1. Unzip to the specified directory [[email protected] software]$ tar -zxvf zookeeper-3.4.10.tar.gz -C /opt/module/ will/opt/module/zookeeper-3.4.< span style="color: #800080;">10/The zoo_sample.cfg in the conf path is modified to zoo.cfg; 2. Change the name of the configuration file in the conf foldermv zoo_sample.cfg zoo.cfg
 Open the zoo.cfg file and modify the dataDir path: 3. Edit zoo.cfg, configure datad ir datadir=/opt/module/zookeeper-3.4.10/ zkData is in/opt/module/zookeeper-3.4.10/Create a zkData folder on this directory [[email protected] zookeeper-3.4.10]$ mkdir zkData 4. Configure cluster machines, each machine is assigned A different Serverid server.1=hadoop101:2888:3888 server.2 =hadoop102:2888:3888 server.3=hadoop103:2888:3888 5. Create a new myid file in the zkData folder, the content is the serverid of the machine  6. Configured Zookeeper LogDIR: Configure bin/zkEnv.sh file ZOO_LOG_DIR= "."Change to a custom log directory /opt/module/zookeeper-3.4.10 /logs
 Start: 7. bin/zkServer.sh start Check whether the process is started [[emailprotected] zookeeper-3.4.10]$ jps ##each This process is to provide services; 　　4020 Jps 　　 4001 QuorumPeerMain View status: [[emailprotected] zookeeper-3.4.10]$ bin/zkServer.sh status Start the client: [[email protected] zookeeper-3.4.10]$ bin/zkCli.sh ##Client to connect to the cluster;
< /span>

　　WATCHER::

　　WatchedEvent state:SyncConnected type:None path:null
　　[zk: localhost:2181(CONNECTED) 0]< /p>

　　[[email protected] zookeeper-3.4.10]$ jps
　　　　3248 ZooKeeperMain ##服务端
　　　　3076 QuorumPeerMain ##Run client
　　/p>

 Exit the client: [zk: localhost:2181(CONNECTED) 0] quit
 Stop Zookeeper [[email protected] zookeeper-3.4.10]$ bin/zkServer.sh stop

Save data+If the notification mechanism kills the server, it will refuse to connect; the default connection is local; if the server is specified at startup, kill its server, it will jump directly; [ zk: localhost:2181(CONNECTED) 1] ls /zookeeper ##This is the node that comes with zookeeper; [quota] [zk: localhost:< span style="color: #800080;">2181(CONNECTED) 2] ls / watch [test2, test40000000004, zookeeper, test1] Create on another client: [zk: localhost:2181(CONNECTED)  0] create /data1 "heihei" ##When creating a node, you must tell it what the data is Created /data1 [zk: localhost:2181(CONNECTED) 3 ] WATCHER:: WatchedEvent state:SyncConnected type:NodeChildrenChanged path:/ Create a second node and nothing It reacted; the observation mechanism of zookeeper is effective once; because each node has a watchingList, the server has the ability to observe at this time;ls / Watch registration observes the root directory, the client will apply, zookeeper keeps the data in the root directory’s watchingList; if the watchingList changes, it will notify all registrations Client , Every notification will be crossed out from the list; common creation-s contains sequence-e Temporary (restart or timeout disappears) [zk: localhost:2181(CONNECTED)  2] create -s /data2 1235 Created /data20000000006 [zk: localhost:2181(CONNECTED)  3] create -e /linshi 001 Created /linshi [zk: localhost:2181(CONNECTED)  6] quit ### Quitting... 2019-01-27 00 :29:27,946 [myid:] -INFO [main:[email protected]684]-Session: 0x1688adaecec0001 closed 2019-01- 27 00:29 span>:27,954 [myid:]-INFO [main-EventThread:[ email protected]519]-EventThread shut down for session: 0x1688adaecec0001 Different clients are independent of each other, only the node created by oneself quit, and this node on another client (Within 2s) disappeared; ===> There are 4 types of nodes: orderly and persistent-s; [zk: localhost:2181(CONNECTED) 1] create -s /order 111 Created /order0000000009 orderly and short-s- e; [zk: localhost:2181(CONNECTED) 2] create -s -e /orderAndShort 222 Created /orderAndShort0000000010 disorderly and persistent; [zk: localhost:2181(CONNECTED) < span style="color: #800080;">3] create /long 333< /span> Created /long Disorderly short-e; [zk: localhost: 2181(CONNECTED) 4] create -e /short 444 Created /short< /span>

