JavaGuide.net

切换主题

微信公众号

  1. 数据存储

  2. 基于java API初探zookeeper的使用

  3. 深入分析Watcher机制的实现原理

  4. Curator客户端的使用，简单高效

数据存储 ​

​ 基于znode，基于文件系统风格的，树形结构的文件模型，和内存数据库差不多，基于增删改查的命令去操作数据库，整个数据库包括整个树形结构的内容，比如说我们的节点目录，节点路径和权限信息，而且zookeeper它会定时去吧这些信息数据存储到磁盘上。

​ 使用DataTree管理整个数据存储，主要用来存储节点路径和数据的内容。底层是一个典型的基于ConcurrentHashMap的一个存储。

事务日志 ​

创造一个节点或者进行事务操作的时候，都会记录一个事务日志，

zoo.cfg 文件中 ， datadir

快照日志 ​

它会记录zookeeper的某个时刻的快照，记录整个zookeeper全量数据的内容，类似于我们数据备份的一个概念，并且他会把这些文件放在指定的目录下去做一个存储，也是基于datadir这样一个路径去做一个存储

运行时日志 ​

bin/zookeeper.out

我们可以通过它来看zookeeper运行时的一些情况。

实际应用过程中，我们会把事务日志单独挂载到一个磁盘上，我们的每一个请求，事务请求都会做一个输入日志的记录，那这个磁盘的读写的性能，将影响到zookeeper本身的操作的性能。

基于Java API初探zookeeper的使用 ​

sh zkCli.sh 基于控制台

xml
<dependency>
   <groupId>com.apache.zookeeper</groupId>
   <artifactId>zookeeper</artifactId>
   <version>3.4.8</version>
</dependency>
<dependency>
   <groupId>com.apache.zookeeper</groupId>
   <artifactId>zookeeper</artifactId>
   <version>3.4.8</version>
</dependency>

引入包以后就可以使用相应的API去做相应的操作，

代码 ​

java
public class ConnectionDemo {
    public static void main(String[] args) {
        try {
            ZooKeeper zooKeeper =
                new ZooKeeper("192.168.136.128:2181," +
                              "192.168.136.129:2181," +
                              "192.168.136.130:2181",
                              4000,null);
            System.out.println(zooKeeper.getState());


        } catch (IOException e) {
            e.printStackTrace();
        } catch (InterruptedException e) {
            e.printStackTrace();
        }
    }
}
public class ConnectionDemo {
    public static void main(String[] args) {
        try {
            ZooKeeper zooKeeper =
                new ZooKeeper("192.168.136.128:2181," +
                              "192.168.136.129:2181," +
                              "192.168.136.130:2181",
                              4000,null);
            System.out.println(zooKeeper.getState());


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

控制台 ​

c
log4j:WARN No appenders could be found for logger (org.apache.zookeeper.ZooKeeper).
log4j:WARN Please initialize the log4j system properly.
log4j:WARN See http://logging.apache.org/log4j/1.2/faq.html#noconfig for more info.
CONNECTING
log4j:WARN No appenders could be found for logger (org.apache.zookeeper.ZooKeeper).
log4j:WARN Please initialize the log4j system properly.
log4j:WARN See http://logging.apache.org/log4j/1.2/faq.html#noconfig for more info.
CONNECTING

连接以后

显示连接状态为 connecting 等待以后connected

引入watcher机制，new watcher() 直接输出链接状态 connected

img

连接，

代码 ​

java
public class ConnectionDemo {
    public static void main(String[] args) {
        try {
            ZooKeeper zooKeeper =
                new ZooKeeper("192.168.136.128:2181," +
                              "192.168.136.129:2181," +
                              "192.168.136.130:2181",
                              4000,null);
            System.out.println(zooKeeper.getState());

            Thread.sleep(1000);

            System.out.println(zooKeeper.getState());

            zooKeeper.close();

        } catch (IOException e) {
            e.printStackTrace();
        } catch (InterruptedException e) {
            e.printStackTrace();
        }
    }
}
public class ConnectionDemo {
    public static void main(String[] args) {
        try {
            ZooKeeper zooKeeper =
                new ZooKeeper("192.168.136.128:2181," +
                              "192.168.136.129:2181," +
                              "192.168.136.130:2181",
                              4000,null);
            System.out.println(zooKeeper.getState());

            Thread.sleep(1000);

            System.out.println(zooKeeper.getState());

            zooKeeper.close();

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

控制台 ​

verilog
log4j:WARN No appenders could be found for logger (org.apache.zookeeper.ZooKeeper).
log4j:WARN Please initialize the log4j system properly.
log4j:WARN See http://logging.apache.org/log4j/1.2/faq.html#noconfig for more info.
CONNECTING
CONNECTED
log4j:WARN No appenders could be found for logger (org.apache.zookeeper.ZooKeeper).
log4j:WARN Please initialize the log4j system properly.
log4j:WARN See http://logging.apache.org/log4j/1.2/faq.html#noconfig for more info.
CONNECTING
CONNECTED

