This is the docker image for intel-hibench implementation. We build a pseudo single-node environment on Ubuntu 14.04. Before building/running, please check the file
$./hibench-docker.conf
for configuration and please set the versions according to your own needs. Otherwise, it will progress by default.
Launch the container by executing:
$./scripts/run_container.sh cdh
OR
$./scripts/run_container.sh open-source
The script will launch all related services and run a workload (by default: wordcount) after initialization. When entering the container of the hibench-docker, you are normally guided to a root user, please modify configurations under the HiBench directory :
#cd ${HIBENCH_HOME}
After running any workload, please check this file
${HIBENCH_HOME}/report/hibench.report
for detailed report.
If there is anything wrong with a certain workload, please check this final configuration file:
${HIBENCH_HOME}/report/${workload}/${method}/conf/${workload}.conf
Remember to replace the workload variable with the actual one (e.g. wordcount) and the method variable with one of ( mapreduce, spark/scala, spark/java, spark/python ).