Hi,
>>so I wanted to try and lower the number to 10 and see how the performance is
The number of mappers is provided as only a hint to the framework, it is not
guaranteed to be that number.
>>I have been digging around in the hadoop source code and it looks like the
>>JobClient actually sets the
Hi Teryl
The number of mapper is determined by the InputFormat you use, in your case,
one way is to merge the files to large file beforehand, or use the
CombineFileInputFormat as your InputFormat.
On Mon, Jan 18, 2010 at 1:05 PM, Teryl Taylor wrote:
> Hi everyone,
>
> I'm playing around with
Hi everyone,
I'm playing around with the Hadoop map reduce library and I'm getting some
odd behaviour. The system is setup on one machine using the pseudo
distributed configuration. I use KFS as my DFS. I have written a
MapReduce program to process a bunch of binary files. The files are
compr