[zk: localhost:2181(CONNECTED) 2] get /test2 watch ##At this time, what is monitored is the content of the node; and ls is the path change of the monitoring node; abcd cZxid = 0x200000003 ctime = Sat Jan 26 15:23:14 CST 2019 mZxid = 0x200000003 span> mtim e = Sat Jan 26 15:23:14 CST 2019 pZxid = 0x400000015 cversion = < span style="color: #800080;">1 dataVersion = 0 aclVersion = 0 ephemeralOwner < /span>= 0x0 dataLength = 4 numChildren = 1 [zk: localhost:2181(CONNECTED) 3] WATCHER:: WatchedEvent state:SyncConnected type:NodeDataChanged path:/test2 [zk: localhost: 2181(CONNECTED) 4] create /test2/test0 123 ##Change the path of the node and create a child node; the value of the byte point has not changed; Created /test2/test0 [zk: localhost:2181(CONNECTED) 5] set /test2 QQ ##set is to change the value of the node cZxid = 0x200000003 ctime = Sat Jan 26 15:23:14 CST 2019 mZxid = 0x400000016 mtime = Sun Jan 27 00:53:19 CST 2019 pZxid = 0x400000015 span> cversion = 1 dataVersion = 1 aclVersion = 0 ephemeralOwner = 0x0 dataLength = 2 numChildren = 1 [zk: localhost:2181(CONNECTED) 6] stat /test2 cZxid = 0x200000003 ## indicates the number Second operation node; 2 shows the second startup of the server; 00000003 The first few operations (hexadecimal) under the current startup created this node; ctime = Sat Jan 26 15:23:14 CST 2019 ##Creation time, long timestamp ; MZxid = 0x400000016 ##The 00000016 modification of the table server at the 4th start The data of this node; mtime = Sun Jan 27 00:< span style="color: #800080;">53:19 CST 2019 #Modification time pZxid = 0x400000015 ##The table server created a child node at the 00000015th time when it was started for the 4th time; cversion = 1 dataVersion = 1< span style="color: #000000;"> #Modified data aclVersion = 0 #access control listAccess control list, access control of the network version; control who can access the node on the network; version 0 is acl ephemeralOwner = 0x0  #If it is a temporary node, 0x0 here will not be 0; if it is a temporary node, it will show the owner; when your owner goes offline, it It disappeared naturally; dataLength = 2 #data length numChildren = < span style="color: #800080;">1 #The number of child nodes; [zk: localhost:2181(CONNECTED) 1] create -e /testXXX 123 Created /testXXX [zk: localhost:2181(CONNECTED) 2] stat /testXXX cZxid = 0x400000019 ctime = Sun Jan 27 01:19: 06 CST 2019 mZxid  span>= 0x400000019 mtime = Sun Jan 27 01:19:06 CST 2019 pZxid = 0x400000019 cversion = 0 dataVersion = 0 aclVersion = 0 ephemeralOwner = 0x1688adaecec0006 span> #sessionid = 0x1688adaecec0006 The session id between client and server; dataLength = < span style="color: #800080;">3 numChildren = 0 [zk: localhost:2181(CONNECTED) 3] delete deletes no child nodes; rmr can delete child nodes; rmr /test2 

