You are viewing a plain text version of this content. The canonical link for it is here.
Posted to mapreduce-user@hadoop.apache.org by db...@aol.com on 2012/01/13 16:32:06 UTC

Newbie Question On Next Generation Map Reduce

I am learning Hadoop and had a question about writing Map Reduce jobs in newer versions of Hadoop.  

Is Next Generation Map Reduce a change just for our system administrator or do I (developer) already writing Map Reduce code have to start doing things differently too (and is there a guide that explains the changes I need to make)?

Re: Newbie Question On Next Generation Map Reduce

Posted by Bejoy Ks <be...@gmail.com>.
HI
     AFAIK on a map reduce application developer perspective there won't be
much changes. The APIs that you use are gonna be the same.  This ensure
that your existing map reduce applications can be deployed on Yarn based
cluster without any code change.

Regards
Bejoy

On Fri, Jan 13, 2012 at 9:02 PM, <db...@aol.com> wrote:

> I am learning Hadoop and had a question about writing Map Reduce jobs in
> newer versions of Hadoop.
>
> Is Next Generation Map Reduce a change just for our system administrator
> or do I (developer) already writing Map Reduce code have to start doing
> things differently too (and is there a guide that explains the changes I
> need to make)?
>

Re: Newbie Question On Next Generation Map Reduce

Posted by Robert Evans <ev...@yahoo-inc.com>.
We are trying very hard to maintain API compatibility and your code should not need anything more then a recompile, which may be required for some code to work properly.  However in practice with anything big like this there will be issues, and we are staring to run into some of them, that we are trying to fix.  I believe that for the most part the command line interface is more or less the same, but 0.23.X is not finished yet so there may be more changes that go in before we get a stable release of 0.23.

--Bobby Evans


On 1/13/12 9:32 AM, "dbadave85@aol.com" <db...@aol.com> wrote:

I am learning Hadoop and had a question about writing Map Reduce jobs in newer versions of Hadoop.

Is Next Generation Map Reduce a change just for our system administrator or do I (developer) already writing Map Reduce code have to start doing things differently too (and is there a guide that explains the changes I need to make)?