I read that setting MAHOUT_LOCAL="TRUE" can set Mahout to run in local mode, implying that I can read files from the local system instead of HDFS. I want to know if running the mahout spark-itemsimilarity job will run in a distributed fashion in the Spark cluster or will it be a single machine execution? I have set the --master option to the Spark Master URL.
PS I want to read from the local file system, not from HDFS. Also, I want to utilize the complete Spark cluster.