q968 发表于 2018-10-29 10:44:01

hadoop系列:zookeeper(3)——zookeeper核心原理(事件)

package com.yinwenjie.test.zookeepertest.test;import java.io.FileNotFoundException;import java.io.IOException;import org.apache.commons.logging.Log;import org.apache.commons.logging.LogFactory;import org.apache.zookeeper.CreateMode;import org.apache.zookeeper.WatchedEvent;import org.apache.zookeeper.Watcher;import org.apache.zookeeper.Watcher.Event.EventType;import org.apache.zookeeper.Watcher.Event.KeeperState;import org.apache.zookeeper.ZooDefs.Ids;import org.apache.zookeeper.ZooKeeper;import org.springframework.util.Log4jConfigurer;import com.yinwenjie.test.zookeepertest.TestZookeeperAgainst;/**  
* 这个测试类测试多个watcher监控某一个znode节点的效果。
  
* servers:192.168.61.129:2181,192.168.61.130:2181,192.168.61.132:2181
  
* 为了验证zk集群的工作效果,我们选择一个节点进行连接(192.168.61.129)。
  
* @author yinwenjie
  
*/public class TestManyWatcher implements Runnable {
  
    static {      try {
  
            Log4jConfigurer.initLogging("classpath:log4j.properties");
  
      } catch (FileNotFoundException ex) {
  
            System.err.println("Cannot Initialize log4j");
  
            System.exit(-1);
  
      }
  
    }    /**
  
   * 日志
  
   */
  
    private static final Log LOGGER = LogFactory.getLog(TestZookeeperAgainst.class);    public static void main(String[] args) throws Exception {
  
      TestManyWatcher testManyWatcher = new TestManyWatcher();      new Thread(testManyWatcher).start();
  
    }    public void run() {      /*
  
         * 验证过程如下:
  
         * 1、验证一个节点X上使用exist方式注册的多个监听器(ManyWatcherOne、ManyWatcherTwo),
  
         *      在节点X发生create事件时的事件通知情况
  
         * 2、验证一个节点Y上使用getDate方式注册的多个监听器(ManyWatcherOne、ManyWatcherTwo),
  
         *      在节点X发生create事件时的事件通知情况
  
         * */
  
      //默认监听:注册默认监听是为了让None事件都由默认监听处理,
  
      //不干扰ManyWatcherOne、ManyWatcherTwo的日志输出
  
      ManyWatcherDefault watcherDefault = new ManyWatcherDefault();
  
      ZooKeeper zkClient = null;      try {
  
            zkClient = new ZooKeeper("192.168.61.129:2181", 120000, watcherDefault);
  
      } catch (IOException e) {
  
            TestManyWatcher.LOGGER.error(e.getMessage(), e);            return;
  
      }      //默认监听也可以使用register方法注册
  
      //zkClient.register(watcherDefault);
  

  
      //1、========================================================
  
      TestManyWatcher.LOGGER.info("=================以下是第一个实验");
  
      String path = "/X";
  
      ManyWatcherOne watcherOneX = new ManyWatcherOne(zkClient , path);
  
      ManyWatcherTwo watcherTwoX = new ManyWatcherTwo(zkClient , path);      //注册监听,注意,这里两次exists方法的执行返回都是null,因为“X”节点还不存在
  
      try {
  
            zkClient.exists(path, watcherOneX);
  
            zkClient.exists(path, watcherTwoX);            //创建"X"节点,为了简单起见,我们忽略权限问题。
  
            //并且创建一个临时节点,这样重复跑代码的时候,不用去server上手动删除)
  
            zkClient.create(path, "".getBytes(), Ids.OPEN_ACL_UNSAFE, CreateMode.EPHEMERAL);
  
      } catch (Exception e) {
  
            TestManyWatcher.LOGGER.error(e.getMessage(), e);            return;
  
      }      //TODO 注意观察日志,根据原理我们猜测理想情况下ManyWatcherTwo和ManyWatcherOne都会被通知。
  

  
      //2、========================================================
  
      TestManyWatcher.LOGGER.info("=================以下是第二个实验");
  
      path = "/Y";
  
      ManyWatcherOne watcherOneY = new ManyWatcherOne(zkClient , path);
  
