Cloudera Docs

Adding a customer managed encryption key for GCP

By default, a Google-managed encryption key is used to encrypt disks and Cloud SQL instances in Data Lake, FreeIPA, and Cloudera Data Hub clusters, but you can optionally configure Cloudera to use a customer-managed encryption key (CMEK) instead.

If you set a CMEK for your GCP environment, then the imported Compute Engine images will be encrypted with the CMEK instead of the default Google-managed key.

To set up a CMEK, you should:

  1. Meet the CMEK prerequisites.
  2. Register a GCP environment in Cloudera via Cloudera web UI or CDP CLI. During environment registration, specify the encryption key that you would like to use.

CMEK prerequisites

Refer to GCP Prerequisites: Customer managed encryption keys.

Create a Cloudera environment with a CMEK

You can pass the CMEK during GCP environment registration in Cloudera via Cloudera web interface or CDP CLI.

Steps

You can register your environment as described in Register a GCP environment from Cloudera UI, just make sure that on the Data Access and Audit page you enable the following:

  1. Under Customer-Managed Encryprion Keys, click Enable Customer-Managed Keys.

  2. In the same section, select the CMEK:

The following screenshot shows the UI options:

The steps below can only be performed via CDP CLI. Create an environment passing the --encryption-key parameter as shows in this example: 
cdp environments create-gcp-environment \
  --no-use-public-ip \
  --environment-name <ENVIRONMENT_NAME> \
  --credential-name <EXISTING_CREDENTIAL-NAME>\
  --region <REGION>\
  --security-access securityGroupIdForKnox=<SG_NAME1>,defaultSecurityGroupId=<SG_NAME2> \
  --public-key <PUBLIC_SSH_KEY>\
  --log-storage storageLocationBase=<LOGS_STORAGE_LOCATION> \
  --existing-network-params networkName=<NETWORK>,subnetNames=<SUBNET>,sharedProjectId=<PROJECT_ID>\
  --workload-analytics \
  --encryption-key <PATH_TO_THE_ENCRYPTION_KEY>

Next, create a Data Lake and IDBroker mappings using the usual commands. Once the environment is running, Cloudera Data Hub clusters can be created using the usual steps.