Why there are no job running on hadoop

Using hadoop1.3.0. I ran the example WordCount correctly in eclipse. But when I enter localhost:50030, it said running jobs and retired jobs were both none. The truth was I had already gotten the correct output. Did I do something wrong with it?

You need not run map-reduce programs from eclipse.

 You need to create a runnable jar of it and run it from a separate VM using hadoop command.