hadoop/hadoop-mapreduce-project/INSTALL

To compile  Hadoop Mapreduce next following, do the following:

Step 1) Install dependencies for yarn

See http://svn.apache.org/repos/asf/hadoop/common/trunk/hadoop-mapreduce/hadoop-yarn/README
Make sure protbuf library is in your library path or set: export LD_LIBRARY_PATH=/usr/local/lib

Step 2) Checkout

svn checkout http://svn.apache.org/repos/asf/hadoop/common/trunk

Step 3) Build common

Go to common directory - choose your regular common build command
Example: mvn clean install package -Pbintar -DskipTests

Step 4) Build HDFS 

Go to hdfs directory
ant veryclean mvn-install -Dresolvers=internal 

Step 5) Build yarn and mapreduce

Go to mapreduce directory
export MAVEN_OPTS=-Xmx512m

mvn clean install assembly:assembly -DskipTests

Copy in build.properties if appropriate - make sure eclipse.home not set
ant veryclean tar -Dresolvers=internal 

You will see a tarball in
ls target/hadoop-mapreduce-0.23.0-SNAPSHOT-all.tar.gz  

Step 6) Untar the tarball in a clean and different directory.
say YARN_HOME. 

Make sure you aren't picking up avro-1.3.2.jar, remove:
  $HADOOP_COMMON_HOME/share/hadoop/common/lib/avro-1.3.2.jar
  $YARN_HOME/lib/avro-1.3.2.jar

Step 7)
Install hdfs/common and start hdfs

To run Hadoop Mapreduce next applications: 

Step 8) export the following variables to where you have things installed:
You probably want to export these in hadoop-env.sh and yarn-env.sh also.

export HADOOP_MAPRED_HOME=<mapred loc>
export HADOOP_COMMON_HOME=<common loc>
export HADOOP_HDFS_HOME=<hdfs loc>
export YARN_HOME=directory where you untarred yarn
export HADOOP_CONF_DIR=<conf loc>
export YARN_CONF_DIR=$HADOOP_CONF_DIR

Step 9) Setup config: for running mapreduce applications, which now are in user land, you need to setup nodemanager with the following configuration in your yarn-site.xml before you start the nodemanager.
    <property>
      <name>nodemanager.auxiluary.services</name>
      <value>mapreduce.shuffle</value>
    </property>

    <property>
      <name>nodemanager.aux.service.mapreduce.shuffle.class</name>
      <value>org.apache.hadoop.mapred.ShuffleHandler</value>
    </property>

Step 10) Modify mapred-site.xml to use yarn framework
    <property>    
      <name> mapreduce.framework.name</name>
      <value>yarn</value>  
    </property>

Step 11) Create the following symlinks in $HADOOP_COMMON_HOME/share/hadoop/common/lib

ln -s $YARN_HOME/modules/hadoop-mapreduce-client-app-0.23.0-SNAPSHOT.jar .	
ln -s $YARN_HOME/modules/hadoop-yarn-api-0.23.0-SNAPSHOT.jar .
ln -s $YARN_HOME/modules/hadoop-mapreduce-client-common-0.23.0-SNAPSHOT.jar .	
ln -s $YARN_HOME/modules/hadoop-yarn-common-0.23.0-SNAPSHOT.jar .
ln -s $YARN_HOME/modules/hadoop-mapreduce-client-core-0.23.0-SNAPSHOT.jar .	
ln -s $YARN_HOME/modules/hadoop-yarn-server-common-0.23.0-SNAPSHOT.jar .
ln -s $YARN_HOME/modules/hadoop-mapreduce-client-jobclient-0.23.0-SNAPSHOT.jar .

Step 12) cd $YARN_HOME

Step 13) bin/yarn-daemon.sh start resourcemanager

Step 14) bin/yarn-daemon.sh start nodemanager

Step 15) bin/yarn-daemon.sh start historyserver

Step 16) You are all set, an example on how to run a mapreduce job is:
cd $HADOOP_MAPRED_HOME
ant examples -Dresolvers=internal 
$HADOOP_COMMON_HOME/bin/hadoop jar $HADOOP_MAPRED_HOME/build/hadoop-mapreduce-examples-0.23.0-SNAPSHOT.jar randomwriter -Dmapreduce.job.user.name=$USER -Dmapreduce.clientfactory.class.name=org.apache.hadoop.mapred.YarnClientFactory -Dmapreduce.randomwriter.bytespermap=0.23.000 -Ddfs.blocksize=536870912 -Ddfs.block.size=536870912 -libjars $YARN_HOME/modules/hadoop-mapreduce-client-jobclient-0.23.0-SNAPSHOT.jar output 

The output on the command line should be almost similar to what you see in the JT/TT setup (Hadoop 0.20/0.21)