AWS EMR setup in Apache Spark - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to specify the EMR cluster's release label.

Apache Spark

emr_cluster = client.run_job_flow(ReleaseLabel='[1]')

Drag options to blanks, or click blank then click option'

Aemr-5.30.0

Bemr-6.7.0

Cemr-3.11.0

Demr-4.0.0

Attempts:

3 left

💡 Hint

Common Mistakes

Using an outdated EMR release label that lacks Spark support.

Misspelling the release label string.

The release label emr-6.7.0 specifies a recent stable EMR version with Apache Spark support.

2fill in blank

medium

Complete the code to set the instance type for the master node in the EMR cluster.

Apache Spark

Instances={'MasterInstanceType': '[1]', 'InstanceCount': 3}

Drag options to blanks, or click blank then click option'

Am5.xlarge

Bt2.micro

Cc4.large

Dr3.2xlarge

Attempts:

3 left

💡 Hint

Common Mistakes

Choosing an instance type too small for the master node.

Confusing master and core node instance types.

The m5.xlarge instance type is a balanced choice for the master node in EMR clusters.

3fill in blank

hard

Fix the error in the bootstrap action configuration to install a custom package.

Apache Spark

BootstrapActions=[{'Name': 'InstallPackages', 'ScriptBootstrapAction': {'Path': '[1]'}}]

Drag options to blanks, or click blank then click option'

Ahttp://mybucket/scripts/install.sh

Bgs://mybucket/scripts/install.sh

C/local/scripts/install.sh

Ds3://mybucket/scripts/install.sh

Attempts:

3 left

💡 Hint

Common Mistakes

Using local file paths instead of S3 URLs.

Using incorrect cloud storage prefixes like gs:// (Google Cloud Storage).

Bootstrap scripts must be stored in an S3 bucket accessible by EMR, so the path must start with s3://.

4fill in blank

hard

Fill both blanks to configure the EMR cluster to use Spot instances for core nodes with a bid price.

Apache Spark

Instances={'CoreInstanceType': '[1]', 'CoreInstanceMarket': '[2]', 'CoreBidPrice': '0.10', 'InstanceCount': 5}

Drag options to blanks, or click blank then click option'

Am5.xlarge

BON_DEMAND

CSPOT

Dc5.large

Attempts:

3 left

💡 Hint

Common Mistakes

Setting CoreInstanceMarket to 'ON_DEMAND' when Spot is intended.

Choosing an instance type not supported for Spot.

Core nodes use m5.xlarge instances and the market type SPOT enables Spot instances for cost savings.

5fill in blank

hard

Fill all three blanks to define an EMR step that runs a Spark application stored in S3.

Apache Spark

Steps=[{'Name': 'SparkApp', 'ActionOnFailure': '[1]', 'HadoopJarStep': {'Jar': '[2]', 'Args': ['spark-submit', '[3]']}}]

Drag options to blanks, or click blank then click option'

ACONTINUE

Bcommand-runner.jar

Cs3://mybucket/apps/my_spark_app.py

DTERMINATE_CLUSTER

Attempts:

3 left

💡 Hint

Common Mistakes

Using TERMINATE_CLUSTER which stops the cluster on failure.

Incorrect Jar name or missing Spark app path.

The step continues on failure (CONTINUE), uses command-runner.jar to run commands, and runs the Spark app located at s3://mybucket/apps/my_spark_app.py.