How to deploy Anbox Cloud with Terraform¶

To deploy Anbox Cloud using Terraform, you need the Juju provider for it.

Note

The terraform plan is a work in progress and makes use of terraform-provider-juju which is also in active development. There may be breaking changes in the future for the terraform plan.

Prerequisites¶

A machine running a supported Ubuntu version.
Your Ubuntu Pro token for an Ubuntu Pro subscription. If you don’t have one yet, speak to your Canonical representative. If you already have a valid Ubuntu Pro token, log in to Ubuntu Pro website to retrieve it.

Caution

The Ubuntu Pro (Infra-only) token does NOT work and will result in a failed deployment. You need an Ubuntu Pro subscription.
A Juju Controller bootstrapped and connected to your Juju Client (CLI). See Add a controller and model for information on how to setup a Juju Controller.

Install Terraform¶

Terraform is an infrastructure as code (IaC) tool that helps you build, change, and version infrastructure safely and efficiently.

To install Terraform, enter the following command:

sudo snap install --classic --channel=latest/stable terraform

Initialize Terraform¶

Clone the reference terraform plan and initialize terraform:

git clone https://github.com/canonical/anbox-cloud-terraform
cd anbox-cloud-terraform
terraform init

The required providers for the terraform plan with their correct versions, including the Juju provider for terraform, are downloaded.

Deploy Anbox Cloud¶

Note

The current reference terraform plan maps a basic Anbox Cloud deployment with a limited amount of configuration options. If you want to customize the deployment based on your current infrastructure requirements, remember to modify the cloned plan as required.

Create a deploy.tfvars file and provide necessary configuration values to the terraform plan. Your deploy.tfvars file could look like this:

ubuntu_pro_token = "<your_ubuntu_pro_token>"
constraints      = ["arch=amd64"]
anbox_channel    = "1.26/stable" // Channel to use for deploying Anbox Cloud
subclusters = [
  {
    name           = "a" // Suffix for the Juju model designated to an Anbox Subcluster
    lxd_node_count = 1  // Number of LXD nodes to deploy in the Anbox Subcluster
  }
]

To preview what will be deployed by the terraform plan, run the following command:

terraform plan -out=tfplan -var-file=deploy.tfvars

If everything looks fine, apply the plan:

terraform apply tfplan -parallelism=1

You should use the terraform option -parallelism=1 to increase the reliability for the plan to successfully apply.

Note

Throughout this guide, you will learn about the different configuration parameters you can set in the deploy.tfvars file for customizing your deployment. Every time you modify this file, remember to preview the changes and apply the plan.

Now, if you run juju models, you will see two Juju models created, namely anbox-controller and anbox-subcluster-a.

Model                Cloud/Region  Type  Status     Machines  Units  Access  Last connection
anbox-controller     juju/default  lxd   available         4      4  admin   1 minute ago
anbox-subcluster-a*  juju/default  lxd   available         6      6  admin   1 minute ago

After deployment is finished, Juju will continue to install software and connect the different parts of the cluster together. This can take several minutes. You can monitor the status by running the following commands:

juju status -m anbox-controller --color
juju status -m anbox-subcluster-a --color

Scale the number of Anbox Subclusters¶

One of the advantages of using Terraform is the ability to manage infrastructure with a persistent state. Terraform remembers and synchronizes the current state of the infrastructure with the expectations of the plan. Terraform generates a diff of the current state and the expected state in the defined plan when you run terraform plan. Use this to scale your Anbox Cloud deployments.

To add another subcluster, modify the deploy.tfvars file:

ubuntu_pro_token = "<your_ubuntu_pro_token>"
constraints      = ["arch=amd64"]
anbox_channel    = "1.26/stable" // Channel to use for deploying Anbox Cloud
subclusters = [
  {
    name           = "a" // Suffix for the Juju model designated to an Anbox Subcluster
    lxd_node_count = 1  // Number of LXD nodes to deploy in the Anbox Subcluster
  },
  { // Add a new subcluster
    name           = "b"
    lxd_node_count = 1
  }
]

Review and apply the plan.

Similarly, to scale down your deployment, remove the subcluster from the configuration and apply the plan.

Scale the number of LXD nodes per Anbox Subcluster¶

Similar to the scaling operation for subclusters, we can also scale the number of LXD nodes per subcluster. To add or remove LXD nodes, modify the value of the lxd_node_count variable.

Enable HA Configuration¶

The terraform plan comes with a recommended high availability (HA) configuration for Anbox Cloud deployments. Enable it by setting enable_ha to true.

Deploy AAR (Anbox Application Registry)¶

To deploy the AAR alongside Anbox Cloud, set the registry_config and deploy_registry parameters:

ubuntu_pro_token = "<your_ubuntu_pro_token>"
constraints      = ["arch=amd64"]
anbox_channel    = "1.26/stable" // Channel to use for deploying Anbox Cloud
subclusters = [
  {
    name           = "a"
    lxd_node_count = 5
    registry_config = { // registry configuration for the subcluster
        mode = "client"
    }
  }
]
deploy_registry = true

This deploys a new Juju model dedicated to the Anbox Application Registry.

Enable observability using COS (Canonical Observability Stack)¶

Set the enable_cos configuration to true to deploy the required resources/charms to integrate Anbox Cloud components with COS.

Note

The plan only deploys the grafana-agent charm internally in the models and creates the required relations with other anbox charms. The plan does NOT create Juju relations for integrating with a COS model deployed externally. If you want to use an external COS model, follow the instructions as described in the COS tutorial to create those relations manually.