zookeeper-3.4.6 集群搭建

摘要: hadoop生态系统和spark生态系统越来越多的开始依赖zookeeper，所以搭建一套zookeeper集群几乎成了大数据平台的必须工作之一

我们采用zookeeper-3.4.6来搭建zookeeper集群，硬件为三台DELL服务器（已经关闭防火墙并且配置好ssh互信）：

192.168.4.33 bi3

192.168.4.34 bi4

192.168.4.35 bi5

1. 解压zookeeper-3.4.6安装包到/home/hadoop/work目录下：/home/hadoop/work/zookeeper-3.4.6

2. 配置conf/zoo.cfg

# the directory where the snapshot is stored.
# do not use /tmp for storage, /tmp here is just
# example sakes.
dataDir=/home/hadoop/work/zookeeper-3.4.6/var/data
dataLogDir=/home/hadoop/work/zookeeper-3.4.6/var/datalog
# the port at which the clients will connect
clientPort=2181
server.1=bi3:2888:3888
server.2=bi4:2888:3888
server.3=bi5:2888:3888

3. 新建目录

mkdir /home/hadoop/work/zookeeper-3.4.6/var/data
mkdir /home/hadoop/work/zookeeper-3.4.6/var/datalog

4. 编辑/home/hadoop/work/zookeeper-3.4.6/var/data/myid，三台机器的内容分别如下：

192.168.4.33 bi3	1
192.168.4.34 bi4	2
192.168.4.35 bi5	3

启动zookeeper集群，分别在三台机器上执行下面的命令

/home/hadoop/work/zookeeper-3.4.6/bin/zkServer.sh start

验证集群是否可用

[hadoop@bi5 apps]$ /home/hadoop/work/zookeeper-3.4.6/bin/zkCli.sh -server 127.0.0.1:2181

进入zookeeper的client命令行模式，在根目录下创建/zktest，然后查看是否创建成功

[zk: 127.0.0.1:2181(CONNECTED) 4] create /zktest test
Created /zktest
[zk: 127.0.0.1:2181(CONNECTED) 5] ls /
[consumers, zktest, rmstore, mytest, spark, yarn-leader-election, spark-test, controller_epoch, dubbo, hbase, hadoop-ha, zookeeper, admin, config, brokers]
[zk: 127.0.0.1:2181(CONNECTED) 6] get /zktest
test
cZxid = 0xf0006967b
ctime = Mon Feb 15 17:19:32 CST 2016
mZxid = 0xf0006967b
mtime = Mon Feb 15 17:19:32 CST 2016
pZxid = 0xf0006967b
cversion = 0
dataVersion = 0
aclVersion = 0
ephemeralOwner = 0x0
dataLength = 4
numChildren = 0