Installation steps for AWS EMR

This document details instructions to install Delight on AWS EMR.

It assumes that you have created an account and generated an access token on the Delight website.

There are two ways (or "launch modes") to run a EMR spark cluster:

the Step execution launch mode creates an ephemeral cluster to run operations on it,
the Cluster launch mode creates a cluster that you can connect to to run Spark applications.

We detail instructions for both cases below.

First option: Step execution launch mode

When creating a cluster in Step execution launch mode, AWS EMR lets you add a Spark Step to run a Spark application:

In the configuration window of the Spark step, add the following lines in the text box named Spark-submit options:

--packages co.datamechanics:delight_<replace-with-your-scala-version-2.11-or-2.12>:latest-SNAPSHOT
--repositories https://oss.sonatype.org/content/repositories/snapshots
--conf spark.extraListeners=co.datamechanics.delight.DelightListener
--conf spark.delight.accessToken.secret=<replace-with-your-access-token>

Don't forget to replace the placeholders! Latest Spark distributions on EMR use Scala version 2.12.

Below is a working example with emr-6.1.0 and Spark 3.0.0:

Here's the content of the text fields in the example screenshot:

Spark-submit options:

--class org.apache.spark.examples.SparkPi
--packages co.datamechanics:delight_2.12:latest-SNAPSHOT
--repositories https://oss.sonatype.org/content/repositories/snapshots
--conf spark.extraListeners=co.datamechanics.delight.DelightListener
--conf spark.delight.accessToken.secret=<replace-with-your-access-token>

Application location:

/usr/lib/spark/examples/jars/spark-examples.jar

Arguments:

Second option: Connect to the master node

Once you've created a long-lived cluster (Cluster launch mode), AWS EMR lets you connect to the master node over SSH.

From there, the spark-submit CLI will be available to you. Please follow the instructions to install Delight with the spark-submit CLI.

Please note that latest Spark distributions on EMR use Scala version 2.12.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!