Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Hadoop, mail # dev - difference between hadoop code and the amazon elastic Mapreduce


Copy link to this message
-
Re: difference between hadoop code and the amazon elastic Mapreduce
Eli Collins 2012-09-09, 20:04
Peter,

Thanks for the info.  Do you guys plan to contribute the rewritten s3
code (assume you're referring to org.apache.hadoop.fs.s3) back to
Apache?

Thanks,
Eli

On Sun, Sep 9, 2012 at 12:38 PM, Sirota, Peter <[EMAIL PROTECTED]> wrote:
> Hi,
>
> The major differences are in s3 file system that has been rewritten in EMR and in Hadoop interactions with S3. Other differences are in detecting various failure conditions.
>
> Outside these it's Apache Hadoop.  Here is a list of patches EMR applied on top of 1.0.3 Hadoop
> http://docs.amazonwebservices.com/ElasticMapReduce/latest/DeveloperGuide/EnvironmentConfig_AMIHadoopPatches.html
>
> Regards,
> Peter
>
>
>
> On Sep 9, 2012, at 11:29 AM, "Momina Khan" <[EMAIL PROTECTED]> wrote:
>
>> hi all!
>>
>> could someone please point out key differences between hadoop code and
>> Amazon's Elastic MapReduce. I am particularly interested in ways that
>> hadoop code is changed/optimized to run on efficiently EC2.
>>
>> cheers!
>> momina