Deploying Polaris on Amazon Web Services (AWS)

Build and launch Polaris using the AWS Startup Script at the location provided in the command below. This script will start an Amazon RDS for PostgreSQL instance, which will be used as the backend Postgres instance holding all Polaris data. Additionally, Polaris will be bootstrapped to use this database and Docker containers will be spun up for Spark SQL and Trino.

The requirements to run the script below are:

  • There must be at least two subnets created in the VPC and region in which your EC2 instance reside. The span of subnets MUST include at least 2 availability zones (AZs) within the same region.
  • Your EC2 instance must be enabled with IMDSv1 or IMDSv2 with 2+ hop limit.
  • The AWS identity that you will use to run this script must have the following AWS permissions:
    • “ec2:DescribeInstances”
    • “rds:CreateDBInstance”
    • “rds:DescribeDBInstances”
    • “rds:CreateDBSubnetGroup”
    • “sts:AssumeRole” on the same role as the Instance Profile role of the EC2 instance on which you are running this script. Additionally, you should ensure that the Instance Profile contains a trust policy that allows the role to trust itself to be assumed.
chmod +x getting-started/assets/cloud_providers/deploy-aws.sh
export ASSETS_PATH=$(pwd)/getting-started/assets/
export CLIENT_ID=root
export CLIENT_SECRET=s3cr3t
./getting-started/assets/cloud_providers/deploy-aws.sh

Next Steps

Congrats, you now have a running instance of1 Polaris! For details on how to use Polaris, check out the Using Polaris page.

Cleanup Instructions

To shut down the Polaris server, run the following commands:

export ASSETS_PATH=$(pwd)/getting-started/assets/
docker compose -p polaris -f getting-started/eclipselink/docker-compose.yml down

To deploy Polaris in a production setting, please review further recommendations at the Configuring Polaris for Production page.