On Sun, Apr 25, 2010 at 4:28 PM, Praveen Sripati <[email protected]> wrote: > > Hi, > > The MapReduce tutorial specifies that > >>> The Hadoop Map/Reduce framework spawns one map task for each InputSplit >>> generated by the InputFormat for the job. > > But, the mapred.map.tasks definition is > >>> The default number of map tasks per job. Ignored when mapred.job.tracker >>> is "local". > > So, is the number of map tasks spawned based on the number of InputSplits > generated or based on the mapred.map.tasks value specified in the > configuration file?
The former.
