Example Terraform Skillet

This Skillet will launch a Terraform project. All user-inputs to the ‘variables’ section will be passed to terraform as terraform variables. Therefore, the ‘variable’ names should match the terraform variable names exactly. Any terraform ‘outputs’ will be automatically captured into the context for subsequent skillets to use.

.meta-cnc.yaml

name: azure_single_pavm

label: Azure Single PAN-OS VM-Series

description: Launch a single Single PAN-OS VM-Series in Azure.

type: terraform

labels:
  terraform_image: registry.gitlab.com/panw-gse/as/terraform_tools:0.11
  collection:
    - Example Skillets

variables:
  - name: admin_username
    description: Admin Username
    default: panhandler
    type_hint: text
  - name: admin_password
    description: Admin Password
    default:
    type_hint: password
  - name: hostname
    description: Hostname
    default: panhandler-vm-01
    type_hint: text
  - name: resource_group
    description: Resource Group
    default: panhandler-unique-value-123
    type_hint: text

Terraform Variables

In this case, our variables from the skillet definition file match the variables that terraform expects. Here is a variables.tf file from this project:

variable "admin_username" {
  description = "PAN-OS NGFW Admin Username"
  default = "admin"
}

variable "admin_password" {
  description = "PAN-OS NGFW Admin Password"
  default = "admin"
}

variable "resource_group" {
  description = "Resource Group to use to build"
  default = "admin"
}

variable "hostname" {
  description = "Host name of the PA VM-Series"
  default = "pavm"
}

Any user input from Panhandler will be passed to terraform as a TFVAR.

Terraform Output Capturing

All terraform ‘outputs’ are automatically captured into the context. Here is a sample ‘outputs.tf’ file:

data "azurerm_public_ip" "pavm_public_ip_address_data" {
  name                = "${azurerm_public_ip.pavm_public_ip.name}"
  resource_group_name = "${azurerm_virtual_machine.pavm.resource_group_name}"
}

output "pavm_public_ip_address" {
  value = "${data.azurerm_public_ip.pavm_public_ip_address_data.ip_address}"
}

This will capture a variable named ‘pavm_public_ip_address’ in the Panhandler skillet context, where it can be used to pre-populate input fields in other skillets, or passed to other skillets via hidden variables, etc.

Snippet Details

The ‘snippets’ section contains all the type specific configuration. Terraform does not require a ‘snippet’ section as the skillet definition file is expected to live in the project root of the terraform project.

Terraform State Files

Terraform keeps its state in a special file on disk called the terraform.tfstate file. Panhandler by default will store the terraform state in a file on the local filesystem in the same directory as the skillet meta-data file. This allows you to destroy or refresh a previously deployed project from the Panhandler GUI.

Deploying Multiple Projects with Panhandler

By default, terraform will only deploy exactly what is proscribed in the various terraform files. That means that if you want to deploy two instances of the same project, you must ‘trick’ terraform into thinking this is a new deployment and not a modification to a previous one. Panhandler allows you to do this via the ‘Override’ option. When deploying a terraform project, if an existing terraform.tfstate file is found, Panhandler will give you the option to ‘override’ the existing state. This will cause Panhandler to backup the existing state and create a new state for this deployment.

Warning

This is a potentially dangerous operation as Terraform can create many resources in your cloud environment that are only tied together via a state file. You must be sure you can destroy all the necessary resources before you continue with the ‘override’ option.

Custom Terraform Images:

Panhandler allows the use of any docker image for Terraform projects. It is often the case that terraform depends on external binaries or libraries for various plugins. For example, the Azure provider requires the ‘az’ binary to be available in the system. To avoid deploying Panhandler with every possible combination of such binaries, you can specify a docker image to use with your terraform project. This is done via a label called: terraform_image. This label should be where the docker engine can pull the image. The entry point must be the terraform binary. This also allows any Terraform version to be supported as well.