Add a `deploy` job to build.yml that needs `build`, so every push to main builds + pushes the images and then deploys them to racknerd2 over the mesh, pinned to the built commit's SHA, with an /api/health gate. Fully hands-off. The standalone deploy.yml stays as a manual workflow_dispatch for deploying/rolling back to an arbitrary tag without rebuilding. deploy/README.md updated to document the auto-deploy flow. Co-authored-by: Samuel James <ssamjame@amazon.com> Co-authored-by: Kiro <noreply@kiro.dev>
4.1 KiB
ArchNest — Build & Deploy (Forgejo Actions → registry → racknerd2)
This pipeline builds the Docker images in Forgejo Actions, pushes them to the Forgejo container registry, and deploys them to racknerd2 (validation host) over the NetBird mesh. racknerd2 only pulls and runs — it never builds (1.9 GiB RAM).
push to main ─► [build.yml]
job: build ─► build + push images ─► registry.snsnetlabs.com/sam/{archnest,archnest-backend}
job: deploy ─► (needs build) ssh racknerd2 ─► compose pull + up -d (this build's SHA) ─► /api/health
│
manual dispatch (any tag / rollback) ─► [deploy.yml] ssh racknerd2 ─► compose pull && up -d
Every push to main auto-builds and auto-deploys to racknerd2. deploy.yml
stays as a manual workflow_dispatch for deploying/rolling back to an arbitrary
tag without rebuilding.
Images
| Image | From | Tags |
|---|---|---|
registry.snsnetlabs.com/sam/archnest |
root Dockerfile (React build → nginx) |
latest, <commit-sha> |
registry.snsnetlabs.com/sam/archnest-backend |
backend/Dockerfile (Fastify) |
latest, <commit-sha> |
registry.snsnetlabs.com is the unproxied (DNS-only) registry host, so large
layers bypass Cloudflare's ~100 MB request-body cap. Pushed images appear at
https://forgejo.snsnetlabs.com/sam/-/packages (web UI, Cloudflare Access SSO).
One-time setup
1. Forgejo Actions secrets (repo or org settings → Actions → Secrets)
FORGEJO_REGISTRY_TOKEN— Forgejo personal access token forsamwith package scope (NOT the account password). Used bybuild.ymlto log in and push.RACKNERD2_SSH_KEY— private SSH key authorized forroot@racknerd2(mesh IP100.96.217.250). Used bydeploy.yml.
2. Runner (forgejo-runner host) — allow Docker builds
The runner runs jobs inside containers and by default has no Docker access.
Enable socket auto-mounting so the build job can build images. Create
/opt/config.yaml (or edit the existing runner config) with at least:
container:
docker_host: "automount" # mounts /var/run/docker.sock into job containers
Generate a full example with forgejo-runner generate-config > /opt/config.yaml,
set docker_host: "automount", point the service at it
(ExecStart=/usr/local/bin/forgejo-runner daemon -c /opt/config.yaml), then
systemctl daemon-reload && systemctl restart forgejo-runner.
3. racknerd2 — prepare the deploy host
Docker Engine + compose plugin are already installed. Then:
mkdir -p /opt/archnest
# copy deploy/docker-compose.yml from this repo to /opt/archnest/docker-compose.yml
# create /opt/archnest/.env from deploy/.env.example and fill in the secrets:
# ARCHNEST_JWT_SECRET = openssl rand -hex 32
# ARCHNEST_SECRET_KEY = openssl rand -hex 32
# ARCHNEST_GUAC_CRYPT_KEY = openssl rand -base64 24 | cut -c1-32
docker login registry.snsnetlabs.com # user: sam, password: the package token
Ports are bound to the mesh IP only (100.96.217.250) — Docker bypasses
ufw, so this is what keeps the app off the public interface. Validate at
http://100.96.217.250:8080.
Running it
- Automatic: push to
main→build.ymlbuilds + pushes both images, then itsdeployjob (needsbuild) pulls this commit's SHA onto racknerd2, restarts the stack, and health-checks/api/health. Fully hands-off. - Manual build: run Build & Push Images from the Actions tab (also triggers the auto-deploy job).
- Manual deploy / rollback: run Deploy to racknerd2, entering any tag
(
latestor a specific commit SHA) to deploy without rebuilding.
Notes / ceilings
- Auto-deploy targets racknerd2 (the validation host) on every push to
main, pinned to the built commit's SHA. If you later add a prod host, gate prod deploys behind a manual approval or a tag/release trigger rather than every push. - Single-arch (amd64) only — both the runner host and racknerd2 are amd64, so no buildx/multi-platform is needed.