0

AS i see the latest stable release of hadoop is 0.20.x. And latest release is 0.23.. Seems there are lot of chanages from .20. to 0.23.x.

We are able to set up small cluster with stable relase(0.20.2) and practicising mapreduce programming.

We have seen lot of new api's added in 0.23.x. In order to explore 0.23.x, we need to setup cluster also with 0.23.x release.

Could you guys point us a documentation, where we can set up cluster with 0.23.x release.

seems 0.23.x is completely different its not like 0.20.x when i untar the tar file. Please give us some book reference/doc where cluster set up is mentioned from begining.

Thanks MRK

Praveen Sripati
  • 32,799
  • 16
  • 80
  • 117
MRK
  • 573
  • 2
  • 8
  • 21
  • 3
    Accept some answers and we will give you a few links... – Thomas Jungblut Dec 12 '11 at 13:15
  • 1
    I know that everyone is busy, even others to respond to your query. But, I think there should be a minimal attempt (I don't say spend an hour) to make sure that the query heading and the actual query are properly written and formatted. This will tell others that you have spent sometime and effort in asking the query and makes them feel to post an answer back. Also, as Thomas mentioned, if you feel the answer is correct upvote and accept it and you will get better responses. Community is like a two-way line, the more you participate the more response you will get. Just my 2c. – Praveen Sripati Dec 13 '11 at 01:19

2 Answers2

2

The major difference between 0.23 and pre-0.23 release is that in 0.23 the resource management and the application life cycle management have been separated. Pre-0.23 allowed only MapReduce applications to run, but 0.23 allows other applications besides MapReduce. Already Hama, Giraph and some other applications have been ported and porting of MPI is in progress.

We have seen lot of new api's added in 0.23.x. In order to explore 0.23.x, we need to setup cluster also with 0.23.x release.

There hasn't been any differences in the user API, so the existing applications should run without any code changes, but configuration file changes are required. 0.23 release is backward compatible from an API perspective.

Here is the consolidated list of MRv2 architecture, videos, articles etc. I will try to keep them updated as I come across new information.

http://www.thecloudavenue.com/p/mrv2resources.html

Praveen Sripati
  • 32,799
  • 16
  • 80
  • 117
0

This is the official documentation for cluster setup in r0.23.0:

http://hadoop.apache.org/common/docs/r0.23.0/hadoop-yarn/hadoop-yarn-site/ClusterSetup.html

Donald Miner
  • 38,889
  • 8
  • 95
  • 118