Stat 对应控制台的各项属性

Watcher实现保证连接成功 ​

java
public class ConnectionDemo {
    public static void main(String[] args) {
        try {
            final CountDownLatch  countDownLatch = new CountDownLatch(1);
            ZooKeeper zooKeeper =
                new ZooKeeper("192.168.136.128:2181," +
                              "192.168.136.129:2181," +
                              "192.168.136.130:2181",
                              4000, new Watcher() {
                                  // 为了保证连接的成功状态
                                  @Override
                                  public void process(WatchedEvent watchedEvent) {
                                      // 连接成功以后触发watcher事件
                                      if(Event.KeeperState.SyncConnected == watchedEvent.getState()){
                                          // 如果收到了服务端的响应事件，连接成功
                                          // 连接成功以后做一个递减。
                                          countDownLatch.countDown();
                                      }
                                  }
                              });
            countDownLatch.await();

            System.out.println(zooKeeper.getState());//CONNECTED

            zooKeeper.close();

        } catch (IOException e) {
            e.printStackTrace();
        } catch (InterruptedException e) {
            e.printStackTrace();
        }
    }
}
public class ConnectionDemo {
    public static void main(String[] args) {
        try {
            final CountDownLatch  countDownLatch = new CountDownLatch(1);
            ZooKeeper zooKeeper =
                new ZooKeeper("192.168.136.128:2181," +
                              "192.168.136.129:2181," +
                              "192.168.136.130:2181",
                              4000, new Watcher() {
                                  // 为了保证连接的成功状态
                                  @Override
                                  public void process(WatchedEvent watchedEvent) {
                                      // 连接成功以后触发watcher事件
                                      if(Event.KeeperState.SyncConnected == watchedEvent.getState()){
                                          // 如果收到了服务端的响应事件，连接成功
                                          // 连接成功以后做一个递减。
                                          countDownLatch.countDown();
                                      }
                                  }
                              });
            countDownLatch.await();

            System.out.println(zooKeeper.getState());//CONNECTED

            zooKeeper.close();

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

控制台 ​

verilog
log4j:WARN No appenders could be found for logger (org.apache.zookeeper.ZooKeeper).
log4j:WARN Please initialize the log4j system properly.
log4j:WARN See http://logging.apache.org/log4j/1.2/faq.html#noconfig for more info.
CONNECTED
log4j:WARN No appenders could be found for logger (org.apache.zookeeper.ZooKeeper).
log4j:WARN Please initialize the log4j system properly.
log4j:WARN See http://logging.apache.org/log4j/1.2/faq.html#noconfig for more info.
CONNECTED

参数都需要穿进去version

Stat.getVersion()

增删改查 ​

java
public class ConnectionDemo {
   public static void main(String[] args) {
       try {
           final CountDownLatch  countDownLatch = new CountDownLatch(1);
           ZooKeeper zooKeeper =
               new ZooKeeper("192.168.136.128:2181," +
                             "192.168.136.129:2181," +
                             "192.168.136.130:2181",
                             4000, new Watcher() {
                                 // 为了保证连接的成功状态
                                 @Override
                                 public void process(WatchedEvent watchedEvent) {
                                     // 连接成功以后触发watcher事件
                                     if(Event.KeeperState.SyncConnected == watchedEvent.getState()){
                                         // 如果收到了服务端的响应事件，连接成功
                                         // 连接成功以后做一个递减。
                                         countDownLatch.countDown();
                                     }
                                 }
                             });
           countDownLatch.await();

           System.out.println(zooKeeper.getState());//CONNECTED

           // 添加节点
           zooKeeper.create("/java-con-darian","232".getBytes(),
                            ZooDefs.Ids.OPEN_ACL_UNSAFE, CreateMode.PERSISTENT);

           Stat stat = new Stat();
           //Stat和我们在控制台看到的属性是一样的。
           // 得到节点的值
           byte[] dataBytes = zooKeeper.getData("/java-con-darian", null, stat);
           System.out.println(new String(dataBytes));

           Thread.sleep(1000);
           // 修改节点的值，乐观锁的概念
           zooKeeper.setData("/java-con-darian", "55".getBytes(),stat.getVersion());

           byte[] updateBytes = zooKeeper.getData("/java-con-darian", null, stat);
           System.out.println(new String(updateBytes));

           zooKeeper.delete("/java-con-darian", stat.getVersion());

           zooKeeper.close();

           // 当前线程进行阻塞
           System.in.read();

