the installation process of hadoop for linux(virtual machine) and use the java API and hadoop to interact

After a few days,i finally finished my homework about the data class.the work is aimed at installing the environment of hadoop on linux,it looks so easy , but there are so many error when i do my work ,so i write down my experience in this blog. first, my linux environment is 32-bit centos 6.7 ,after add the user and configure the ssh. i start to install jdk1.7 for os,hadoop need the jdk,when i use the yum which the tools was used to install some software to install jdk,it occured the fist problem,through the internet,the problem is the source of yum , the yum need to change the source of aliyun , you need use the code berfore to change the source
then you can install jdk ffor your linux successfully.then you can install hadoop,i have a suggesstion about the hadoop version 2.x is a good choice, at first,using the hadoop 3.x led the unmatch of java version and hadoop version,if you already use the 3.x ,just go on,if you can not check the hadoop version in your computer,if you are a novice in linux, you had better to change a new virtual machine ,choose the hadoop 2.x and then repeat the process that you had done before,but i never install the jdk1.8 and then install the hadoop 3.x,so you can try it.
if you can install the hadoop sucessfully,then you just need a ide to write code and then use java API and hadoop to interact , you can choose the intellijidea and eclipse , i choose the eclipse as my ide,after installing the ide,it semms that it`s almost over,then it occuered a inexplicable problem , when i use eclipse to open a file ,the eclipse flashback crazily.use the almost idea in internet,the problem can not be solved,when i change the ide ,another ide can not launch because the linux lack the realpath tools,when i search how to install realpath in linux,there are not a way can help me ,so i decide to change the linux from centos to ubuntu ,and repeat the process above all.
when i use the ubuntu to install hadoop,there is no any serious problem,just one point,we need to allocate a little more memory to the ubuntu,because the eclipse need more memeory.
if you have some question,you can email me(2489321713@qq.com).