Skip to content

README.md

Latest commit

 

History

History
101 lines (76 loc) · 3.29 KB

README.md

File metadata and controls

101 lines (76 loc) · 3.29 KB

Module: management/storage

Downloads OS images from public URLs and registers them in Harvester HCI, making them selectable when provisioning VMs. Run this module whenever you want to add or update images available to product teams.

When to Use

Apply this module during initial datacenter setup (after harvester-integration) and again whenever a new OS image is needed (e.g. adding Rocky Linux alongside the existing Ubuntu image).

Images are downloaded by Harvester directly from the URL — the Terraform runner does not need to download them.

Requirements

Name Version
terraform >= 1.3
harvester/harvester ~> 1.7

Prerequisites

  • Harvester cluster up and accessible via kubeconfig
  • harvester provider configured at the root level

Usage

module "storage" {
  source = "github.com/wso2-enterprise/open-cloud-datacenter//modules/management/storage?ref=v0.1.x"

  harvester_namespace = "default"

  managed_images = {
    "ubuntu-22-04" = {
      display_name = "Ubuntu 22.04 LTS (Jammy)"
      url          = "https://cloud-images.ubuntu.com/jammy/current/jammy-server-cloudimg-amd64.img"
    }
    "rocky-9" = {
      display_name = "Rocky Linux 9"
      url          = "https://dl.rockylinux.org/pub/rocky/9/images/x86_64/Rocky-9-GenericCloud.latest.x86_64.qcow2"
    }
  }
}

How Consumers Reference Images

The storage module is applied by operators in the management phase. Consumers (workload provisioning) reference the images via Terraform remote state — using the same key they were registered under:

# In the management phase (operator)
module "storage" {
  source = "..."
  managed_images = {
    "ubuntu-22-04" = { ... }
    "rocky-9"      = { ... }
  }
}

# In a workload/tenant phase (consumer) — pulls image ID from management state
data "terraform_remote_state" "management" {
  backend = "local"
  config  = { path = "../02-management/terraform.tfstate" }
}

module "tenant_vm_cluster" {
  source              = "..."
  harvester_image_name = data.terraform_remote_state.management.outputs.image_ids["ubuntu-22-04"]
}

The value of image_ids["ubuntu-22-04"] is the Harvester image resource path (namespace/name) that VM and cluster provisioning resources accept directly.

Inputs

Name Description Type Default Required
harvester_namespace Harvester namespace to register images into string "default" no
managed_images Map of image key → { display_name, url } map(object({ display_name = string, url = string })) {} no

The map key becomes the image resource name in Harvester. Use lowercase, hyphen-separated names (e.g. "ubuntu-22-04") — Harvester uses this as the Kubernetes resource name.

Outputs

Name Description
image_ids Map of image key → Harvester image ID (namespace/name). Keys match managed_images input.

Notes

  • Removing an entry from managed_images will delete that image from Harvester. Ensure no VMs are using the image before removing it.
  • Image download is async — Harvester starts pulling the image when the resource is created. Large images (>1 GB) may take several minutes to become Active.
  • The url field accepts .img, .qcow2, and .iso formats.