[TOC]
### 1\. 安装环境要求
> you need to install Storm's dependencies on Nimbus and the worker machines. These are:
>
> 1. Java 7+ (Apache Storm 1.x is tested through travis ci against both java 7 and java 8 JDKs)
> 2. Python 2.6.6 (Python 3.x should work too, but is not tested as part of our CI enviornment)
按照[官方文档](http://storm.apache.org/releases/1.2.2/Setting-up-a-Storm-cluster.html) 的说明:storm 运行依赖于 Java 7+ 和 Python 2.6.6 +,所以需要预先安装这两个软件。由于这两个软件在多个框架中都有依赖,其安装步骤单独整理至 :
* [Linux 环境下 JDK 安装](https://github.com/heibaiying/BigData-Notes/blob/master/notes/installation/Linux%E4%B8%8BJDK%E5%AE%89%E8%A3%85.md)
* [Linux 环境下 Python 安装](https://github.com/heibaiying/BigData-Notes/blob/master/notes/installation/Linux%E4%B8%8BPython%E5%AE%89%E8%A3%85.md)
### 2\. 下载并解压
下载并解压,官方下载地址:[storm.apache.org/downloads.h…](http://storm.apache.org/downloads.html)
~~~
# tar -zxvf apache-storm-1.2.2.tar.gz
复制代码
~~~
### 3\. 配置环境变量
~~~
# vim /etc/profile
复制代码
~~~
添加环境变量:
~~~
export STORM_HOME=/usr/app/apache-storm-1.2.2
export PATH=$STORM_HOME/bin:$PATH
复制代码
~~~
使得配置的环境变量生效:
~~~
# source /etc/profile
复制代码
~~~
### 4\. 启动相关进程
因为要启动多个进程,所以统一采用后台进程的方式启动。进入到 `${STORM_HOME}/bin` 目录下,依次执行下面的命令:
~~~
# 启动zookeeper
nohup sh storm dev-zookeeper &
# 启动主节点 nimbus
nohup sh storm nimbus &
# 启动从节点 supervisor
nohup sh storm supervisor &
# 启动UI界面 ui
nohup sh storm ui &
# 启动日志查看服务 logviewer
nohup sh storm logviewer &
复制代码
~~~
### 5\. 验证是否启动成功
验证方式一:jps 查看进程:
~~~
[root@hadoop001 app]# jps
1074 nimbus
1283 Supervisor
620 dev_zookeeper
1485 core
9630 logviewer
复制代码
~~~
验证方式二: 访问 8080 端口,查看 Web-UI 界面:
![https://github.com/heibaiying](https://user-gold-cdn.xitu.io/2019/9/21/16d52da5fc871a03?imageView2/0/w/1280/h/960/format/webp/ignore-error/1)
作者:heibaiying
链接:https://juejin.cn/post/6844903950035091464
来源:掘金
著作权归作者所有。商业转载请联系作者获得授权,非商业转载请注明出处。