Skip to content

prerequisites.md

Latest commit

 

History

History
156 lines (120 loc) · 5.83 KB

prerequisites.md

File metadata and controls

156 lines (120 loc) · 5.83 KB
Raw

Prerequisites

Requirements

  • RAM: 8GB runs the smaller curated scenarios; the full techvault-operational stack needs more than 20GB
  • 20GB+ disk
  • Docker Engine 20.10+ on native Linux, or Docker Desktop on macOS, Windows, or Linux
  • Docker Compose 2.0+ (docker compose version)
  • Docker Buildx (docker buildx version)
  • Python 3.11+ (for the CLI)
  • OpenSSH client. ssh-keygen must be on PATH. aptl lab start generates the lab SSH keys (the control-plane key and, for pack-backed scenarios, the scenario's ssh_key_bundle keypairs) with it, and hardens them per-platform (POSIX mode on Linux/macOS, NTFS ACLs via icacls on Windows). Preinstalled on Linux and macOS; on Windows enable the built-in OpenSSH Client optional feature (Settings → Apps → Optional features), or use Git for Windows / WSL2.
  • Node.js 18+ and npm (for the MCP servers, the AI-agent control plane that aptl lab start builds via mcp/build-all-mcps.sh; without them the lab still boots but reports degraded with MCP servers unavailable)
  • Git (only for the from-source dev install; pipx install aptl-labs needs no clone)

Install Docker

Native Linux Docker Engine:

curl -fsSL https://get.docker.com | sh
sudo usermod -aG docker $USER

Sign out and back in after changing Docker group membership.

The official Docker installer above includes Compose and Buildx. If you use Ubuntu's distribution packages instead, install all three explicitly:

sudo apt install docker.io docker-compose-v2 docker-buildx

Docker CE repositories name the last package docker-buildx-plugin instead.

macOS (Docker Desktop): Install Docker Desktop and allocate enough memory in Settings -> Resources. The full techvault-operational stack needs more than 20GB.

macOS (Colima alternative, no Docker Desktop): If you cannot use Docker Desktop (licensing, corporate policy, or preference), Colima runs the same Docker Engine in a lima VM and APTL supports it directly. The APTL host check calls out this path when Docker Buildx is missing; the full setup is:

brew install docker docker-buildx docker-compose colima
mkdir -p ~/.docker/cli-plugins
ln -sf "$(brew --prefix docker-buildx)/bin/docker-buildx" ~/.docker/cli-plugins/docker-buildx
ln -sf "$(brew --prefix docker-compose)/bin/docker-compose" ~/.docker/cli-plugins/docker-compose
colima start --cpu 4 --memory 8 --disk 60

Bump the resources for the full techvault-operational stack (see the RAM/disk requirements above). colima start also sets the active docker context to colima; verify with docker context ls.

Windows: Install Docker Desktop with the WSL2 backend enabled. Run APTL from PowerShell, Windows Terminal, Git Bash, or a WSL2 shell; keep Docker Desktop running before aptl lab start.

Linux Docker Desktop: Install Docker Desktop and use the Desktop-managed engine. It behaves like the macOS/Windows Docker VM for host sysctls.

System Config

aptl lab start enforces vm.max_map_count only when Docker is a native Linux engine. Docker Desktop on macOS, Windows, or WSL2 manages the setting inside its Linux VM, so there is no host sysctl step for those platforms.

Native Linux Docker Engine:

# Required for OpenSearch
sudo sysctl -w vm.max_map_count=262144
echo 'vm.max_map_count=262144' | sudo tee -a /etc/sysctl.conf

Check ports available:

netstat -tlnp | grep -E "(443|2027|8443|9000|9001|9200|55000)"

You do not have to free these by hand: aptl lab start probes every published host port and, if a default is already in use (Windows reserves UDP 5353 for mDNS; an editor's automatic port-forwarding may hold others), publishes that service on a free port instead and prints the real ports under "Host port remaps" in the start summary. Use the reported ports (for example, the Wazuh Dashboard URL, or dig @localhost -p <reported-port> techvault.local SOA). Pin a specific port with the matching APTL_HP_* / APTL_DNS_HOST_PORT variable to override.

Python environment

Install the CLI into a virtualenv, not the system Python. Modern Debian/Ubuntu/WSL2 hosts mark the system interpreter as externally managed and block system-wide pip under PEP 668, so pip install -e . against the system Python fails with error: externally-managed-environment.

For released installs on any OS, prefer pipx install aptl-labs.

macOS gotcha—pipx bound to the system Python 3.9. aptl-labs requires Python 3.11+ (declared in pyproject.toml). If your pipx was installed against the Command Line Tools Python (/usr/bin/python3, which is 3.9), pipx install aptl-labs fails with:

ERROR: Could not find a version that satisfies the requirement aptl-labs (from versions: none)

The real cause is the "Ignored the following versions that require a different python version" line further up in pip's output—every published aptl-labs release is filtered out by the Python-version gate. Recover with a scoped standalone Python that pipx fetches just for this venv:

pipx install --python 3.12 --fetch-missing-python aptl-labs

Alternatively, brew install [email protected] and use that interpreter.

For source installs, create a virtualenv. On Debian/Ubuntu/WSL2, install the venv module first (Debian ships it separately from python3):

sudo apt install python3-venv   # or python3-full

Then create and activate the virtualenv from the repo root on Linux/macOS:

python3 -m venv .venv && source .venv/bin/activate

On Windows PowerShell:

py -3.12 -m venv .venv
.\.venv\Scripts\Activate.ps1

.venv is gitignored. Re-run source .venv/bin/activate in each new shell on Linux/macOS or .\.venv\Scripts\Activate.ps1 in each new PowerShell before using aptl.

Verify

docker --version
docker compose version
docker buildx version
docker ps