {
            @Override
            public void process(WatchedEvent watchedEvent) {

            }
        });
    }

    private void Regist(String hostname) throws InterruptedException, KeeperException {
        String server = zooKeeper.create("/servers/"+hostname, hostname.getBytes(), ZooDefs.Ids.OPEN_ACL_UNSAFE, CreateMode.EPHEMERAL_SEQUENTIAL);
        System.out.println(hostname + " is online");
    }

    private void business() throws InterruptedException {
        Thread.sleep(Long.MAX_VALUE);
    }
}

DistributeClient类：

public class DistributeClient {
    private String connectString = "hadoop102:2181,hadoop103:2181,hadoop104:2181";
    private int sessionTimeout = 2000;
    private ZooKeeper zooKeeper;

    public static void main(String[] args) throws IOException, InterruptedException, KeeperException {
        DistributeClient Client = new DistributeClient();
        //1.获取zk连接
        Client.getConnect();
        //2.监听/severs下面子节点的增加和删除
        Client.getServerList();
        //3.业务逻辑（sleep）
        Client.business();
    }

    private void getConnect() throws IOException {
        zooKeeper = new ZooKeeper(connectString, sessionTimeout, new Watcher() {
            @Override
            public void process(WatchedEvent watchedEvent) {
                //监听完成一次后，立刻再监听一次
                try {
                    getServerList();
                } catch (InterruptedException e) {
                    e.printStackTrace();
                } catch (KeeperException e) {
                    e.printStackTrace();
                }
            }
        });
    }

    private void getServerList() throws InterruptedException, KeeperException {
        List<String> children = zooKeeper.getChildren("/servers", true);

        ArrayList<String> servers = new ArrayList<>();

        for (String child : children) {
            //将获取到的节点名拼接成绝对路径，以此来获取到节点中的数据
            byte[] data = zooKeeper.getData("/servers/" + child, false, null);
            //将获取到的数据存入集合中
            servers.add(new String(data));
        }
        //打印集合
        System.out.println(servers);
    }

    private void business() throws InterruptedException {
        //Thread.sleep(80000);
        //客户端在执行业务的同时一直在监听集群
        //sleep设置的时间太短会导致获取不到监听
        Thread.sleep(Long.MAX_VALUE);
    }
}

测试：

1. 先执行DistributeClient类，启动客户端

   

2. 执行DistributeServer：

   1. mian函数需要先传入参数：传入hadoop102

      

   

   2. 执行DistributeServer：Hadoop102上线

      

3. 观察客户端的变化：监听到Hadoop102上线，并打印集合