TensorFlowOnSpark安装教程

• 各个环境版本
  OS：Ubuntu16.04 LTS
  java：jdk 1.8
  scala:2.10.4
  hadoop:2.7.3
  spark:2.1.0
  TensorFlowOnSpark:0.12.1

• 安装Hadoop
  需要配置3台虚拟机，分别为master，slave，slave2。先配置主节点master，别的slave节点的虚拟机可copy master。

  安装SSH并配置无密码登录
  

sudo apt-get install openssh-server
cd ~/.ssh/
ssh-keygen -t rsa
cat ./id_rsa.pub >> ./authorized_keys

安装java环境
下载jdk：http://www.oracle.com/technetwork/java/javase/downloads/jdk8-downloads-2133151.html
配置环境变量

vim .bashrc

添加export JAVA_HOME=/usr/lib/jvm/java-8-openjdk-amd64

source .bashrc

输入java -version,检验是否配置成功

zangtt@master:~$ java -version
openjdk version "1.8.0_121"
OpenJDK Runtime Environment (build 1.8.0_121-8u121-b13-0ubuntu1.16.04.2-b13)
OpenJDK 64-Bit Server VM (build 25.121-b13, mixed mode)

HADOOP
下载：http://hadoop.apache.org/releases.html，解压到指定目录

Hadoop集群配置：
1.core-site.xml