       } catch (IOException e) {
           e.printStackTrace();
       } catch (InterruptedException e) {
           e.printStackTrace();
       } catch (KeeperException e) {
           e.printStackTrace();
       }
public class ConnectionDemo {
   public static void main(String[] args) {
       try {
           final CountDownLatch  countDownLatch = new CountDownLatch(1);
           ZooKeeper zooKeeper =
               new ZooKeeper("192.168.136.128:2181," +
                             "192.168.136.129:2181," +
                             "192.168.136.130:2181",
                             4000, new Watcher() {
                                 // 为了保证连接的成功状态
                                 @Override
                                 public void process(WatchedEvent watchedEvent) {
                                     // 连接成功以后触发watcher事件
                                     if(Event.KeeperState.SyncConnected == watchedEvent.getState()){
                                         // 如果收到了服务端的响应事件，连接成功
                                         // 连接成功以后做一个递减。
                                         countDownLatch.countDown();
                                     }
                                 }
                             });
           countDownLatch.await();

           System.out.println(zooKeeper.getState());//CONNECTED

           // 添加节点
           zooKeeper.create("/java-con-darian","232".getBytes(),
                            ZooDefs.Ids.OPEN_ACL_UNSAFE, CreateMode.PERSISTENT);

           Stat stat = new Stat();
           //Stat和我们在控制台看到的属性是一样的。
           // 得到节点的值
           byte[] dataBytes = zooKeeper.getData("/java-con-darian", null, stat);
           System.out.println(new String(dataBytes));

           Thread.sleep(1000);
           // 修改节点的值，乐观锁的概念
           zooKeeper.setData("/java-con-darian", "55".getBytes(),stat.getVersion());

           byte[] updateBytes = zooKeeper.getData("/java-con-darian", null, stat);
           System.out.println(new String(updateBytes));

           zooKeeper.delete("/java-con-darian", stat.getVersion());

           zooKeeper.close();

           // 当前线程进行阻塞
           System.in.read();

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

控制台 ​

c
CONNECTED
232
55
CONNECTED
232
55

事件机制 ​

我们对一个节点发起订阅以后，那么对这个节点发起的任何变化，它都会有一个触发，发给触发前绑定的客户端，相当于一个发布订阅的功能。

通过watcher机制，保证刚开始的连接状态是connected.

Watcher机制是zookeeper里边非常重要的一个特性，那么我们在zookeeper上创建节点的时候，我们可以绑定监听事件，比如说，我们可以监听节点的变更，删除，子节点变更的一些事件，通过这些事件，可以实现zookeeper的分布式锁 ，集群管理的一些功能。

Watcher特性：当数据发生变化的时候，zookeeper会产生一个watcher事件，并且发送到客户端，但客户端只会收到一次通知，如果后续这个节点再次发生变化，那么之前设置的watcher的客户端不会再次收到消息（watcher机制是一次性操作）。可以通过循环监听达到永久的监听效果。

如何注册事件机制 ​

通过三个操作来绑定事件

getData、Exists、getChildren

getData：通过一个结点获得一个节点的数据

Exists：判断这个节点是否存在

getChildren：可以拿到当前节点的所有子节点

​ 通过设置节点的路径，相当于给节点绑定了事件

如何触发事件？ ​

​ 凡是事务类型的操作，都会触发监听操作 create/delete/setData

zookeeper.exists(event.getPath().true);

true ：默认使用全局的默认事件。

代码 ​

java
public class WatcherDemo {
    public static void main(String[] args) {
        try {
            final CountDownLatch countDownLatch = new CountDownLatch(1);
            final ZooKeeper zooKeeper =
                new ZooKeeper("192.168.136.128:2181," +
                              "192.168.136.129:2181," +
                              "192.168.136.130:2181",
                              4000, new Watcher() {
                                  // 为了保证连接的成功状态
                                  @Override
                                  public void process(WatchedEvent watchedEvent) {
                                      System.out.println("全局的默认的watcher事件："
                                                         + watchedEvent.getType()
                                                         + "----->>>>>>"
                                                         + watchedEvent.getPath());
                                      // 连接成功以后触发watcher事件
                                      if (Event.KeeperState.SyncConnected == watchedEvent.getState()) {
                                          // 如果收到了服务端的响应事件，连接成功
                                          // 连接成功以后做一个递减。
                                          countDownLatch.countDown();
                                      }
                                  }
                              });
            countDownLatch.await();

            zooKeeper.create("/zk-tmp-darian", "1".getBytes(),
                             ZooDefs.Ids.OPEN_ACL_UNSAFE, CreateMode.PERSISTENT);

            // exists   getdata getchildren

            // 绑定事件
            // 我们在创建zookeeper的时候创建了一个默认的匿名内部类，
            // 如果说设置成 true 的话，所有的事件都会触发到内部类里边
            // zooKeeper.exists("/zi-tmp-darian",true);

            // 绑定自己的watcher事件
            // 通过 exists 绑定事件
            Stat stat = zooKeeper.exists("/zk-tmp-darian", new Watcher() {
                @Override
                public void process(WatchedEvent watchedEvent) {
                    System.out.println(watchedEvent.getType() + "--->>>>" + watchedEvent.getPath());

                    try {
                        // 通过再一次绑定事件去实现永久监听的一个效果
                        zooKeeper.exists(watchedEvent.getPath(), true);
                    } catch (KeeperException e) {
                        e.printStackTrace();
                    } catch (InterruptedException e) {
                        e.printStackTrace();
                    }
                }
            });

