feat: Add examples for PAIF automation

examples/workflows/private-ai-foundation/README.md

@@ -0,0 +1,34 @@
+# Private AI Foundation via Terraform - Samples
+
+This repository contains sample automation for enabling a Tanzu Kubernetes Cluster with NVIDIA GPUs in VMware Cloud Foundation.
+
+The configuration is divided into several steps which are intended to be executed in order.
+Each step is designed to be atomic and can be executed independently from the rest provided that its prerequisites in terms of infrastructure are in place.
+
+These examples also use the [vSphere Terraform Provider](https://github.com/hashicorp/terraform-provider-vsphere)
+
+## Workflow Overview
+
+### Starting State
+
+This configuration is intended to be applied on an environment with a configured management domain.
+
+### Desired State
+
+Using these samples you can:
+
+* Create a cluster image with vGPU drivers provided by NVIDIA
+* Create a subscribed content library with custom container images
+* Deploy a workload domain with vSAN storage and NSX network backing
+* Create an NSX Edge Cluster
+* Enable vSphere Supervisor on a cluster
+* Configure a vSphere Namespace and a Virtual Machine Class with your vGPU drivers
+
+## Contents
+
+### The [steps](https://github.com/vmware/terraform-provider-vcf/-/tree/main/examples/workflows/private-ai-foundation/steps) directory contains the sample configuration divided into a number of examples/workflows/private-ai-foundation/steps
+#### [Step 1](https://github.com/vmware/terraform-provider-vcf/-/blob/main/examples/workflows/private-ai-foundation/steps/01) - Create a vLCM cluster image with NVIDIA GPU drivers
+#### [Step 2](https://github.com/vmware/terraform-provider-vcf/-/blob/main/examples/workflows/private-ai-foundation/steps/02) - Export the cluster image and create a workload domain with it
+#### [Step 3](https://github.com/vmware/terraform-provider-vcf/-/blob/main/examples/workflows/private-ai-foundation/steps/03) - Create an NSX Edge Cluster
+#### [Step 4](https://github.com/vmware/terraform-provider-vcf/-/blob/main/examples/workflows/private-ai-foundation/steps/04) - Create a subscribed Content Library
+#### [Step 5](https://github.com/vmware/terraform-provider-vcf/-/blob/main/examples/workflows/private-ai-foundation/steps/05) - Enable Supervisor, create a vSphere Namespace and a VM class

examples/workflows/private-ai-foundation/steps/01/main.tf

@@ -0,0 +1,55 @@
+# This step generates a custom host image with vGPU drivers
+# on the vCenter for the management domain.
+# The source for the offline software depot for this step has to
+# contain the drivers.
+
+terraform {
+  required_providers {
+    vsphere = {
+      source = "hashicorp/vsphere"
+    }
+  }
+}
+
+# Connect to the vCenter Server backing the management domain
+provider "vsphere" {
+  user                 = var.vcenter_username
+  password             = var.vcenter_password
+  vsphere_server       = var.vcenter_server
+}
+
+# Read a datacenter. Can be any datacenter
+data "vsphere_datacenter" "dc" {
+  name = var.datacenter_name
+}
+
+# Retrieve the list of available host images from vLCM
+# It is also valid to base your custom image on the build of a particular host
+# but this scenario is not automated
+data "vsphere_host_base_images" "base_images" {}
+
+# Create an offline software depot
+# The source for the depot should contain the vGPU drivers 
+resource "vsphere_offline_software_depot" "depot" {
+  location = var.depot_location
+}
+
+# Create a compute cluster
+# It will remain empty and its sole purpose is to be used by vLCM to configure
+# a custom image with the GPU drivers
+resource "vsphere_compute_cluster" "image_source_cluster" {
+  name            = var.cluster_name
+  datacenter_id   = data.vsphere_datacenter.dc.id
+
+  # The "host_image" block enables vLCM on the cluster and configures a custom image with the provided settings
+  # It is recommended to add this block after you have configured your depot and retrieved the list of base images
+  # so that you can select the correct values
+  # This example uses the first available image and the first available component
+  host_image {
+    esx_version = data.vsphere_host_base_images.base_images.version.0
+    component {
+      key = vsphere_offline_software_depot.depot.component.0.key
+      version = vsphere_offline_software_depot.depot.component.0.version.0
+    }
+  }
+}

examples/workflows/private-ai-foundation/steps/01/variables.tf

@@ -0,0 +1,23 @@
+variable "vcenter_username" {
+  description = "Username used to authenticate against the vCenter Server"
+}
+
+variable "vcenter_password" {
+  description = "Password used to authenticate against the vCenter Server"
+}
+
+variable "vcenter_server" {
+  description = "FQDN or IP Address of the vCenter Server for the management domain"
+}
+
+variable "datacenter_name" {
+  description = "The name of the datacenter where the new cluster will be created"
+}
+
+variable "cluster_name" {
+  description = "The name of the compute cluster"
+}
+
+variable "depot_location" {
+  description = "The URL where the contents for the offline software depot are hosted"
+}

