terraform-harvester-mkdocs-site

Important

This repository is still under construction!

A Terraform module for serving a mkdocs documentation website on Condenser.

It will deploy a virtual machine using the Harvester Terraform provider. The VM will be provided with a script to build and serve the website.

This template is suitable for deploying on Condenser.

Usage

1. Use this template when creating a new repo. If creating a self-contained module, name your repo according to the module naming convention of terraform-<PROVIDER>-<NAME>.
2. Change CODEOWNERS to you or your Team.

Deployment

Create a new file env.tfvars with the following contents to configure the variables for the module:

img_display_name   = "almalinux-9.3" # Display name of an image in the harvester-public namespace
prefix             = "terraform-harvester-vm"
namespace          = "my-ns" # A namespace in the cluster
public_key         = "my-key" # Your key in the namespace
network_name       = "my-ns/my-net" # A network in the namespace; this can also be left empty
baseos_repo_url    = "" # A URL for the baseos repo
appstream_repo_url = "" # A URL for the appstream repo
mkdocs_repo_url    = "" # A HTTPS URL for the documentation Git repo, e.g. https://github.com/UCL-ARC/condenser-mkdocs.git
mkdocs_repo_branch = "main" # Optional; specify a branch. Defaults to main.
network = {
  ip      = "" # Choose an IP address for the VM
  iface   = "eth0" # eth0 is a required interface
  dns     = "" # point to DNS for the network
  gateway = "" # point to Gateway for the network
}

Obtain a suitable kubeconfig file to access the Harvester cluster. Then you can deploy this module as follows:

KUBECONFIG=/path/to/kubeconfig.yaml terraform apply -var-file=env.tfvars

Wait for the deployment to complete and for the VM to finish restarting. Then log in to the VM and become the root user to run the script to build and serve the website.

sudo su -
./build_mkdocs_site.sh

The Terraform module outputs will suggest an SSH command to tunnel to the VM and serve the website to your localhost.

You can destroy the VM like so:

KUBECONFIG=/path/to/kubeconfig.yaml terraform apply -destroy -var-file=env.tfvars

Requirements

Name	Version
terraform	>= 1.8.5
cloudinit	2.3.5
harvester	0.6.4
random	3.6.3

Providers

Name	Version
cloudinit	2.3.5
harvester	0.6.4
random	3.6.3
terraform	n/a

Modules

No modules.

Resources

Name	Type
harvester_cloudinit_secret.cloud-config	resource
harvester_virtualmachine.vm	resource
random_id.secret	resource
terraform_data.vm_tags	resource
cloudinit_config.server_user_data	data source
harvester_image.img	data source
harvester_ssh_key.mysshkey	data source

Inputs

Name	Description	Type	Default	Required
appstream_repo_url	URL where the AppStream RPMs can be accessed	string	n/a	yes
baseos_repo_url	URL where the BaseOS RPMs can be accessed	string	n/a	yes
img_display_name	Display name of an image in the harvester-public namespace	string	n/a	yes
local_port	Port on your local machine where the site can be served through an SSH tunnel	number	3000	no
mkdocs_repo_branch	Name of a branch in the mkdocs repository	string	"main"	no
mkdocs_repo_url	HTTPS URL for a publicly-accessible mkdocs repository	string	n/a	yes
namespace	Namespace that the SSH public key and network are already deployed in, and that the VM will be deployed in	string	n/a	yes
network	Harvester VM network to add NICs for	object({ ip = string iface = string dns = string gateway = string })	n/a	yes
network_name	Name of a network in the specified namespace	string	n/a	yes
prefix	Prefix for the VM name	string	n/a	yes
public_key	Name of an SSH key in the specified namespace	string	n/a	yes
vm_count	n/a	number	1	no

Outputs

Name	Description
site_url	n/a
ssh_tunnel	n/a
vm_ids	n/a

---