README.md

EMR Serverless Java SDK Example

This example shows how to call the EMR Serverless API using the Java SDK.

In it, we use a new maven project with the latest preview jar for EMR Serverless.

Pre-requisites

• Maven 3
• Access to EMR Serverless Preview
• An Amazon S3 bucket in the us-east-1 region
• A preview version of EMR Serverless Java SDK jar
  1. Download 'java-sdk.jar`
  2. Install to your local maven repo

The steps below were tested on macOS and require curl and mvn.

# Download the java-sdk.jar and rename it
curl -o aws-java-sdk-emrserverless.jar https://elasticmapreduce.s3.amazonaws.com/emr-serverless-preview/artifacts/latest/dev/sdk/java-sdk.jar

# Install the preview jar in your local maven repo
mvn install:install-file \
    -Dfile=aws-java-sdk-emrserverless.jar \
    -DgroupId=com.amazonaws \
    -DartifactId=aws-java-sdk-emrserverless \
    -Dversion=PREVIEW \
    -Dpackaging=jar

Example Java Usage

The example below will:

• Create a new EMR Serverless Application
• Start a new Spark job with a sample SparkPi application
• Stop and delete your Application when done

It is intended as a high-level demo of how to call the EMR Serverless API from the Java SDK.

In the myapp folder.

• Ensure you install the necessary dependencies

mvn install

• Run the sample app with your own S3 bucket and IAM role

mvn exec:java -Dexec.mainClass="com.example.myapp.App -Dexec.args="--bucket <S3_BUCKET> --role-arn arn:aws:iam::123456789012:role/emr-serverless-job-role"

Once the job is running, you can also view Spark logs.

# View Spark logs
aws s3 ls s3://<S3_BUCKET>/emr-serverless/logs/applications/<application_id>/jobs/<job_run_id>/

Or copy the stdout to view the results.

aws s3 cp s3://<S3_BUCKET>/emr-serverless/logs/applications/<application_id>/jobs/<job_run_id>/SPARK_DRIVER/stdout.gz - | gunzip

If you like, you can build a jar and run it independently and modify the Spark job arguments as well.

mvn package
java -cp target/myapp-1.0-SNAPSHOT.jar com.example.myapp.App -h