examples/workflows/private-ai-foundation/steps/02/main.tf

@@ -0,0 +1,193 @@
+# This step creates a new workload domain with the custom image
+# from step 1.
+
+terraform {
+  required_providers {
+    vcf = {
+      source = "vmware/vcf"
+    }
+    vsphere = {
+      source = "hashicorp/vsphere"
+    }
+  }
+}
+
+# Connect to the SDDC Manager
+provider "vcf" {
+  sddc_manager_host     = var.sddc_manager_host
+  sddc_manager_username = var.sddc_manager_username
+  sddc_manager_password = var.sddc_manager_password
+}
+
+# Connect to the vCenter Server backing the management domain
+provider "vsphere" {
+  user                 = var.vcenter_username
+  password             = var.vcenter_password
+  vsphere_server       = var.vcenter_server
+}
+
+# Request the same datacenter which you created your cluster on in step 1
+data "vsphere_datacenter" "dc" {
+  name = var.datacenter_name
+}
+
+# Request the compute cluster which you created in step 1
+data "vsphere_compute_cluster" "image_source_cluster" {
+  name            = var.source_cluster_name
+  datacenter_id   = data.vsphere_datacenter.dc.id
+}
+
+# Configure a network pool for your hosts
+resource "vcf_network_pool" "domain_pool" {
+  name = "engineering-pool"
+  network {
+    gateway = "192.168.10.1"
+    mask    = "255.255.255.0"
+    mtu     = 9000
+    subnet  = "192.168.10.0"
+    type    = "VSAN"
+    vlan_id = 100
+    ip_pools {
+      start = "192.168.10.5"
+      end   = "192.168.10.50"
+    }
+  }
+  network {
+    gateway = "192.168.11.1"
+    mask    = "255.255.255.0"
+    mtu     = 9000
+    subnet  = "192.168.11.0"
+    type    = "vMotion"
+    vlan_id = 100
+    ip_pools {
+      start = "192.168.11.5"
+      end   = "192.168.11.50"
+    }
+  }
+}
+
+# Commission 3 hosts for the new domain
+resource "vcf_host" "host1" {
+  fqdn            = var.esx_host1_fqdn
+  username        = "root"
+  password        = var.esx_host1_pass
+  network_pool_id = vcf_network_pool.domain_pool.id
+  storage_type    = "VSAN"
+}
+resource "vcf_host" "host2" {
+  fqdn            = var.esx_host2_fqdn
+  username        = "root"
+  password        = var.esx_host2_pass
+  network_pool_id = vcf_network_pool.domain_pool.id
+  storage_type    = "VSAN"
+}
+resource "vcf_host" "host3" {
+  fqdn            = var.esx_host3_fqdn
+  username        = "root"
+  password        = var.esx_host3_pass
+  network_pool_id = vcf_network_pool.domain_pool.id
+  storage_type    = "VSAN"
+}
+
+# Extract a vLCM personality (a cluster image) from the cluster you created in step 1
+# This will be applied on the new workload domain
+# It is crucial that you do this before creating the domain as it is not possible to enable vLCM afterwards
+resource "vcf_cluster_personality" "custom_image" {
+    name      = "custom-image-3"
+    cluster_id = data.vsphere_compute_cluster.image_source_cluster.id
+    domain_id = var.management_domain_id
+}
+
+# Create a workload domain
+resource "vcf_domain" "wld01" {
+  name = var.workload_domain_name
+
+  vcenter_configuration {
+    datacenter_name = "${var.workload_domain_name}-datacenter"
+    fqdn            = var.workload_vcenter_fqdn
+    gateway         = "10.0.0.250"
+    ip_address      = var.workload_vcenter_address
+    name            = "${var.workload_domain_name}-vcenter"
+    root_password   = var.vcenter_root_password
+    subnet_mask     = "255.255.252.0"
+  }
+
+  cluster {
+    name = "vi-cluster"
+
+    host {
+      id          = vcf_host.host1.id
+      license_key = var.esx_license_key
+    }
+
+    host {
+      id          = vcf_host.host2.id
+      license_key = var.esx_license_key
+    }
+
+    host {
+      id          = vcf_host.host3.id
+      license_key = var.esx_license_key
+    }
+
+    vds {
+      name = "${var.workload_domain_name}-vds01"
+
+      portgroup {
+        name           = "${var.workload_domain_name}-vds01-PortGroup-Mgmt"
+        transport_type = "MANAGEMENT"
+      }
+
+      portgroup {
+        name           = "${var.workload_domain_name}-vds01-PortGroup-vMotion"
+        transport_type = "VMOTION"
+      }
+
+      portgroup {
+        name           = "${var.workload_domain_name}-vds01-PortGroup-VSAN"
+        transport_type = "VSAN"
+      }
+    }
+
+    vsan_datastore {
+      datastore_name = "${var.workload_domain_name}-vsan"
+      license_key    = var.vsan_license_key
+    }
+
+    geneve_vlan_id = "112"
+    cluster_image_id = vcf_cluster_personality.custom_image.id
+  }
+
+  nsx_configuration {
+    license_key                = var.nsx_license_key
+    nsx_manager_admin_password = var.nsx_manager_admin_password
+
+    # You need to prepare the DNS entries for these hostnames before running Terraform
+    # You are free to modify the FQDNs
+    nsx_manager_node {
+      fqdn        = "nsx-mgmt-wld-1.vrack.vsphere.local"
+      gateway     = "10.0.0.250"
+      ip_address  = var.nsx_manager_node1_address
+      name        = "nsx-mgmt-wld-1"
+      subnet_mask = "255.255.252.0"
+    }
+
+    nsx_manager_node {
+      fqdn        = "nsx-mgmt-wld-2.vrack.vsphere.local"
+      gateway     = "10.0.0.250"
+      ip_address  = var.nsx_manager_node2_address
+      name        = "nsx-mgmt-wld-2"
+      subnet_mask = "255.255.252.0"
+    }
+
+    nsx_manager_node {
+      fqdn        = "nsx-mgmt-wld-3.vrack.vsphere.local"
+      gateway     = "10.0.0.250"
+      ip_address  = var.nsx_manager_node3_address
+      name        = "nsx-mgmt-wld-3"
+      subnet_mask = "255.255.252.0"
+    }
+    vip      = var.nsx_manager_vip_address
+    vip_fqdn = "nsx-manager-wld.vrack.vsphere.local"
+  }
+}