public String createClusterHadoop(
String name, int nodes, String instance, Map<String, Boolean> apps) {
RunJobFlowResult result;
StepFactory stepFactory = new StepFactory();
StepConfig enabledebugging =
new StepConfig()
.withName("Enable debugging")
.withActionOnFailure("TERMINATE_JOB_FLOW")
.withHadoopJarStep(stepFactory.newEnableDebuggingStep());
List<Application> applications = new ArrayList<Application>();
if (apps.get("hive")) {
Application hive = new Application().withName("Hive");
applications.add(hive);
}
if (apps.get("hue")) {
Application hue = new Application().withName("Hue");
applications.add(hue);
}
if (apps.get("spark")) {
Application spark = new Application().withName("Spark");
applications.add(spark);
}
RunJobFlowRequest request =
new RunJobFlowRequest()
.withName(name)
.withReleaseLabel("emr-4.1.0")
.withSteps(enabledebugging)
.withApplications(applications)
.withServiceRole("EMR_DefaultRole")
.withJobFlowRole("EMR_EC2_DefaultRole")
.withInstances(
new JobFlowInstancesConfig()
.withInstanceCount(nodes)
.withKeepJobFlowAliveWhenNoSteps(true)
.withMasterInstanceType(instance)
.withSlaveInstanceType(instance));
result = emr.runJobFlow(request);
return result.getJobFlowId();
}
