Storm 安装及搭建集群环境
搭建storm的集群环境
官方下载地址址:http://storm.apache.org/注意
:本博主使用版本 Storm 1.1.3 Released
a) 在hadoop-001主机解压apache-storm-1.1.3
tar -zxvf apache-storm-1.1.3.tar.gz
b) 配置环境变量
vi ~/.bash_profile
export STORM_HOME=/home/hadoop/bigdatasoftware/apache-storm-1.1.3
PATH=$PATH:$STORM_HOME/bin
c) 编写conf/storm-env.sh文件
export JAVA_HOME=/home/hadoop/bigdatasoftware/jdk1.8.0_161
export STORM_CONF_DIR=/home/hadoop/bigdatasoftware/apache-storm-1.1.3/conf
d) 分别在不同的主机hadoop-001、hadoop-002、hadoop-003,新建目录
mkdir -p /home/hadoop/stormdata
e) 编辑conf/storm.yaml文件
storm.zookeeper.servers:
- "hadoop-001"
- "hadoop-002"
- "hadoop-003"
storm.local.dir: "/home/hadoop/stormdata"
nimbus.host: "hadoop-001"
supervisor.slots.ports:
- 6700
- 6701
- 6703
- 6704
f) 将storm分发到不同的主机
scp -r /home/hadoop/bigdatasoftware/apache-storm-1.1.3 hadoop@hadoop-002:/home/hadoop/ bigdatasoftware /
scp -r /home/hadoop/ bigdatasoftware /apache-storm-1.1.3 hadoop@hadoop-003:/home/hadoop/ bigdatasoftware /
scp -r ~/.bash_profile hadoop@hadoop-002:~/.bash_profile
scp -r ~/.bash_profile hadoop@hadoop-003:~/.bash_profile
分发完成脚本文件,必须 source一下
source ~/.bash_profile
g) 启动集群的storm环境(注意必须要在bin目录下启动)
nohup sh storm nimbus &(hadoop-master主机启动)
nohup sh storm ui & (hadoop-master主机启动)
nohup sh storm supervisor &(hadoop-001主机启动、hadoop-003主机启动、hadoop-002主机启动)
nohup sh storm logviewer &(hadoop-001主机启动、hadoop-003主机启动、hadoop-002主机启动)
h) 查看进程(输入jps)
i) 打开浏览器:输入http://hadoop-001:8080/index.html(注意配置C:\Windows\System32\drivers\etc\ hosts)