            // 通过修改的食物类型操作来触发监听事件
            stat = zooKeeper.setData("/zk-tmp-darian", "222".getBytes(), stat.getVersion());

            Thread.sleep(1000);

            zooKeeper.delete("/zk-tmp-darian", stat.getVersion());

            System.in.read();

        } catch (InterruptedException e) {
            e.printStackTrace();
        } catch (IOException e) {
            e.printStackTrace();
        } catch (KeeperException e) {
            e.printStackTrace();
        }
    }
}
public class WatcherDemo {
    public static void main(String[] args) {
        try {
            final CountDownLatch countDownLatch = new CountDownLatch(1);
            final ZooKeeper zooKeeper =
                new ZooKeeper("192.168.136.128:2181," +
                              "192.168.136.129:2181," +
                              "192.168.136.130:2181",
                              4000, new Watcher() {
                                  // 为了保证连接的成功状态
                                  @Override
                                  public void process(WatchedEvent watchedEvent) {
                                      System.out.println("全局的默认的watcher事件："
                                                         + watchedEvent.getType()
                                                         + "----->>>>>>"
                                                         + watchedEvent.getPath());
                                      // 连接成功以后触发watcher事件
                                      if (Event.KeeperState.SyncConnected == watchedEvent.getState()) {
                                          // 如果收到了服务端的响应事件，连接成功
                                          // 连接成功以后做一个递减。
                                          countDownLatch.countDown();
                                      }
                                  }
                              });
            countDownLatch.await();

            zooKeeper.create("/zk-tmp-darian", "1".getBytes(),
                             ZooDefs.Ids.OPEN_ACL_UNSAFE, CreateMode.PERSISTENT);

            // exists   getdata getchildren

            // 绑定事件
            // 我们在创建zookeeper的时候创建了一个默认的匿名内部类，
            // 如果说设置成 true 的话，所有的事件都会触发到内部类里边
            // zooKeeper.exists("/zi-tmp-darian",true);

            // 绑定自己的watcher事件
            // 通过 exists 绑定事件
            Stat stat = zooKeeper.exists("/zk-tmp-darian", new Watcher() {
                @Override
                public void process(WatchedEvent watchedEvent) {
                    System.out.println(watchedEvent.getType() + "--->>>>" + watchedEvent.getPath());

                    try {
                        // 通过再一次绑定事件去实现永久监听的一个效果
                        zooKeeper.exists(watchedEvent.getPath(), true);
                    } catch (KeeperException e) {
                        e.printStackTrace();
                    } catch (InterruptedException e) {
                        e.printStackTrace();
                    }
                }
            });

            // 通过修改的食物类型操作来触发监听事件
            stat = zooKeeper.setData("/zk-tmp-darian", "222".getBytes(), stat.getVersion());

            Thread.sleep(1000);

            zooKeeper.delete("/zk-tmp-darian", stat.getVersion());

            System.in.read();

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

控制台 ​

verilog
全局的默认的watcher事件：None----->>>>>>null
NodeDataChanged--->>>>/zk-tmp-darian
全局的默认的watcher事件：NodeDeleted----->>>>>>/zk-tmp-darian
全局的默认的watcher事件：None----->>>>>>null
NodeDataChanged--->>>>/zk-tmp-darian
全局的默认的watcher事件：NodeDeleted----->>>>>>/zk-tmp-darian

Watcher事件机制： ​

时间类型 ​

​ None（-1） 客户端链接状态发生变化的时候，会受到none的事件

​ NodeCreated（1） 创建节点的事件。比如说zk-persis-mic

​ NodeDeleted（2） 删除节点的事件。

​ NodeDataChanged（3） 节点数据发生变更

​ NodeChildrenChanged（4） 子节点被创建、被删除、会发生事件触发

触发条件 ​

zk-persis-mic（监听事件）zk-persis-mic/child（监听事件）
create(/zk-persis-mic)NodeCreated（exists/getData）无
delete(/zk-persis-mic)NodeDelete（exists/getData）无
setData(/zk-persis-mic)NodeDataChanged（exists/getData）无
create(/zk-persis-mic/children)NodeChildrenChanged(getChildren)NodeCreated
delete(/zk-persis-mic/children)NodeChildrenChanged(getChildren)NodeDelete
setData(/zk-persis-mic/children)NodeDataChanged

事件的实现原理 ​

通过exists方法去绑定的事件

入口就是exists

Zookeeper类中的exists方法

java
public Stat exists(final String path, Watcher watcher)
   throws KeeperException, InterruptedException{
   final String clientPath = path;
   PathUtils.validatePath(clientPath);

   // the watch contains the un-chroot path
   WatchRegistration wcb = null;
   if (watcher != null) {
       wcb = new ExistsWatchRegistration(watcher, clientPath);
   }

   final String serverPath = prependChroot(clientPath);

