Designs that include Terraform configurations require access to a Terraform runtime environment that you integrate with the Cloud Assembly on-premises product.

How to add a Terraform runtime

The runtime environment consists of a Kubernetes cluster that runs Terraform CLI commands to perform requested operations. In addition, the runtime collects logs and returns the results from Terraform CLI commands.

The vRealize Automation on-premises product requires users to configure their own Terraform runtime Kubernetes cluster. Only one Terraform runtime per organization is supported. All Terraform deployments for that organization use the same runtime.

  1. Verify that you have a Kubernetes cluster on which to run the Terraform CLI.
    • All users can supply a kubeconfig file to run the Terraform CLI on an unmanaged Kubernetes cluster.
    • Enterprise license users have the option to run the Terraform CLI on a Kubernetes cluster managed by vRealize Automation.

      In Cloud Assembly, go to Infrastructure > Resources > Kubernetes, and verify that you have a Kubernetes cluster. See How do I work with Kubernetes in Cloud Assembly if you need to add one.

  2. If the Kubernetes cluster is newly added or modified, wait for its data collection to complete.

    Data collection retrieves the list of namespaces and other information, and might take up to 5 minutes depending on provider.

  3. After data collection completes, go to Infrastructure > Connections > Integrations > Add Integration, and select the Terraform Runtime card.
  4. Enter settings.
    Figure 1. Example Terraform runtime integration
    Example Terraform runtime integration for Advanced license users
    Setting Description
    Name Give the runtime integration a unique name.
    Description Explain what the integration is for.
    Terraform Runtime Integration:
    Runtime type (Enterprise only) Enterprise license users may select whether to run the Terraform CLI on a Kubernetes cluster managed by vRealize Automation or an unmanaged one.
    Kubernetes kubeconfig (all users)

    For an unmanaged Kubernetes cluster, paste in the entire contents of the kubeconfig file for the external cluster.

    To use an external Kubernetes runtime with a proxy server, see How to add proxy support.

    This option is available for all users.
    Kubernetes cluster (Enterprise only)

    For Kubernetes managed by vRealize Automation, select the cluster in which to run the Terraform CLI.

    The cluster and its kubeconfig file must be reachable. You can validate access to kubeconfig with a GET on /cmx/api/resources/k8s/clusters/{clusterId}/kube-config.

    This option is only available for Enterprise licenses.
    Kubernetes namespace Select the namespace to use within the cluster, for creating pods that run the Terraform CLI.
    Runtime Container Settings:
    Image Enter the path to the container image of the Terraform version that you want to run.
    Note: The VALIDATE button doesn't check for the container image.
    CPU request Enter the amount of CPU for running containers. Default is to 250 millicores.
    CPU limit Enter the maximum allowable CPU for running containers. Default is to 250 millicores.
    Memory request Enter the amount of memory for running containers. Default is 512 MB.
    Memory limit Enter the maximum allowable memory for running containers. Default is 512 MB.
  5. Click VALIDATE and adjust settings as needed.
  6. Click ADD.

Settings are cached. After adding the integration, you can modify settings such as the cluster or namespace, but it might take up to 5 minutes for a change to be detected and for the Terraform CLI to run under the new settings.

Troubleshooting the Terraform runtime

Some Terraform configuration deployment problems might be related to the runtime integration.

Problem Cause Resolution
Validation fails with an error stating that the namespace is invalid. You modified the cluster but left the previous namespace in the UI. Always reselect a namespace after modifying the cluster selection.
The namespace drop down is empty or doesn't list newly added namespaces. Data collection for the cluster has not completed. Data collection takes up to 5 minutes after entering or modifying the cluster and up to 10 minutes when entering or modifying the namespace. For a new cluster with existing namespaces, wait up to 5 minutes for data collection to complete.

For a new namespace in an existing cluster, wait up to 10 minutes for data collection to complete.

If the problem continues, remove the cluster and re-add it under Infrastructure > Resources > Kubernetes.
Terraform CLI containers are created in a previous cluster, previous namespace, or with previous runtime settings, even after the integration account was updated. The Kubernetes API client used by vRealize Automation is cached for 5 minutes. Changes might need up to 5 minutes to take effect.
Validation or a Terraform deployment operation fails with an error stating that kubeconfig is not available. Sometimes these errors occur because the cluster isn't reachable from vRealize Automation.

In other cases, user credentials, tokens, or certificates are invalid.

 The kubeconfig error can occur for a number of reasons and might require engagement with technical support for troubleshooting.

How to add proxy support

To have your external Kubernetes runtime cluster connect through a proxy server, follow these steps.

  1. Log in to your external Kubernetes cluster server.
  2. Create an empty folder.
  3. In the new folder, add the following lines to a new file named Dockerfile. 
    FROM projects.registry.vmware.com/vra/terraform:latest as final
ENV https_proxy=protocol://username:password@proxy_host:proxy_port
ENV http_proxy=protocol://username:password@proxy_host:proxy_port
ENV no_proxy=.local,.localdomain,localhost
  4. Modify the placeholder values so that the https_proxy and http_proxy environment variables include the proxy server settings that you use to access the internet.

    The protocol will be http or https according to what your proxy server uses, which might not match the environment variable name of https_proxy or http_proxy.

  5. Save and close Dockerfile.
  6. From the empty folder, run the following command. Depending on your account privileges, you might need to run the command in sudo mode.

    docker build --file Dockerfile --tag custom-terraform-runtime:1.0 .

    The command creates a local custom-terraform-runtime:1.0 Docker image.

  7. In Cloud Assembly, under Infrastructure > Connections > Integrations, go to your Terraform runtime integration.
  8. Create or edit the runtime container settings to use the custom-terraform-runtime:1.0 image:
    Terraform runtime proxy image