I have Spark application written in Java and executing it via AWS EMR. I want to get the ID of the EMR cluster inside my Java code. I have tried using below:

String emrClusterID = System.getenv("EMR_CLUSTER_ID");

but it returns null cluster ID. I do not want to use the EMR APIs to get the list of running clusters and then get the ID from there because I do not know inside the code what is the cluster name and there can be multiple clusters in Running state with the same name. So, how can I get the cluster ID of the running cluster from inside the code in Spark Java application?

CodePudding user response：

You can read and parse the JSON file /mnt/var/lib/info/job-flow.json on the EMR servers local filesystem.

The attribute jobFlowId is the clusterId.

A basic implementation (missing error handling) could be something like this:

import java.io.File;
import java.util.Map;
import com.fasterxml.jackson.databind.ObjectMapper;

public class EmrInfo {

  static final File EMR_JOB_FLOW = new File("/mnt/var/lib/info/job-flow.json");

  public static String getEmrId() {
      ObjectMapper mapper = new ObjectMapper();
      Map<?, ?> map = mapper.readValue(EMR_JOB_FLOW, Map.class);
      return map.getOrDefault("jobFlowId", "UNKNOWN_ID");
  }
}