Create a test user named "testuser" on the client host, the Linux cluster, and in HDFS, then log in to the client host as user.

hdfs dfs -put /tmp/sample-test.txt /user/testuser

Export the required environment variables for Mahout:

export JAVA_HOME=<your jdk home install location here>
export HADOOP_HOME=/usr/hdp/current/hadoop-client
export MAHOUT_HOME=/usr/hdp/current/mahout-client
export PATH="$PATH":$HADOOP_HOME/bin:$MAHOUT_HOME/bin
export CLASSPATH="$CLASSPATH":$MAHOUT_HOME

Upload a few megabytes of natural-language plain text to the client host as /tmp/sample-test.txt.

Transfer the sample-test.txt file to a subdirectory of the testuser's HDFS home directory.

hdfs dfs -mkdir /user/testuser/testdata
hdfs dfs -put /tmp/sample-test.txt /user/testuser/testdata

Create a mahout test output directory:

hdfs dfs -mkdir /user/testuser/mahouttest

Use the following command to instruct Mahout to convert the plain text file sample-test.txt into a sequence file that is in the output directory mahouttest:

mahout seqdirectory --input /user/testuser/testdata --output /user/ testuser/mahouttest -ow --charset utf-8