/terraform-digitalocean-droplet

Terraform module to create a Digital Ocean Droplet using Terraform with desirable additional features.

Primary LanguageHCLApache License 2.0Apache-2.0

Terraform + Digital Ocean Droplets

Terraform module to create a Digital Ocean Droplet using Terraform with desirable additional features.

The module is a helpful convenience wrapper around the standard DigitalOcean provider using cloud-init to provide additional features:-

  • easily remount existing volumes
  • easily create an initial passwordless sudo user directly with sshkeys
  • easily add additional user_data scripts to bootstrap the Droplet

Usage

This module allows you to establish a Droplet on Digital Ocean as shown in the example below:-

module "terraform-digitalocean-droplet" {
  source  = "verbnetworks/droplet/digitalocean"

  # required variables
  # ===
  hostname = "awesomehost"
  digitalocean_region = "sgp1"

  # optional variables
  # ===
  initial_user = "ndejong"
  initial_user_sshkeys = [ 
    trimspace(file("~/.ssh/authorized_keys")) 
  ]
  
  user_data = "#cloud-config\npackages:\n - python3"
  
  # digitalocean_volume0 = "[mount-point]:[mount-device]:[volume-id]:[mount-fstype]"
  # digitalocean_volume0 = "/mnt:/dev/disk/by-id/scsi-0DO_Volume_example01:0010c05e-20ad-10e0-9007-00000c113408:ext4"
}

# optional outputs
output "hostname" { value = module.terraform-digitalocean-droplet.hostname }
output "region" { value = module.terraform-digitalocean-droplet.region }
output "ipv4_address" { value = module.terraform-digitalocean-droplet.ipv4_address }
output "volume0" { value = module.terraform-digitalocean-droplet.volume0 }

Note on Droplet startup

When a Droplet first starts it can take time before the cloudinit scripts complete. Thus, you may notice in the first few minutes when the Droplet starts that the initial_user is unable to authenticate.

Note on the digitalocean_volume0 attribute

It is optionally possible to mount a pre-existing volume to your Droplet by providing the mount details in the following format [mount-point]:[mount-device]:[volume-id]:[mount-fstype]

Obtaining the [volume-id] value requires using the Digital Ocean API, you can consider using digitalocean-api-query as a cli tool using a "volumes" argument to discover this value.

$ digitalocean-api-query volumes | jq .volumes
[
  {
    "id": "0010c05e-20ad-10e0-9007-00000c113408",
    "name": "example01",
    "created_at": "2018-03-07T02:12:46Z",
...

Note on the digitalocean_volume1 attribute

The fact that volumes are not expressed as a list is an implementation detail left for another time. The issue is root in the fact that (older) Terraform has limitations in working with lists. The hard-coded limit of two volumes is not ideal but good enough for most situations.

Versions and testing

The v0.15 verbnetworks/droplet/digitalocean module has been tested and is known to with the following

  • Terraform versions
    • v0.13.0
    • v0.14.7
  • Terraform provider digitalocean/digitalocean versions
    • v2.5.1
  • DigitalOcean images
    • ubuntu-20-10-x64, ubuntu-20-04-x64, ubuntu-18-04-x64, ubuntu-16-04-x64
    • centos-8-x64, centos-7-x64
    • fedora-33-x64, fedora-32-x64
    • debian-10-x64, debian-9-x64

The older v0.13 (and prior) verbnetworks/droplet/digitalocean versions are compatible with Terraform up to v0.12

Authors

Module managed by

License

Apache 2 Licensed. See LICENSE for full details.