Skip to content

index.md

Latest commit

 

History

History
49 lines (36 loc) · 2.82 KB

index.md

File metadata and controls

49 lines (36 loc) · 2.82 KB
Raw

LabLink CLI

The lablink command is a CLI-driven alternative to the lablink-template repository. It deploys the same allocator infrastructure to AWS — cloud resources are identical either way — but drives OpenTofu from your own machine instead of GitHub Actions.

!!! note "Installing" The CLI is on PyPI as lablink-cli. Install it with uv tool install lablink-cli (or pip install lablink-cli) — see Installation.

Note the package is `lablink-cli` while the command it installs is `lablink`.

Two providers

The CLI can deploy in either of two modes, selected by the provider field in your config:

provider Allocator runs on Client machines Needs AWS?
aws (default) EC2, provisioned by OpenTofu lablink client launch provisions them for you Yes
manual docker-compose, on a machine you already have you register your own boxes with lablink client register No

Everything below compares the AWS path against the template repo. If you'd rather run LabLink on hardware you already own — lab workstations, an on-prem server, a scheduler-hosted workload — see Bring-Your-Own Clients instead.

CLI vs. template repo

Both paths deploy the same allocator service and manage the same set of AWS resources. The practical difference is how much of the deployment you own and configure yourself.

Template repo CLI
What you maintain A full repo forked from lablink-template (OpenTofu .tf files, GitHub Actions workflows, configs) A single config.yaml
Customization surface OpenTofu resources, CI workflow, and config values Whatever the config.yaml schema exposes, including instance type, images, region, DNS, and SSL
Where OpenTofu runs GitHub Actions Your machine
Where state lives S3 backend in your AWS account S3 backend in your AWS account
How you trigger a deploy Push to test or run the deployment workflow lablink deploy
Secrets management GitHub repository secrets AWS credentials on your machine, passwords prompted
Who can deploy Anyone with repo access Whoever has the AWS creds locally
Best for Custom AWS resources or CI workflows Standard deployments without a repository to maintain

Both paths let you select a custom client Docker image or AMI through config.yaml. You can switch paths later because they use the same config schema for those settings.

Next steps

  1. Install the CLI
  2. Run your first deployment (AWS) — or bring your own clients (no AWS)
  3. Manage an existing deployment
  4. Full command reference: CLI Reference