zookeeper 的Java客户端

zookeeper服务器有三种Java客户端:
1.Zookeeper,Zookeeper官方提供的原生Java客户端
2.Zkclient,在原生Zookeeper基础上进行扩展的开源第三方Java客户端
3.Curator,Netflix公司在原生zookeeper基础上开源的Java客户端

---

1.zookeeper客户端:
文档如下:
https://zookeeper.apache.org/doc/current/api/index.html

maven依赖

<!-- https://mvnrepository.com/artifact/org.apache.zookeeper/zookeeper/3.5.5-->
<dependency>
    <groupId>org.apache.zookeeper</groupId>
    <artifactId>zookeeper</artifactId>
    <version>3.5.5</version>
</dependency>

客户端连接成功时显示事件状态为SyncConnected,类型None

image.png

zookeeper的Java客户端操作zookeeper代码如下

public class ZookeeperClient01 {

    private static final String ZK_ADDRESS = "127.0.0.1:2181";

    private static final String ZNODE = "/ZNODE";
    //到计数器,默认倒数1下
    private static CountDownLatch countDownLatch = new CountDownLatch(1);
    public static void main(String[] args) throws IOException, InterruptedException, KeeperException {
        //创建zookeeper客户端对象
        ZooKeeper zooKeeper = new ZooKeeper(ZK_ADDRESS, 500000, new Watcher() {
            /**
             * 观察与zookeeper有没有连接上
             * 这是个异步操作,非主线程执行
             * @param watchedEvent
             */
            @Override
            public void process(WatchedEvent watchedEvent) {
                //获取事件的状态
                Event.KeeperState state = watchedEvent.getState();

                //获取事件类型
                Event.EventType type= watchedEvent.getType();
                //校验连接状态和类型,连接上的状态为SyncConnected,类型为None
                if (Event.KeeperState.SyncConnected == state){
                    if (Event.EventType.None == type){
                        System.out.println("连接成功");
                        countDownLatch.countDown();
                    }
                }
            }
        });
        // 倒数计数器没有倒数完成,不能执行下面的代码.确保zookeeper对象创建成功
        countDownLatch.await();

        //开始操作zookeeper


        //获取状态
        Stat stat = zooKeeper.exists(ZNODE,false);
        System.out.println(stat);

        if (stat == null){
            //创建节点,节点存在则创建失败
            //参数 为 节点,数据,访问控制类型,创建类型(持久节点,临时节点,有序无序)
            //返回值为节点路径 /ZNODE
            String path = zooKeeper.create(ZNODE,"zookeeper".getBytes(),ZooDefs.Ids.OPEN_ACL_UNSAFE, CreateMode.PERSISTENT);
            System.out.println("节点创建结果:"+path);
            //创建子节点
            //String chlidPath = zooKeeper.create(ZNODE + "/child","chlid".getBytes(),ZooDefs.Ids.OPEN_ACL_UNSAFE,CreateMode.PERSISTENT);
            //System.out.println("子节点创建结果" + chlidPath);
        }else {
            //修改数据,返回该节点最新的状态
            Stat updateStat = zooKeeper.setData(ZNODE, "zookeeper2".getBytes(), stat.getVersion());
            //打印更新时间
            System.out.println(updateStat.getMtime());
        }

        //获取数据 节点路径,是否进行观察,状态(将stat更新为获取的当前节点最新的状态)
        byte[] bytes = zooKeeper.getData(ZNODE,false,stat);
        System.out.println(new String(bytes));
        List<String> children = zooKeeper.getChildren(ZNODE, false, stat);
        //[CHLID]
        children.forEach((String child)->{
                    System.out.println(children);
                });
    }
} 

在测试的时候经常出现这个异常
Exception in thread "main" org.apache.zookeeper.KeeperException$ConnectionLossException: KeeperErrorCode = ConnectionLoss for /ZNODE
可能是:1.防火墙的问题,关闭安装zookeeper的机器 防火墙
2.可能是在没有完全建立成功zookeeper对象时,进行了zookeeper操作,所以代码中采用了一个倒数的机制

zookeeper客户端是zookeeper自带的,编程较为繁琐,会话容易超时,重连繁琐.
节点数据是二进制,需要转换.
zookeeper客户端和服务器端的会话是异步,为了保证在做相关操作时,zookeeper客户端与服务端已经连接成功,我们需要做状态和类型校验,或者借助类似于CountDownLatch工具类.

---

-------------------------------------------------------------------------------------------------------------------------
2.Zkclient
在原生zookeeper基础上进行扩展的第三方Java客户端
Github：https://github.com/sgroschupf/zkclient

maven依赖

<!--https://mvnrepository.com/artifact/com.101tec/zkclient/0.11-->
<dependency>
    <groupId>com.101tec</groupId>
    <artifactId>zkclient</artifactId>
    <version>0.11</version>
</dependency>

代码如下

public class ZkClientDemo {

    private static final String ZK_ADDRESS = "127.0.0.1:2181";

    private static final String ZNODE = "/zkclient";
    public static void main(String[] args) {
        //创建zookeeper连接
        ZkClient client = new ZkClient(ZK_ADDRESS);
        //可以重写序列化器,并在创建连接的时候指定序列化器
        ZkClient mySerializerClient = new ZkClient(ZK_ADDRESS,5000,15000,new MyZkSerializer());
        //判断节点是否存在
        if (!client.exists(ZNODE)){
            //创建持久化节点 ,初始化数据
            client.createPersistent(ZNODE,"zkclient");
            String chlid = client.create(ZNODE + "/chlid", "chlid", CreateMode.PERSISTENT);
            System.out.println(chlid);
        }else {
            //修改节点数据,并返回该节点的状态
            Stat znodeStat = client.writeDataReturnStat(ZNODE, "znode", -1);
            System.out.println(znodeStat);
        }
        //获取子节点  子节点:chlid
        List<String> children = client.getChildren(ZNODE);
        children.forEach((String node) -> {
            System.out.println("子节点:" + node);
        });
        //获取节点数据
        String readData = client.readData(ZNODE);
        Stat stat = new Stat();
        //获取节点数据时更新节点状态
        client.readData(ZNODE + "/chlid",stat);
        System.out.println(readData);
    }
}

---

---