TaskTracker are slave nodes that do the actual computation
related to a job in MapReduce part in Hadoop. It has a loop that sends
heartbeat method calls to jobtracker every 5 seconds , indicating its existence
and status of its task. JobTracker assigns
a new task when the old one is completed. Some TaskTracker performs map task
and after completing the map task, the reduce task is started.The output of the reducer task is the final output which is
submitted to the client by the JobTracker.TaskTracker has fixed number of slots, for map task and
reduce task.
No comments:
Post a Comment