虚拟机上安装你的第一个hadoop测试集群
git仓库
https://github.com/fencex/hadoop-vagrant
目标
我们的目标是安装五个节点的集群,3台data node 2台namenode
5 nodes
| NN | DN | ZK | ZKFC | JN | RM | NM | |
|---|---|---|---|---|---|---|---|
| node01 | √ | √ | √ | √ | √ | ||
| node02 | √ | √ | √ | √ | √ | ||
| node03 | √ | √ | √ | √ | |||
| node04 | √ | √ | |||||
| node05 | √ | √ |
准备
获取 vagrant
下载 jdk, zookeeper, hadoop 二进制文件
https://www.oracle.com/technetwork/java/javase/downloads/jdk8-downloads-2133151.htmlget jdk-8u212-linux-x64.tar.gzwget http://mirror.bit.edu.cn/apache/zookeeper/zookeeper-3.4.14/zookeeper-3.4.14.tar.gzwget http://mirror.bit.edu.cn/apache/hadoop/common/hadoop-2.8.5/hadoop-2.8.5.tar.gz
开始安装
git clone git@github.com:fencex/hadoop-vagrant.git
cd hadoop-vagrant
## generate key for nodes
ssh-keygen -f ssh-key
cat ssh-key.pub >> authorized_keys
## add your pub key
cat ~/.ssh/id_rsa.pub >> authorized_keys
## add vagrant pub key
ssh-keygen -y -f ~/.vagrant.d/insecure_private_key >> authorized_keys
## fire up
./command/up.sh
第一次初始化
## 在 node01 node02 node03 执行
zkServer.sh start
## 在 node01 执行
hadoop-daemon.sh start journalnode
## node04
hdfs namenode -format
hadoop-daemon.sh start namenode
## node05
hdfs namenode -bootstrapStandby
hadoop-daemon.sh start namenode
## node05
hdfs zkfc -formatZK
## then stop all
stop-all.sh
开机
## node01 node02 node03
zkServer.sh start
## node01
start-dfs.sh
## node01
start-yarn.sh
vagrant的相关操作
## up
./command/up.sh
## suspend
./command/suspend.sh
## resume suspended
./command/resume.sh
## shutdown
./command/halt.sh
检查是否安装成功

欢迎订阅公众号「大数据学习指北」,
记住能力越大,薪资越高
💰
👆
