The FILE and HDFS are indicators of the type of FileSystem the counter
is for. LocalFileSystem => FILE, and DistributedFileSystem => HDFS. In
Map+Reduce jobs, LocalFileSystem is used for the transient data, so it
would show up information about how many bytes were written and read
across the phases.
On Sun, Apr 24, 2011 at 4:52 PM, Pedro Costa <psdc1978@gmail.com> wrote:
> Hi,
>
> the Hadoop MapReduce counters has the parameters FILE_BYTES_READ,
> FILE_BYTES_WRITTEN, HDFS_BYTES_READ and HDFS_BYTES_WRITTEN.
> What the FILE and HDFS counters represents?
>
>
> Thanks,
>
>
>
> --
> Pedro
>
--
Harsh J
|