Hi, You can use two options: - From operational point of view, using JobTracker link (you can open it in a browser). Here you can see the current running progress which shows you the number of tasks and related counters as well - From coding point of view, you can use Job.getNumReduceTasks() (name of API might not be exactly same). Basically, for a running job, these statistics should be associated with a "Job", so check the API's related to "Job"
On Mon, Mar 5, 2012 at 1:42 AM, 张建 <edit...@gmail.com> wrote: > dear > I'm sorry for my poor in English. > I'm confused about something. > Recently, i read a paper names "TeraByte Sort on Apache Hadoop" > which was written by Owen O'Malley. And i see the graph "Running Tasks > Over Time" ,how can i know the tasks of map, reduce ,shuffle, merge at a > certain time ? > Thank you for your help ! > -- Regards, Piyush Kansal