   RequestHeader h = new RequestHeader();
   h.setType(ZooDefs.OpCode.exists);
   ExistsRequest request = new ExistsRequest();
   request.setPath(serverPath);
   request.setWatch(watcher != null);
   SetDataResponse response = new SetDataResponse();
   ReplyHeader r = cnxn.submitRequest(h, request, response, wcb);
   if (r.getErr() != 0) {
       if (r.getErr() == KeeperException.Code.NONODE.intValue()) {
           return null;
       }
       throw KeeperException.create(KeeperException.Code.get(r.getErr()),
               clientPath);
   }

   return response.getStat().getCzxid() == -1 ? null : response.getStat();
}
public Stat exists(final String path, Watcher watcher)
   throws KeeperException, InterruptedException{
   final String clientPath = path;
   PathUtils.validatePath(clientPath);

   // the watch contains the un-chroot path
   WatchRegistration wcb = null;
   if (watcher != null) {
       wcb = new ExistsWatchRegistration(watcher, clientPath);
   }

   final String serverPath = prependChroot(clientPath);

   RequestHeader h = new RequestHeader();
   h.setType(ZooDefs.OpCode.exists);
   ExistsRequest request = new ExistsRequest();
   request.setPath(serverPath);
   request.setWatch(watcher != null);
   SetDataResponse response = new SetDataResponse();
   ReplyHeader r = cnxn.submitRequest(h, request, response, wcb);
   if (r.getErr() != 0) {
       if (r.getErr() == KeeperException.Code.NONODE.intValue()) {
           return null;
       }
       throw KeeperException.create(KeeperException.Code.get(r.getErr()),
               clientPath);
   }

   return response.getStat().getCzxid() == -1 ? null : response.getStat();
}

我们client在zookeeper server上注册一个事件

ZkWatcher Manager 上保存客户端的事件注册

img

源码地址：Zookeeper核心类中exists(

分析原理：

在客户端调用里边就是实例化，

调用增删改查方法

一定有一个机制区队这个队列去做一个数据的处理

OutgoingQueue

有一个入队肯定有一个出对

数据不是放大队列中就完事了，肯定有机制去对队列做一个出队列

判断是引导流程走向的一个方式！

Ctrl + alt + B 点到方法上，可以看到有几个实现

NettyServerCnxn

服务端处理请求的类

receiveMessage()方法

通过链式的process对请求去做一个链式的处理，把不同的业务去做一个分离

预处理 > 同步处理 > finally处理

img

PIPE（管道模式）

Linux 的管道的处理，一个请求过来，进入不同的管道，做不同的处理，最终返回一个结果。

类似于链式风格的一种设计。

Zookeeper中大部分都在用多线程的方式异步化流程。

一定有一个线程去处理队列 submittRequests。

Hashmap<watcher, String path>

img

Curator的操作 ​

主要是对我们数据节点操作的一个封装

​ 原生操作会有很多不便之处，虽然说灵活性比较好，但是我们要做一些事情，比如说，要去绑定事件很复杂，我们要去创建一个连接也很复杂，那怎么去简化？

​ Netflix公司开源了一个zookeeper客户端，他跟原生的客户端相比，它是一个更高层次的抽象，他不光是API层次的一个抽象，同时还封装了一套基于应用场景的API，也就是说zookeeper它本身能够实现分布式锁，leader选举这些功能，那么curator对这些功能做了一些封装，我们直接调用API就好了。就可以完成leader选举的这样一个机制，那么我们怎么实现这样一个功能。

基于fluent风格的一个机制

CuratorFramework curatorFamework = CuratorFrameworkFactory

.builder()

.connectString()

.sessionTimeoutMS(4000)

.retryPolicy(new ExponetialBackoffRetry(1000,3))

.namespace(“curator”)

.build()

// 结果： /curator/mic/node1

// 原生api中，必须是逐层创建，也就是父节点必须存在，子节点才能创建

Curator的增删改查 ​

java
package com.gupao.study.vip;

import org.apache.curator.framework.CuratorFramework;
import org.apache.curator.framework.CuratorFrameworkFactory;
import org.apache.curator.retry.ExponentialBackoffRetry;
import org.apache.zookeeper.CreateMode;
import org.apache.zookeeper.data.Stat;

/**
* <br>类说明 :
* <br>属性说明：
* <br>作者：Darian
**/
public class CuratorDemo {
   public static void main(String[] args) throws Exception {
       CuratorFramework curatorFramework = CuratorFrameworkFactory
           .builder()
           .connectString("192.168.136.128:2181," +
                          "192.168.136.129:2181," +
                          "192.168.136.130:2181")
           .sessionTimeoutMs(4000)
           // 一个重试机制，递减的重试机制
           .retryPolicy(new ExponentialBackoffRetry(1000, 3))
           // 指定命名空间，我们会把某个业务放在一个大的命名空间下面
           // 因为我们要区分每个场景下的业务划分
           .namespace("/curator")
           .build();

       curatorFramework.start();

