Audience: System Administrators
Content Summary: This simple deployment guide familiarizes users with Immuta on EMR. This guide is only meant to be deploy clusters for non-production purposes, such as demos or proof-of-concept. For more robust deployments, please see the main installation guide for Immuta on EMR.
Deprecation notice
Support for this integration has been deprecated.
AWS CLI (v1.16.x
or greater) installed in a bash environment.
The CLI should be configured to use a role that is able to fully manage EMR, IAM, and S3 resources. This can be a user role in a local environment or an instance role on an EC2 instance.
Resource IDs for your chosen AWS VPC subnet and EMR-managed security groups.
Be sure that your master and worker security groups are configured for bi-directional communication with your Immuta instance.
An instance of Immuta that is reachable from your chosen AWS VPC.
A username and password for the Immuta archives site. If you need these, reach out to your Immuta support professional.
First, download the quickstart script.
Next, run the script. Note that you will be prompted for input variables. If a variable is not required, you can press enter to use the displayed default value.
See below for an example of the script being run and prompting for variables. Note that any input in the example is simply for demonstration purposes; you will need to provide your own values.
The immuta-emr-quickstart.sh
script will prompt the user for input variables to configure the AWS resources required for the cluster. These variables are represented by the environment variables listed below. Exporting these environment variables prior to running the script will skip the prompts.
CLUSTER_NAME
Optional. The name of the EMR cluster to be created.
Default: immuta-quickstart
.
EMR_VERSION
Optional. The EMR version of the cluster. Current supported versions are 5.17.0
- 5.23.0
.
Default: 5.23.0
.
IMMUTA_VERSION
Optional. The full Immuta version to be installed on the cluster.
Default: 2023.4.12_20240507.
IMMUTA_INSTANCE_URL
Required. The URL of the Immuta instance that will drive policies on the cluster.
AWS_REGION
Optional. The AWS Region that the cluster will run in.
Default: us-east-1
.
INSTANCE_COUNT
Optional. The number of instances (master + worker) in the cluster.
Default: 3
.
INSTANCE_TYPE
Optional. The type of instance for cluster nodes.
Default: m5.xlarge
.
AWS_KEY_NAME
Required. The name of the SSH keypair in AWS that will be used to connect to the cluster.
AWS_SUBNET_ID
Required. The ID string of the subnet that the cluster will run in.
SERVICE_SECURITY_GROUP
Required. The ID string of the security group for the cluster's EMR services.
MASTER_SECURITY_GROUP
Required. The ID string of the security group for the cluster's master node.
WORKER_SECURITY_GROUP
Required. The ID string of the security group for the cluster's worker nodes.
ARCHIVE_USERNAME
Required. The username for Immuta Archives.
ARCHIVE_PASSWORD
Required. The password for Immuta Archives.
BOOTSTRAP_BUCKET
Optional. The S3 bucket where bootstrap artifacts will be stored. If the specified bucket does not exist, a new one will be created with default private ACLs.
Default: immuta-emr-bootstrap-$AWS_ACCOUNT_ID-$AWS_REGION
.
DATA_BUCKET
Optional. The S3 bucket where partitioned data is stored. If the specified bucket does not exist, a new one will be created with default private ACLs.
Default: immuta-emr-data-$AWS_ACCOUNT_ID-$AWS_REGION
.
KADMIN_PASSWORD
Optional. The Kerberos admin password that will be used to create Kerberos principals on the cluster's dedicated internal KDC.
Default: random.
HDFS_SYSTEM_TOKEN
Optional. The HDFS System Token that the cluster will use to securely communicate with the Immuta instance. You should generate this value in the Immuta Configuration UI before creating your cluster.
Default: random.
You will need to copy the immuta.keytab
and krb5.conf
files from the cluster and upload them to your Immuta instance using the Immuta Configuration UI.
The quickstart bootstrap automatically seeds the cluster with three user principals for you to use while familiarizing yourself with the Immuta platform and data policies: owner
, consumer1
, and consumer2
. The default Kerberos password for these users is immuta-quickstart
.
You can associate these users with your Immuta users by following this guide. Note that only the owner
principal will have access to the data in your chosen S3 data bucket, so this is the principal that you should use to create your data sources in Immuta.