Mahout has a script under the bin folder of the installation. Notice line 120 onwards of the following script:
# CLASSPATH initially contains $MAHOUT_CONF_DIR, or defaults to $MAHOUT_HOME/src/conf
CLASSPATH=${CLASSPATH}:$MAHOUT_CONF_DIR
if [ "$MAHOUT_LOCAL" != "" ]; then
echo "MAHOUT_LOCAL is set, so we don't add HADOOP_CONF_DIR to classpath."
elif [ -n "$HADOOP_CONF_DIR" ] ; then
echo "MAHOUT_LOCAL is not set; adding HADOOP_CONF_DIR to classpath."
CLASSPATH=${CLASSPATH}:$HADOOP_CONF_DIR
fiWe can set HADOOP_HOME and HADOOP_CONF_DIR to launch the Mahout job (algorithm) on the Hadoop cluster.
Just before the algorithm command, set the two previously mentioned parameters using the export command:
export HADOOP_HOME=<ur hadoop location> export HADOOP_CONF_DIR=$HADOOP_HOME/conf
The Mahout launcher script helps to launch the job locally or on a cluster.