       // 创建结果： /curator/darian/node1
       // 原生的 API 中，必须是逐层创建，也就是父节点必须存在，子节点才能创建
       curatorFramework.create()
           .creatingParentContainersIfNeeded()
           .withMode(CreateMode.PERSISTENT)
           .forPath("/darian/node1", "2".getBytes());

       // 查询节点，得到节点的 Stat 信息
       Stat stat = new Stat();
       curatorFramework.getData()
           .storingStatIn(stat)
           .forPath("/darian/node1");

       // 更新操作
       curatorFramework.setData()
           .withVersion(stat.getVersion())
           .forPath("/darian/node1", "232".getBytes());


       // 去删除某个节点
       curatorFramework.delete()
           .deletingChildrenIfNeeded()
           .forPath("/darian/node1");
   }
}
package com.gupao.study.vip;

import org.apache.curator.framework.CuratorFramework;
import org.apache.curator.framework.CuratorFrameworkFactory;
import org.apache.curator.retry.ExponentialBackoffRetry;
import org.apache.zookeeper.CreateMode;
import org.apache.zookeeper.data.Stat;

/**
* <br>类说明 :
* <br>属性说明：
* <br>作者：Darian
**/
public class CuratorDemo {
   public static void main(String[] args) throws Exception {
       CuratorFramework curatorFramework = CuratorFrameworkFactory
           .builder()
           .connectString("192.168.136.128:2181," +
                          "192.168.136.129:2181," +
                          "192.168.136.130:2181")
           .sessionTimeoutMs(4000)
           // 一个重试机制，递减的重试机制
           .retryPolicy(new ExponentialBackoffRetry(1000, 3))
           // 指定命名空间，我们会把某个业务放在一个大的命名空间下面
           // 因为我们要区分每个场景下的业务划分
           .namespace("/curator")
           .build();

       curatorFramework.start();

       // 创建结果： /curator/darian/node1
       // 原生的 API 中，必须是逐层创建，也就是父节点必须存在，子节点才能创建
       curatorFramework.create()
           .creatingParentContainersIfNeeded()
           .withMode(CreateMode.PERSISTENT)
           .forPath("/darian/node1", "2".getBytes());

       // 查询节点，得到节点的 Stat 信息
       Stat stat = new Stat();
       curatorFramework.getData()
           .storingStatIn(stat)
           .forPath("/darian/node1");

       // 更新操作
       curatorFramework.setData()
           .withVersion(stat.getVersion())
           .forPath("/darian/node1", "232".getBytes());


       // 去删除某个节点
       curatorFramework.delete()
           .deletingChildrenIfNeeded()
           .forPath("/darian/node1");
   }
}

CuratorFramework.create()

.creatingPatentsIfNeeded()

.withMode(CreateMode.PERSISTENT)

.forPath(“/mic/node1”,”1”.getBytes());

//删除

curatorFramework

.delete()

.deletingChildrenIfNeeded()

.forPath(“/mic/node1”);

Stat stat = new Stat();

curatorFramework

.getData()

.storingStatIn(stat).forPath(“/mic/node1”)

curatorFramework.setData().

withVersion(sta.getVersion())

.forPath(“/mic/node1”,”xx”.getBytes);

curatorFramework.close(); //关闭连接

curator 事件的高度封装，

一定有一个节点特性

PathChildCache 监听一个节点下子节点的创建、删除、更新

（CHILD-ADDED,CHILD_UPDATE,CHILD_REMOVED）

NodeCache 监听一个结点的更新和创建时间

​ （Receive Event:/mic）

TreeCache zonghe1PathChildCache和NodeCache的特性

Curator的watcher机制 ​

java
package com.gupao.study.vip;

import org.apache.curator.framework.CuratorFramework;
import org.apache.curator.framework.CuratorFrameworkFactory;
import org.apache.curator.framework.recipes.cache.*;
import org.apache.curator.retry.ExponentialBackoffRetry;

/**
* <br>类说明 : 事件的高度封装
* <p>
* curator封装了三种事件的机制
* PathChildCache   监听一个节点下节点的创建、删除、更新
* NodeCache        监听一个节点的更新和创建事件
* TreeCache        综合 PatchChildCache 和 NodeCache 的特性
*
* <br>属性说明：
* <br>作者：Darian
**/
public class CuratorWatcherDemo {
   public static void main(String[] args) throws Exception {
       CuratorFramework curatorFramework = CuratorFrameworkFactory
           .builder()
           .connectString("192.168.136.128:2181," +
                          "192.168.136.129:2181," +
                          "192.168.136.130:2181")
           .sessionTimeoutMs(4000)
           .retryPolicy(new ExponentialBackoffRetry(1000, 3))
           .namespace("curator")
           .build();

       curatorFramework.start();


