Benchmarks for an Hadoop Cluster

http://www.michael-noll.com/blog/2011/04/09/benchmarking-and-stress-testing-an-hadoop-cluster-with-terasort-testdfsio-nnbench-mrbench/

Benchmarking consists of running a number of tests stressing the Hadoop cluster to measure its maximum computing capabilities and validate its configuration.

In the Hadoop context of this script, Benchmarking consists of determining the DataNode's resources (disk, cpu, ram), running the following tests multiple times and extract reporting informations :

• TeraSort suite

  module.exports = use: test_user: module: 'ryba/commons/test_user', local: true, implicit: true configure: 'ryba/benchmark/configure' commands: 'benchmark' : [ 'ryba/benchmark/discover' 'ryba/benchmark/terasort' ]