flink hadoop 从0~1分布式计算与大数据项目实战（4）zookeeper内部原理流程简介以及java curator client操作集群注册，读取

1.zookeeper 集群（至少个节点）启动后，会自动选举出一个leader,其他节点为follower 跟随节点
2.client 连接给定的单地址或者集群，连接集群时，会挑选一个可用的节点进行连接，整个节点可能时leader，也可能时follower
3.如果是读请求，则直接从当前连接的节点获取指定的service实例
4.如果时写请求
4.1.如果当前连接的是leader,则leader会将信息同步到所有其他节点并且得到ack反馈后commit消息，返回给client 成功
4.2.如果当前连接的是follower,则会先把请求转发给leader,然后会继续4.1的流程

@Component
public class RegisterZookeeper implements ApplicationRunner {
    @Value("${zookeeper.address}")
    private String zkAddress;
    @Override
    public void run(ApplicationArguments args) throws Exception {
        CuratorFramework client = CuratorFrameworkFactory.newClient(zkAddress, new RetryOneTime(1000));
        client.start();
        client.blockUntilConnected();
        ServiceInstance<Object> instance = ServiceInstance.builder().name("test1").address("10.188.17.28").port(7777).build();
        ServiceDiscovery<Object> serviceDiscovery = ServiceDiscoveryBuilder.builder(Object.class).client(client)
                .basePath("/zookeeper-test").build();
        serviceDiscovery.registerService(instance);
        serviceDiscovery.start();
        System.out.println("service register ok");
    }
}

zookeeper:
  address: hadoop1:2181,hadoop2:2181,hadoop3:2181

public class ZookeeperTester {
    private static  ZooKeeper Zk;
    public static void main(String[] args) throws Exception {
        CuratorFramework client = CuratorFrameworkFactory.newClient("hadoop1:2181,hadoop1:2181,hadoop1:2181", new RetryOneTime(1000));
        client.start();
        client.blockUntilConnected();
        ServiceDiscovery<Object> serviceDiscovery = ServiceDiscoveryBuilder.builder(Object.class).client(client)
                .basePath("/zookeeper-test").build();
        Collection<ServiceInstance<Object>> list = serviceDiscovery.queryForInstances("test1");
        list.forEach((instance)->{
            String servicePath = instance.getAddress()+":"+instance.getPort();
            System.out.println(servicePath);
        });
        Thread.sleep(10000);
    }
}