/
Uninstall Process: Removing LHO from your AWS account

Uninstall Process: Removing LHO from your AWS account

This document provides step-by-step instructions for uninstalling Lakehouse Optimizer from your environment. Properly uninstalling the software ensures that all components are removed from your system. Before uninstalling LHO, ensure that you have administrative privileges.

 

  1. Log in to LHO and navigate to Settings > Provisioning & Permissions

  2. Unpublish all published workspaces by toggling the switch off next to each workspace name.
    This action will remove the LHO Agent from Databricks, leaving no trace behind.

    2025-08-05_09h47_11-20250805-164800.png

  3. Log in to your AWS CloudShell environment.

  4. Download the installation archive from the deployment backup S3 bucket.

    Screenshot 2025-08-05 at 18.49.20.png

    1. During deployment, an S3 bucket is created to store all deployment files.

    2. You can find this bucket in your S3 console, in the region where LHO was deployed.

    3. The bucket is named <name-prefix>-tfstate-bkp, where <name-prefix> corresponds to the name-prefix variable provided during deployment.

    4. The name-prefix can be derived from the Route 53 A-record that points to your LHO instance. For example, if the URL is https://lho.test.com, then the name-prefix is lho.

  5. If you're using AWS CloudShell for this process, upload the installation archive to your CloudShell environment.

  6. Unzip the archive in a directory of your choice within the environment.
    unzip install-files-backup.zip

  7. Run terraform init in the directory where you unzipped the archive.

    1. Run terraform destroy in the same directory to begin tearing down the deployment.
      When prompted for the Databricks service principal token, leave the field empty — it is not required for the destroy operation.

    2. This will remove all resources created during the deployment, including the EC2 instance, IAM profile, SQL database, and others.

  8. Navigate to DynamoDB and delete all tables with names starting with bplm-.

    Screenshot 2025-08-05 at 18.29.45.png

  9. Navigate to Amazon SQS and delete all queues with names containing/starting with bplm.

    Screenshot 2025-08-05 at 18.29.15.png

  10. Remove the bp-lakehouse-monitor volume from your main Unity Catalog inside your Databricks environment

    Screenshot 2025-10-15 at 12.39.56.png

  11. Remove the LHO IAM user from IAM inside your AWS console.

  12. Remove any DynamoDB tables and SQS queues starting with bplm.