[[emailprotected] zookeeper-3.4.10]$ ssh hadoop103 /opt/module/zookeeper-3.4.10/bin/zkServer.sh status 
ZooKeeper JMX enabled by default Using config: /opt/module/zookeeper-3.4.10/bin/../conf/zoo.cfg Error contacting service. It is probably not running. ### Note that there must be a space in the source [[email protected] zookeeper-3.4.10]$ ssh hadoop103  source /etc/profile && /opt/module/zookeeper-3.4.10/bin/zkServer.sh status ZooKeeper JMX enabled by default  Using config: /opt/module/zookeeper-3.4.10/bin/ ../conf/zoo.cfg Mode: follower [[email protected] zookeeper-3.4.10] $ hadoop 103 To configure the environment variables of JAVA_HOME:: #JAVA_HOME export JAVA_HOME=/opt/module/jdk1.8.0_144 export PATH=$ PATH:$JAVA_HOME/bin [[email protected] zookeeper-3.4.10]$ ssh hadoop103 /opt/module/zookeeper-3.4.10/bin/< span style="color: #000000;">zkServer.sh status ZooKeeper JMX enabled by default Using config: /opt/module/zookeeper-3.4.10/bin/../conf/zoo.cfg Mode: follower< /span>

public class Zookeeper{



 private ZooKeeper zkClient;

 public static final String CONNECT_STRING = "hadoop101:2181,hadoop102:2181,hadoop103:2181";

 public static final int SESSION_TIMEOUT = 2000;

 @Before

 public void before() throws IOException {

 //cluster address; session expiration time; anonymous inner class, callback Function; The main process will not stop, and the root node will be notified through the callback function.

 zkClient = new ZooKeeper(CONNECT_STRING, SESSION_TIMEOUT, new Watcher() { //When creating a ZooKeeper client

 public void process(WatchedEvent event) {

 System.out.println("default callback function"); //No node is monitored, write or not write

}

 });

}

 //Create

 @Test

 public void create() throws KeeperException, InterruptedException {

 String s = zkClient.create("/APITest", "123".getBytes(),

 ZooDefs.Ids.OPEN_ACL_UNSAFE, CreateMode.EPHEMERAL_SEQUENTIAL);//node Permission access; Temporary ordered node

 System.out.println(s);

 Thread.sleep(Long.MAX_VALUE);

}

 //recursive --> can be called repeatedly

 @Test

 public void getChildren() throws KeeperException, InterruptedException {

 List children = zkClient.getChildren("/", new Watcher() {//watch: true will monitor, call the default callback function, and monitor once is effective;

 // You can also write new Watch without the default callback function ;

 public void process(WatchedEvent watchedEvent) {

 try {

 System.out.println("Own callback function");

 getChildren(); //Can monitor repeatedly; monitor the root directory watcher .process(pair.event);

} catch (KeeperException e) {

 e.printStackTrace();

} catch (InterruptedException e) {

 e.printStackTrace();

}

}

 });

 for (String child: children) {

 System.out.println(child);

}

 System.out.println("==================");

}

 @Test //Can be called repeatedly and monitored repeatedly

 public void testGet() throws KeeperException, InterruptedException {

 getChildren();

 Thread.sleep(Long.MAX_VALUE); //The main thread is Blocking, indicating that the callback is not the main thread



}

 // Determine whether znode exists

 @Test

 public void exist() throws KeeperException, InterruptedException {

 Stat stat = zkClient.exists("/zookeeper1", false) ;

 if (stat == null){

 System.out.println("Node does not exist");

 }else{

 System.out.println(stat.getDataLength());

}

}

}

ClientCnxn.java: sendThread = new SendThread(clientCnxnSocket); //connect is sendThread responsible for network connection communication; eventThread = new EventThread(); 　　　　　　　//listener is eventThread responsible for monitoring. Two child threads are created here; class SendThread extends ZooKeeperThread public class ZooKeeperThread extends Thread public  void start() {sendThread.start(); The thread that sends information to zookeeper by the client; eventThread.start(); is notified by zookeeper changes eventThread is responsible for receiving event changes; eventThread is responsible for the callback function called. Zookeeper changes it and sends this change to eventThread }