       // 当前节点的创建、删除事件监听
       // addListenerWithNodeCache(curatorFramework, "/darian");

       // 子节点的增加、修改、删除的事件监听
       // addlistenerWithPathChildCache(curatorFramework,"/darian");

       // 综合节点的监听事件
       addListerWithTreeCache(curatorFramework, "/darian");

       System.in.read();

   }

   /**
    * NodeCache + PathChildrenCache
    **/
   public static void addListerWithTreeCache(CuratorFramework curatorFramework, String path) throws Exception {
       final TreeCache treeCache = new TreeCache(curatorFramework,path);

       TreeCacheListener treeCacheListener = new TreeCacheListener() {
           @Override
           public void childEvent(CuratorFramework curatorFramework, TreeCacheEvent treeCacheEvent) throws Exception {
               System.out.println(treeCacheEvent.getType()+"-->>"+treeCacheEvent.getData());
           }
       };

       treeCache.getListenable().addListener(treeCacheListener);
       treeCache.start();
       /*
       INITIALIZED-->>null
NODE_ADDED-->>ChildData{path='/darian', stat=51539607621,51539607621,1531222950294,1531222950294,0,0,0,0,1,0,51539607621
, data=[49]}
NODE_ADDED-->>ChildData{path='/darian/node1', stat=51539607622,51539607622,1531222969992,1531222969992,0,0,0,0,1,0,51539607622
, data=[49]}
NODE_ADDED-->>ChildData{path='/darian/node1/node1', stat=51539607623,51539607623,1531222989856,1531222989856,0,0,0,0,1,0,51539607623
, data=[49]}
NODE_REMOVED-->>ChildData{path='/darian/node1/node1', stat=51539607623,51539607623,1531222989856,1531222989856,0,0,0,0,1,0,51539607623
, data=[49]}
NODE_UPDATED-->>ChildData{path='/darian/node1', stat=51539607622,51539607625,1531222969992,1531223116477,1,2,0,0,1,0,51539607624
, data=[50]}
NODE_REMOVED-->>ChildData{path='/darian/node1', stat=51539607622,51539607625,1531222969992,1531223116477,1,2,0,0,1,0,51539607624
, data=[50]}
NODE_REMOVED-->>ChildData{path='/darian', stat=51539607621,51539607621,1531222950294,1531222950294,0,2,0,0,1,0,51539607626
, data=[49]}
        */
   }


   /**
    * 监听对应节点下的子节点的变化
    * 创建
    * 删除
    * 更新
    */
   public static void addlistenerWithPathChildCache(CuratorFramework curatorFramework, String path) throws Exception {
       final PathChildrenCache pathChildrenCache =
           new PathChildrenCache(curatorFramework, path, true);
       PathChildrenCacheListener pathChildrenCacheListener = new PathChildrenCacheListener() {
           @Override
           public void childEvent(CuratorFramework curatorFramework, PathChildrenCacheEvent pathChildrenCacheEvent) throws Exception {
               System.out.println("receive Event:"
                                  + pathChildrenCacheEvent.getType()
                                  + "-->>"
                                  + pathChildrenCacheEvent.getData());
           }
       };

       pathChildrenCache.getListenable().addListener(pathChildrenCacheListener);
       pathChildrenCache.start(PathChildrenCache.StartMode.NORMAL);
       /*
receive Event:CONNECTION_RECONNECTED-->>null
receive Event:CHILD_ADDED-->>ChildData{path='/darian/node1', stat=51539607611,51539607611,1531222192297,1531222192297,0,0,0,0,1,0,51539607611
, data=[49]}
receive Event:CHILD_REMOVED-->>ChildData{path='/darian/node1', stat=51539607611,51539607611,1531222192297,1531222192297,0,0,0,0,1,0,51539607611
, data=[49]}
        */
   }

   /**
    * 监听这个节点的变化
    * 创建
    * 更新
    **/
   public static void addListenerWithNodeCache(CuratorFramework curatorFramework, String path) throws Exception {
       final NodeCache nodeCache = new NodeCache(curatorFramework, path, false);

       NodeCacheListener nodeCacheListener = new NodeCacheListener() {
           @Override
           public void nodeChanged() throws Exception {

               System.out.println("Receive Event:" + nodeCache.getCurrentData().getPath());
           }
       };

       nodeCache.getListenable().addListener(nodeCacheListener);
       nodeCache.start();
   }
}
package com.gupao.study.vip;

import org.apache.curator.framework.CuratorFramework;
import org.apache.curator.framework.CuratorFrameworkFactory;
import org.apache.curator.framework.recipes.cache.*;
import org.apache.curator.retry.ExponentialBackoffRetry;

