3. Start MapReduce

Execute these commands from the JobTracker server:

Login as $HDFS_USER
/usr/lib/hadoop/bin/hadoop fs -mkdir /mapred
/usr/lib/hadoop/bin/hadoop fs -chown -R mapred /mapred

Login as $MAPRED_USER
/usr/lib/hadoop/bin/hadoop-daemon.sh --config $HADOOP_CONF_DIR start jobtracker

Execute these commands from the JobHistory server:

Login as $MAPRED_USER
/usr/lib/hadoop/bin/hadoop-daemon.sh --config $HADOOP_CONF_DIR start historyserver

Execute these commands from all TaskTracker nodes:

Login as $MAPRED_USER
/usr/lib/hadoop/bin/hadoop-daemon.sh --config $HADOOP_CONF_DIR start tasktracker