Updating the Deployment
    • 13 Nov 2024
    • 1 Minute to read
    • Dark
      Light

    Updating the Deployment

    • Dark
      Light

    Article summary

    From time to time Acante will need updates to the jobs deployed within the customer Databricks environment. Some of these changes will require an upgrade to the deployment, such as:

    • Change in permissions assumed by the Service Principals to access additional telemetry
    • New configuration parameters that enable additional control of the deployed components
    • Bug fixes

    Material changes are infrequent and a monthly cadence of updating the deployment should be adequate. The upgrade sequence is as follows:

    1. Download the latest Notebooks from the Configurations page of the UI
    2. The update for the notebooks MUST be done via Databricks CLI to ensure a clean overwrite. Use CLI version v0.222.0 or later to ensure a clean overwrite:
    databricks workspace --profile <profile-name> import "<notebook path in Databricks workspace>/Acante Provisioning" --file "<file path on local disk>/Acante Provisioning.py" --overwrite --language PYTHON
    
    databricks workspace --profile <profile-name> import "<notebook path in Databricks workspace>/Acante Metadata discovery" --file "<file path on local disk>/Acante Metadata discovery.py" --overwrite --language PYTHON
    

    <profile-name> for the primary workspace where Acante notebooks were deployed
    <notebook path in Databricks workspace> is the location in the workspace where the notebooks were imported into.

    Reach out to Acante team if you encounter errors as it may mean the overwrite has not occurred cleanly.

    1. Unless specifically noted, no changes are required to the input variables file.
    2. Update the PAT token in the Acante Secret as follows:
    • [FOR AWS] databricks auth login --host https://accounts.cloud.databricks.com --account-id <DATABRICKS_ACCOUNT_ID>
    • [FOR AZURE] databricks auth login --host https://accounts.azuredatabricks.net --account-id <DATABRICKS_ACCOUNT_ID>
    • Using databricks CLI version v0.222.0 or higher, run:
      databricks auth token --profile ACCOUNT-<DATABRICKS_ACCOUNT_ID>

    Take the access token output from the step above and update the previously created Acante token key acante_admin_token (or other token key that you can see in the Acante Input Variables file).

    1. Log in to the primary Databricks workspace you are onboarding as Account Admin, and proceed with Steps 5 & 6 (only) as noted in the Getting Started guide