Docker Registry: Local Image Cache

The Instructor's Perspective

In the Army, we had “supply points.” A Docker Registry is your supply point for container images. Instead of downloading the same nginx or python image from Docker Hub every time, you cache it locally. It’s faster for your lab and it reduces the load on the public infrastructure.

Why a Local Registry?

• Speed: Pulling images from your local LAN is near-instant compared to downloading from the internet.
• Bandwidth Discipline: Reduces your external bandwidth usage, especially when you’re deploying multiple stacks.
• Security: You can store your own private images securely without making them public.
• Caching (Pull-Through): Acts as a transparent proxy for Docker Hub, automatically caching requested images.

Image Supply Reliability (The PACE Plan)

Operational Discipline

P (Primary): Local Docker Registry (Docker container with a Registry-UI). A (Alternate): Direct pull from public registries (Docker Hub, GitHub Container Registry, etc.). C (Contingency): docker save and docker load to move images between nodes manually via NAS. E (Emergency): Rebuild the images from source (Dockerfile) on each node.

Standard Operating Procedure (SOP): Setting Up the Supply Point

1. Deploy: Use Docker to spin up a Registry container.
2. Configure Pull-Through Cache: Set up the registry to proxy Docker Hub.
3. Add Registry-UI: Deploy a web-based UI (like joxit/docker-registry-ui) to browse your local images.
4. Configure Clients: Update your Docker daemon configuration (daemon.json) to use your local registry as a mirror.
5. Verify: Pull an image and check the registry logs (or UI) to ensure it was cached.

Check for Understanding

• Why is it helpful to have a Registry-UI? (Hint: Think about “visibility”).
• How does a “Pull-Through Cache” save bandwidth for your entire lab?

---

Related: The Stack, index, Apt Cacher NG