**Applicable only to GPU PaaS deployments**

# Install Baseline SKU Package

This document describes how to install the baseline SKU templates and bundled cluster images into the _default-catalog_ project of the _default_ organization for an air-gapped controller deployment.

---

## Prerequisites

- **Package URLs**  
  - URL to download the installer package for profiles/templates and its dependencies  
  - URL to download the images used by templates (only needed for airgapped controller)  
- Access to ops-console of the target Rafay Controller with Partner Admin access  
- A loader VM/host (Ubuntu, AMD64) with:  
  - **Disk**: 150 GB free space minimum

> The cluster image bundle alone is approximately **85 GB**. Additional space is required for the extracted package contents.
  
  - Sufficient CPU and memory for image extraction and upload  
  - Network access to the controller endpoint and Ops Console  
  - Python 3.x installed on the loader VM

---

## Setup

### Step 1: Create a Partner API Key

Create a partner API key from the ops-console platform before proceeding.

1. Login to your controller ops-console (e.g., `https://ops-console.<your-domain>`)

> Your user ID must have **Partner Admin** access to create a partner API key.

2. Navigate to **Users** and find your username (it should have a Partner Admin role).
3. Click the Option Menu (⋮) on the row.  
4. Select **Manage Keys** under Actions.

5. Click **New API Key** to generate the partner API key.  
6. Store the key securely for future use.

---

### Step 2: Download the Package

1. Create a working directory and navigate into it:

```
    mkdir -p ~/ncp-templates
    cd ~/ncp-templates
    ```

2. Download the SKU package and cluster image bundle:

```
    wget https://<controller-url>/Automation/ncp-templates-v3.1.38-3.tar.gz
    wget https://<controller-url>/Automation/rafay-cluster-images-3.1.38.tar.gz
    ```

3. Extract the SKU package:

```
    tar -xvzf ncp-templates-v3.1.38-3.tar.gz
    ```

4. Navigate into the extracted directory:

```
    cd ncp-templates
    ```

After extraction, the directory will contain the following files:

```
    Changelog.md
    loader-utility.py
    config.yaml
    catalog-templates/
    ```

> **Note:** Do not extract the cluster image tarball (`rafay-cluster-images-3.1.38.tar.gz`) manually. The loader utility handles this automatically when `load_images` is set to `True` in the config.

---

### Step 3: Update the Configuration

Open the `config.yaml` file and update the required values:

## environment
    hostenv: <testbed-controller-fqdn>
    opsenv: <ops-console-fqdn>

## API key
    apikey: <partner-api-key>

## Project name
    projectname: default-catalog

untar: True
    load_images: True
    images_file_name: rafay-cluster-images-3.1.38.tar.gz
    ```

#### Configuration Field Reference

| Field | Description |
| --- | --- |
| `hostenv` | Console URL of the target controller (e.g., `console.<example.com>`) |
| `opsenv` | Ops Console URL (e.g., `ops-console.<example.com>`) |
| `apikey` | Partner API key generated from the ops-console |
| `projectname` | Target project for loading templates. Default is `default-catalog` |
| `untar` | Set to `True` to allow the loader to extract the package automatically before processing |
| `load_images` | Set to `True` to load bundled cluster images to the controller nexus registry. Set to `False` for non-airgapped environments where templates will refer to the Rafay nexus registry instead |
| `images_file_name` | Filename of the downloaded cluster image tarball |

> **Important:** Ensure `images_file_name` exactly matches the downloaded image bundle filename. If it does not match, the image upload step will fail even if the template creation step succeeds.
> **Optional:** Update `projectname` if templates need to be loaded into a project other than `default-catalog`.

---

### Step 4: Run the Loader Utility

From the package directory, run:

```
    python3 loader-utility.py
    ```

The utility performs the following actions in sequence:

1. Validates template YAML files
2. Creates environment templates
3. Shares templates where required
4. Uploads the bundled cluster images
5. Loads SKU definitions into the default catalog

During the image upload step, you will see a progress bar similar to the following:

```
    [Archive] Extracted to /tmp/rafay-cluster-images
    Uploading images...
    
    [                                                  ]  3%        1/27
    [#                                                 ] 14%        4/27
    [######                                            ] 55%       15/27
    [##########                                        ] 77%       21/27
    [##############                                    ]100%       27/27
    
    [Archive] Extracting archive /home/ubuntu/demo-templates/demo-catalog-templates/demo-rafay-cluster-assets.tar.gz to /tmp/rafay-cluster-assets
    Extracting Archive: <-----> 100.00%
    [Archive] Extracted to /tmp/rafay-cluster-assets
    
    Your Rafay cluster blueprint has initialized successfully!
    ```

> **Note:** The image upload step can take **30–45 minutes** while all 27 Rafay cluster images (~85 GB) are imported. Do not interrupt the process until you see the `Your Rafay cluster blueprint has initialized successfully!` message.

---

## Verify the Upload

Monitor the command output for success messages in the following format:

```
    Validating file... catalog-templates/env-templates/<name>.yaml
    Creating EnvironmentTemplate --> <name> with sharing
    Environment Template '<name>' created successfully.
    ```

After the process completes, confirm the following:
- SKU templates are available in the **Template Catalog** from the Infra Portal
- Required environment templates are visible
- Cluster images are available for provisioning
- Profiles are accessible from **Profile Catalog** in SKU Studio

---