Thoughts on Hadoop application development under Eclipse

Recently in the study of Hadoop, first through the online a lot of tutorials to build a fully distributed environment, in the distributed environment to run WordCount is also successful, then intend to write their own Mr to deal with the actual business, before the real development began to build a good development environment ah, So I think of Eclipse, but then we have a lot of problems:

1, the first is to look for a Hadoop plug-in, installation can directly build Mr Project can also run, but did not expect to use Hadoop plug-in with the development of windows, but the 2 problems encountered is the corresponding version of the plug-in is difficult to find, their compilation is also more trouble, But if you can compile your own success that is best, then installed a desktop version of Ubuntu, and then installed on this server eclipse, but due to plug-in problems, the plug-in cannot establish a connection.

2, can not use plug-ins, also do not have Windows environment, I under the eclipse of Linux under the new Java Project has imported all the package and server-side configuration files, but found that the wordcount can run, their own written Mr Program will not run, Later found that the remote development and debugging understanding of the wrong, to run Mr First to package him into a jar and then upload to the distributed environment before you can run on Eclipse, view the running log and so on. The main reference is the following article

http://qindongliang1922.iteye.com/blog/2036619

This article is from the "Seavar" blog, make sure to keep this source http://4397407.blog.51cto.com/4387407/1435289