Ulhas, As Roman said the documentation for 0.23 release is spread out. I tried to consolidate the information I come across. The below link has references to Architecture, Code, Video, Build, Deploy etc. It would be better to get sometime like this into the Apache documentation.
http://hadoop-tips.blogspot.com/p/mrv2-resources.html Regards, Praveen On Wed, Dec 7, 2011 at 1:21 AM, Roman Shaposhnik <r...@apache.org> wrote: > Hi Ulhas, > > On Tue, Dec 6, 2011 at 12:20 AM, Ulhas Shringare > <shringareulha...@gmail.com> wrote: > > Hello, > > I need to know steps to install Hadoop 0.23 , every where there are steps > > to install hadoop 0.20. > > And there are bit difference between them.Like conf/hadoop-env config > file > > is not there. > > Please provide me suitable information or link. > > I agree with you that the information on configuring .23 is scarce (it > is actually > available if you dig for it, but you have to be persistent). It also > depends on > what OS you would like to run it on. > > If you happen to be on RPM-based system, you can utilize the > *pre-release* packages > that Bigtop is publishing over here: > > http://bigtop01.cloudera.org:8080/view/Hadoop%200.23/job/Bigtop-23-matrix/label=centos5/lastSuccessfulBuild/artifact/output/ > > http://bigtop01.cloudera.org:8080/view/Hadoop%200.23/job/Bigtop-23-matrix/label=sles11/lastSuccessfulBuild/artifact/output/ > > You can simply follow the instructions on how to install that are > published on our > wiki, provided that you change the URL for the repo file to the ones > mentioned above: > > https://cwiki.apache.org/confluence/display/BIGTOP/How+to+install+Hadoop+distribution+from+Bigtop > > This will give you a pseudo distributed cluster that works and you can > inspect what > kind of settings we had to put in place. All the config files are > found in the usual Linux > place of /etc/hadoop/conf and /etc/yarn/conf. > > If you want to deploy to a multinode configuration, you can use the puppet > code > that we've created: > > https://svn.apache.org/repos/asf/incubator/bigtop/branches/hadoop-0.23/bigtop-deploy/puppet/ > > You simply need to run the following command on every node of your cluster: > FACTER_hadoop_head_node="FQDN" puppet --modulepath > bigtop-deploy/puppet/modules bigtop-deploy/puppet/manifests/site.pp > where FQDN is a fully qualified domain name for a node that you want all > the > "centralized" services to run on (NameNode, etc.) > > Hope this helps. And remember -- instead of jobtracker there's now > resoruce manager > and instead of tasktracker there's now a nodemanger. So the services are > name > accordingly. > > Thanks, > Roman. >
