dom-cli 0.1.3

Digital Ocean manager with CLI functionality

DigitalOcean Manager (DOM) CLI

dom is a command-line tool to manage your DigitalOcean resources. From GPU droplets to volumes, it provides an intuitive interface to simplify your cloud management tasks, especially for AI workloads.

---

Installation

Install dom using pip:

pip install dom-cli

---

Prerequisites

Before using the dom CLI, you need to set up your DigitalOcean API token.
Export the DIGITALOCEAN_TOKEN environment variable with your token, ensuring it has the following permissions:

• actions: read
• block_storage: read, delete, create
• block_storage_action: read, create
• droplet: read, update, delete, create
• regions: read
• sizes: read
• tag: create, read, delete
• ssh_key: read

---

Usage

The general syntax of the dom command is:

dom [OPTIONS] COMMAND [ARGS]...

Workflow

1. Installation: Install the CLI tool using pip.

2. Initialize a Project: Create a new project in the current directory using:

   dom init
   

   This command will create the following project structure:

   ├── config.yaml
   ├── cloud-configs
   │   └── base.yaml
   ├── droplets
   │   ├── cpu-mini.json
   │   └── nvidia-h100.json
   └── volumes
       └── models.json
   

config.yaml

The config.yaml file contains the main configuration for the project. Below is an example:

# DigitalOcean datacenter region slug (str).
# Full detail here: https://docs.digitalocean.com/platform/regional-availability/
digitalocean_region: nyc2

# List of (int) protected droplets IDs.
# You can't stop/delete a protected droplet.
# You can't detach a volume from a protected droplet.
protected_droplets: []

# List of (str) protected volumes IDs.
# You can't delete a protected volume.
protected_volumes: []

# JSON indentation level (int) for raw outputs.
json_indent: 4

# Action ping interval (float) in seconds.
ping_interval: 1

Project Directories

• cloud-configs/: Directory for cloud-init YAML files used to define tasks and configurations for droplets.
• droplets/: Directory for droplet templates, including configurations for CPUs and GPUs.
• volumes/: Directory for volume templates, such as those for storing AI models.

---

Examples

Droplet Management

dom handles GPU droplets and is optimized for managing automated AI tasks on DigitalOcean GPU droplets.

Create a Droplet

Example of droplet creation using the nvidia-h100 template:

dom droplet create nvidia-h100 ai-tasks-01 \
  --key [some key id] \
  --volume [some volume id] \
  --cloud-config name-of-my-cloud-init

The above command creates a droplet with all the configurations and scripts needed for running AI tasks on GPU droplets, using a cloud-init file (name-of-my-cloud-init.yaml) defined under the cloud-configs/ directory.

Volume Management

dom also supports creating and managing volumes for storing data, such as AI models.

Example: Create a Volume

Example using the models.json configuration file:

cat volumes/models.json 
{
    "name": null,
    "size_gigabytes": 100,
    "description": "Models volume",
    "region": null,
    "filesystem_type": "ext4",
    "filesystem_label": "models"
}

dom volume create models aimodels

This creates a volume named aimodels with a size of 100GB, formatted with the ext4 filesystem, and labeled as models. The volume can then be shared across multiple droplet instances.

---

License

This project is licensed under the Apache License 2.0. See the LICENSE file for details.