      ManyWatcherTwo watcherTwoY = new ManyWatcherTwo(zkClient , path);      //注册监听,注意,这里使用两次getData方法注册监听,"Y"节点目前并不存在
  
      try {
  
            zkClient.getData(path, watcherOneY, null);
  
            zkClient.getData(path, watcherTwoY, null);
  
      } catch (Exception e) {
  
            TestManyWatcher.LOGGER.error(e.getMessage(), e);
  
      }      //TODO 注意观察日志,因为"Y"节点不存在,所以getData就会出现异常。watcherOneY、watcherTwoY的注册都不起任何作用。
  
      //然后我们在报了异常的情况下,创建"Y"节点,根据原理,不会有任何watcher响应"Y"节点的create事件
  
      try {
  
            zkClient.create(path, "".getBytes(), Ids.OPEN_ACL_UNSAFE, CreateMode.EPHEMERAL);
  
      } catch (Exception e) {
  
            TestManyWatcher.LOGGER.error(e.getMessage(), e);            return;
  
      }      //下面这段代码可以忽略,是为了观察zk的原理。让守护线程保持不退出
  
      synchronized(this) {            try {                this.wait();
  
            } catch (InterruptedException e) {
  
                TestManyWatcher.LOGGER.error(e.getMessage(), e);
  
                System.exit(-1);
  
            }
  
      }
  
    }
  
}/**
  
* 这是默认的watcher实现。
  
* @author yinwenjie
  
*/class ManyWatcherDefault implements Watcher {    /**
  
   * 日志
  
   */
  
    private static Log LOGGER = LogFactory.getLog(ManyWatcherDefault.class);    public void process(WatchedEvent event) {
  
      KeeperState keeperState = event.getState();
  
      EventType eventType = event.getType();
  
      ManyWatcherDefault.LOGGER.info("=========默认监听到None事件:keeperState = "
  
                + keeperState + ":eventType = " + eventType);
  
    }
  
}/**
  
* 这是第一种watcher
  
* @author yinwenjie
  
*/class ManyWatcherOne implements Watcher {    /**
  
   * 日志
  
   */
  
    private static Log LOGGER = LogFactory.getLog(ManyWatcherOne.class);    private ZooKeeper zkClient;    /**
  
   * 被监控的znode地址
  
   */
  
    private String watcherPath;    public ManyWatcherOne(ZooKeeper zkClient , String watcherPath) {      this.zkClient = zkClient;      this.watcherPath = watcherPath;
  
    }    public void process(WatchedEvent event) {      try {            this.zkClient.exists(this.watcherPath, this);
  
      } catch (Exception e) {
  
            ManyWatcherOne.LOGGER.error(e.getMessage(), e);
  
      }
  
      KeeperState keeperState = event.getState();
  
      EventType eventType = event.getType();      //这个属性是发生事件的path
  
      String eventPath = event.getPath();
  

  
      ManyWatcherOne.LOGGER.info("=========ManyWatcherOne监听到" + eventPath + "地址发生事件:"
  
                + "keeperState = " + keeperState + ":eventType = " + eventType);
  
    }
  
}/**
  
* 这是第二种watcher
  
* @author yinwenjie
  
*/class ManyWatcherTwo implements Watcher {    /**
  
   * 日志
  
   */
  
    private static Log LOGGER = LogFactory.getLog(ManyWatcherOne.class);    private ZooKeeper zkClient;    /**
  
   * 被监控的znode地址
  
   */
  
    private String watcherPath;    public ManyWatcherTwo(ZooKeeper zkClient, String watcherPath) {      this.zkClient = zkClient;      this.watcherPath = watcherPath;
  
    }    public void process(WatchedEvent event) {      try {            this.zkClient.exists(this.watcherPath, this);
  
      } catch (Exception e) {
  
            ManyWatcherTwo.LOGGER.error(e.getMessage(), e);
  
      }
  
      KeeperState keeperState = event.getState();
  
      EventType eventType = event.getType();      //这个属性是发生事件的path
  
      String eventPath = event.getPath();
  

  
      ManyWatcherTwo.LOGGER.info("=========ManyWatcherTwo监听到" + eventPath + "地址发生事件:"
  
                + "keeperState = " + keeperState + ":eventType = " + eventType);
  
    }
  
}123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209


页: [1]
查看完整版本: hadoop系列:zookeeper(3)——zookeeper核心原理(事件)