The explosion of accessible human-generated information necessitates automated analytical processing to cluster, classify, and filter this information. The MapReduce paradigm has emerged as a popular approach to handling large-scale analysis, farming out requests to a cluster of nodes that first perform filtering and transformation of the data (map) and then aggregate the results (reduce). The Data Analytics benchmark is included in CloudSuite to cover the increasing importance of machine learning tasks analyzing large amounts of data in datacenters using the MapReduce framework. It is composed of Mahout, a set of machine learning libraries, running on top of Hadoop, an open-source implementation of MapReduce.
The benchmark consists of running a Naive Bayes classifier on a Wikimedia dataset. It uses Hadoop version 2.7.3 and Mahout version 0.12.2.
To obtain the images:
$ docker pull cloudsuite/hadoop $ docker pull cloudsuite/data-analytics
The benchmark is designed to run on a Hadoop cluster, where the single master runs the driver program, and the slaves run the mappers and reducers.
First, create a network to isolate your Hadoop cluster:
$ docker network create hadoop-net
Start the master with:
$ docker run -d --net hadoop-net --name master --hostname master \ cloudsuite/data-analytics master
Start a number of slaves with:
$ docker run -d --net hadoop-net --name slave01 --hostname slave01 \ cloudsuite/hadoop slave $ docker run -d --net hadoop-net --name slave02 --hostname slave02 \ cloudsuite/hadoop slave ...
Note that it is important to set hostnames on docker containers, and that the hostname should be the same as the name. If master’s name/hostname isn’t
master, then it should be supplied to slaves when running the containers as an argument after
Run the benchmark with:
$ docker exec master benchmark