zookeeper 集群搭建

1.上传zk安装包

2.解压

3.配置（先在一台节点上配置）
3.1添加一个zoo.cfg配置文件
$ZOOKEEPER/conf
mv zoo_sample.cfg zoo.cfg

3.2修改配置文件（zoo.cfg）
initLimit=10
syncLimit=5
clientPort=2888
tickTime=2000
dataDir=/home/hadoop/software/zookeeper/data
dataLogDir=/home/hadoop/software/zookeeper/logs

server.1=Master:2888:3888
server.2=Slave0:2888:3888
server.3=Slave1:2888:3888
server.4=Slave2:2888:3888

解释：

zookeeper的默认配置文件为zookeeper/conf/zoo_sample.cfg，需要将其修改为zoo.cfg。其中各配置项的含义，解释如下：

1.tickTime：CS通信心跳时间
Zookeeper 服务器之间或客户端与服务器之间维持心跳的时间间隔，也就是每个 tickTime 时间就会发送一个心跳。tickTime以毫秒为单位。
tickTime=2000  

2.initLimit：LF初始通信时限
集群中的follower服务器(F)与leader服务器(L)之间初始连接时能容忍的最多心跳数（tickTime的数量）。
initLimit=5  

3.syncLimit：LF同步通信时限
集群中的follower服务器与leader服务器之间请求和应答之间能容忍的最多心跳数（tickTime的数量）。
syncLimit=2  
 
4.dataDir：数据文件目录
Zookeeper保存数据的目录，默认情况下，Zookeeper将写数据的日志文件也保存在这个目录里。
dataDir=/home/michael/opt/zookeeper/data  

5.clientPort：客户端连接端口
客户端连接 Zookeeper 服务器的端口，Zookeeper 会监听这个端口，接受客户端的访问请求。
clientPort=2181 

6.服务器名称与地址：集群信息（服务器编号，服务器地址，LF通信端口，选举端口）
这个配置项的书写格式比较特殊，规则如下：
server.N=YYY:A:B 

3.3在（dataDir=/home/hadoop/software/zookeeper/data）创建一个myid文件，里面内容是server.N中的N（server.2里面内容为2）
echo "5" > myid

3.4将配置好的zk拷贝到其他节点 
scp -r /home/hadoop/software/zookeeper/  Slave0:/home/hadoop/software/
scp -r /home/hadoop/software/zookeeper/  Slave1:/home/hadoop/software/

scp -r /home/hadoop/software/zookeeper/  Slave2:/home/hadoop/software/

3.5 注意：在其他节点上一定要修改myid的内容
在Slave0应该讲myid的内容改为2 
在Slave1应该讲myid的内容改为3 

在Slave2应该讲myid的内容改为4

4.启动集群
分别启动zk
./zkServer.sh start

 

安装zookeeper

hadoop@Master:~/software/zookeeper/bin$ ./zkServer.sh start
JMX enabled by default
Using config: /home/hadoop/software/zookeeper/bin/../conf/zoo.cfg
grep: /home/hadoop/software/zookeeper/bin/../conf/zoo.cfg: No such file or directory
grep: /home/hadoop/software/zookeeper/bin/../conf/zoo.cfg: No such file or directory
mkdir: cannot create directory \u2018\u2019: No such file or directory
Starting zookeeper ... ./zkServer.sh: line 131: /zookeeper_server.pid: Permission denied
FAILED TO WRITE PID

1. 修改权限：chmod 

hadoop@Master:~/software/zookeeper/bin$ cd ~/software/
hadoop@Master:~/software$ sudo chmod a+xwr zookeeper

2 . 重命令： mv

hadoop@Slave1:~/software/zookeeper/conf$ ls
configuration.xsl  log4j.properties  zoo_sample.cfg
hadoop@Slave1:~/software/zookeeper/conf$ mv zoo_sample.cfg zoo.cfg
hadoop@Slave1:~/software/zookeeper/conf$ ls
configuration.xsl  log4j.properties  zoo.cfg

7、启动zookeeper，zookeeper需要各个服务器单独启动，分别在master，slave0，slave1，slave2启动

hadoop@Master:~$ cd /home/hadoop/software/zookeeper/bin/
hadoop@Master:~/software/zookeeper/bin$ ls
null.next     zkCli.cmd  zkEnv.sh                zkServer.sh
README.txt    zkCli.sh   zkServer.cmd            zookeeper.out
zkCleanup.sh  zkEnv.cmd  zkServer-initialize.sh
hadoop@Master:~/software/zookeeper/bin$ zkServer.sh start
JMX enabled by default
Using config: /home/hadoop/software/zookeeper/bin/../conf/zoo.cfg
Starting zookeeper ... STARTED
hadoop@Master:~/software/zookeeper/bin$ jps
16178 QuorumPeerMain
16905 Jps

出错解答：  Error contacting service. It is probably not running.

hadoop@Master:~/software/zookeeper/bin$ ./zkServer.sh status
JMX enabled by default
Using config: /home/hadoop/software/zookeeper/bin/../conf/zoo.cfg
Client port found: 2181
Error contacting service. It is probably not running.

我的 zoo.cfg 文件，如下： 

initLimit=10
syncLimit=5
clientPort=2888
tickTime=2000
dataDir=/home/hadoop/software/zookeeper/data
dataLogDir=/home/hadoop/software/zookeeper/logs
dynamicConfigFile=/home/hadoop/software/zookeeper/bin/../conf/zoo.cfg.dynamic

server.1=Master:2888:3888
server.2=Slave0:2888:3888
server.3=Slave1:2888:3888
server.4=Slave2:2888:3888

试了网上很多方法，查看每台机子的/home/hadoop/software/zookeeper/data/myid， 确保每台机子的myid没出错

OK！

以后停止，启动，状态查看方式：

/usr/local/sbin/xcall.sh /home/hadoop/software/zookeeper/bin/zkServer.sh stop
/usr/local/sbin/xcall.sh /home/hadoop/software/zookeeper/bin/zkServer.sh start
/usr/local/sbin/xcall.sh /home/hadoop/software/zookeeper/bin/zkServer.sh status