Currently working with Hadoop 2.0.4-alpha-rc2 and a stable snapshot of Mahout 0.8
Confirmed working: pagerank wordcount sort bayes kmeans
The rest are untested
Steps to set up:
-
Modify HiBench/bin/hibench-config.sh
-
Make sure that hadoop is installed into Maven: navigate to hadoop root and mvn install
-
Install Mahout: navigate to HiBench/common/mahout-trunk and run the following command:
$ mvn -Phadoop-0.23 clean install -DskipTests -Dhadoop.version=2.0.4-alpha -Dmahout.skip.distribution=false
Further points to note:
- All jobs are set up to use the 'default' queue. Changing this is difficult.