mapreduce_on_yarn.txt

### RUN A MAP REDUCE JOB ON YARN 
### USING hadoop-2.7.3 AND CENTOS7 

# If needed : follow this tutorial for Hadoop 2.7 single node installation and configuration on RHEL7/CentOS7
# http://fibrevillage.com/storage/610-hadoop-2-7-single-node-installation-and-configuration-on-rhel7-centos7

cd /usr/local/hadoop-2.7.3

# Use jps (Java Virtual Machine Process Status Tool) to display the nodes
jps

# Start YARN
sbin/start-yarn.sh

# Display the nodes
# Node manager must now appear
jps

# List of programs (example)
bin/hadoop jar share/hadoop/mapreduce/hadoop-mapreduce-examples-2.7.3.jar 

# Run a MapReduce job using one a program 
# Example : using pi with number of maps : 4 and damples per map : 4
bin/hadoop jar share/hadoop/mapreduce/hadoop-mapreduce-examples-2.7.3.jar pi 4 4

# In a web browser, enter :   http://ip_addr:8088/cluster
# Access port 8088 for information about cluster and all applications

# Stop the daemons
sbin/stop-yarn.sh


### IF NEEDED : Solve the error “file not find exception” 

# Amend JAVA_HOME variable
# Note :  Use /JAVA_HOME to look for the variable
# TO DO : remove the brackets after equal in both lines
vi libexec/hadoop-config.sh

# Stop the daemons
sbin/stop-yarn.sh

### RUN AGAIN A MAP REDUCE JOB
# Restart the daemons
sbin/start-yarn.sh

# Run a MapReduce job
bin/hadoop jar share/hadoop/mapreduce/hadoop-mapreduce-examples-2.7.3.jar pi 4 4