/**
* <br>类说明 : 事件的高度封装
* <p>
* curator封装了三种事件的机制
* PathChildCache   监听一个节点下节点的创建、删除、更新
* NodeCache        监听一个节点的更新和创建事件
* TreeCache        综合 PatchChildCache 和 NodeCache 的特性
*
* <br>属性说明：
* <br>作者：Darian
**/
public class CuratorWatcherDemo {
   public static void main(String[] args) throws Exception {
       CuratorFramework curatorFramework = CuratorFrameworkFactory
           .builder()
           .connectString("192.168.136.128:2181," +
                          "192.168.136.129:2181," +
                          "192.168.136.130:2181")
           .sessionTimeoutMs(4000)
           .retryPolicy(new ExponentialBackoffRetry(1000, 3))
           .namespace("curator")
           .build();

       curatorFramework.start();


       // 当前节点的创建、删除事件监听
       // addListenerWithNodeCache(curatorFramework, "/darian");

       // 子节点的增加、修改、删除的事件监听
       // addlistenerWithPathChildCache(curatorFramework,"/darian");

       // 综合节点的监听事件
       addListerWithTreeCache(curatorFramework, "/darian");

       System.in.read();

   }

   /**
    * NodeCache + PathChildrenCache
    **/
   public static void addListerWithTreeCache(CuratorFramework curatorFramework, String path) throws Exception {
       final TreeCache treeCache = new TreeCache(curatorFramework,path);

       TreeCacheListener treeCacheListener = new TreeCacheListener() {
           @Override
           public void childEvent(CuratorFramework curatorFramework, TreeCacheEvent treeCacheEvent) throws Exception {
               System.out.println(treeCacheEvent.getType()+"-->>"+treeCacheEvent.getData());
           }
       };

       treeCache.getListenable().addListener(treeCacheListener);
       treeCache.start();
       /*
       INITIALIZED-->>null
NODE_ADDED-->>ChildData{path='/darian', stat=51539607621,51539607621,1531222950294,1531222950294,0,0,0,0,1,0,51539607621
, data=[49]}
NODE_ADDED-->>ChildData{path='/darian/node1', stat=51539607622,51539607622,1531222969992,1531222969992,0,0,0,0,1,0,51539607622
, data=[49]}
NODE_ADDED-->>ChildData{path='/darian/node1/node1', stat=51539607623,51539607623,1531222989856,1531222989856,0,0,0,0,1,0,51539607623
, data=[49]}
NODE_REMOVED-->>ChildData{path='/darian/node1/node1', stat=51539607623,51539607623,1531222989856,1531222989856,0,0,0,0,1,0,51539607623
, data=[49]}
NODE_UPDATED-->>ChildData{path='/darian/node1', stat=51539607622,51539607625,1531222969992,1531223116477,1,2,0,0,1,0,51539607624
, data=[50]}
NODE_REMOVED-->>ChildData{path='/darian/node1', stat=51539607622,51539607625,1531222969992,1531223116477,1,2,0,0,1,0,51539607624
, data=[50]}
NODE_REMOVED-->>ChildData{path='/darian', stat=51539607621,51539607621,1531222950294,1531222950294,0,2,0,0,1,0,51539607626
, data=[49]}
        */
   }


   /**
    * 监听对应节点下的子节点的变化
    * 创建
    * 删除
    * 更新
    */
   public static void addlistenerWithPathChildCache(CuratorFramework curatorFramework, String path) throws Exception {
       final PathChildrenCache pathChildrenCache =
           new PathChildrenCache(curatorFramework, path, true);
       PathChildrenCacheListener pathChildrenCacheListener = new PathChildrenCacheListener() {
           @Override
           public void childEvent(CuratorFramework curatorFramework, PathChildrenCacheEvent pathChildrenCacheEvent) throws Exception {
               System.out.println("receive Event:"
                                  + pathChildrenCacheEvent.getType()
                                  + "-->>"
                                  + pathChildrenCacheEvent.getData());
           }
       };

       pathChildrenCache.getListenable().addListener(pathChildrenCacheListener);
       pathChildrenCache.start(PathChildrenCache.StartMode.NORMAL);
       /*
receive Event:CONNECTION_RECONNECTED-->>null
receive Event:CHILD_ADDED-->>ChildData{path='/darian/node1', stat=51539607611,51539607611,1531222192297,1531222192297,0,0,0,0,1,0,51539607611
, data=[49]}
receive Event:CHILD_REMOVED-->>ChildData{path='/darian/node1', stat=51539607611,51539607611,1531222192297,1531222192297,0,0,0,0,1,0,51539607611
, data=[49]}
        */
   }

   /**
    * 监听这个节点的变化
    * 创建
    * 更新
    **/
   public static void addListenerWithNodeCache(CuratorFramework curatorFramework, String path) throws Exception {
       final NodeCache nodeCache = new NodeCache(curatorFramework, path, false);

       NodeCacheListener nodeCacheListener = new NodeCacheListener() {
           @Override
           public void nodeChanged() throws Exception {

               System.out.println("Receive Event:" + nodeCache.getCurrentData().getPath());
           }
       };

       nodeCache.getListenable().addListener(nodeCacheListener);
       nodeCache.start();
   }
}

friends