GitHub Runners¶
Self-hosted GitHub Actions runners for CI/CD workflows with GPU acceleration support.
Overview¶
| Attribute | Value |
|---|---|
| Image | ghcr.io/atrawog/bazzite-ai-pod-githubrunner:stable |
| Size | ~5GB |
| GPU | Yes (for GPU-accelerated workflows) |
| Port | None (outbound connection to GitHub) |
Quick Start¶
| Step | Command | Description | Recording |
|---|---|---|---|
| 1 | ujust runners config | Configure runner | |
| 2 | ujust runners start | Start runner | |
| 3 | ujust runners status | Check status |
Prerequisites¶
Configure with your GitHub repository and runner token:
You'll need a runner registration token from your repository's Settings > Actions > Runners.
Lifecycle Commands¶
| Command | Description | Recording |
|---|---|---|
ujust runners config | Configure settings | |
ujust runners start | Start runner | |
ujust runners status | Check status | |
ujust runners logs | View logs | |
ujust runners stop | Stop runner | |
ujust runners delete | Remove config |
Multiple Instances¶
Run multiple runners for parallel job execution:
# First runner
ujust runners config --repo-url=https://github.com/username/repo
ujust runners start
# Second runner (same repo)
ujust runners config -n 2 --repo-url=https://github.com/username/repo
ujust runners start -n 2
# Third runner (different repo)
ujust runners config -n 3 --repo-url=https://github.com/username/other-repo
ujust runners start -n 3
GPU Support¶
Runners include GPU access for workflows that require it:
- ML model training workflows
- CUDA/PyTorch testing
- GPU-accelerated builds
GPU is automatically detected and available to workflow steps.
Pre-installed Tools¶
The runner image includes:
- Python, Node.js, Go, Rust
- Docker (for container workflows)
- Common build tools (gcc, make, cmake)
- ML libraries (PyTorch, Transformers)
See Also¶
- Runners Command Reference - All commands and flags
- Runners Recordings - Watch command demos
- GPU Setup - GPU configuration