Module 13 Lesson 3: System Maintenance

Docker is a "Storage Hungry" system. It keeps every image layer, every volume, and every log file forever unless you tell it otherwise. A healthy developer machine or production server needs regular maintenance.

1. The "Prune" Hierarchy

The prune command is your main weapon.

• docker container prune: Deletes all stopped containers. (Low risk).
• docker network prune: Deletes all unused networks. (Low risk).
• docker image prune: Deletes only "Dangling" images (those with no name/tag). (Medium risk).
• docker system prune: Deletes containers, networks, and dangling images.
• docker system prune -a: Deletes EVERYTHING not currently tied to a running container. (High risk—you'll have to download your images again).

---

2. Managing the Log Bloat

Every time a container prints a line of code, Docker saves it to a .json file on your hard drive. If you have a chatty app, these logs can grow to 50GB!

The Global Fix (Docker Desktop):

In your settings, find the "Log rotation" section and limit individual log files to 10MB.

The Docker Compose Fix:

services:
  app:
    # ...
    logging:
      driver: "json-file"
      options:
        max-size: "10m"
        max-file: "3"

(This keeps only the last 3 files of 10MB each, ensuring your app never uses more than 30MB for logs).

---

3. High-Performance Maintenance: df and dui

Docker provides its own disk usage tool:

docker system df

It shows you a summary of how much space is used by images, containers, and volumes, and how much is "Reclaimable" (not in use).

---

Exercise: The Big Clean

1. Run docker system df. How much "Reclaimable" space do you have?
2. Run docker system prune (without the -a).
3. Check docker system df again. How much space did you save?
4. Navigate to your newest project. Add the "Logging" block from Section 2 to your docker-compose.yml.
5. Why is it dangerous to run a production server without log rotation? (Research: "Log file disk exhaustion").

---

Summary

Regular maintenance is what separates a "Brittle" system from a "Resilient" one. By automating your prunes and setting strict limits on your log files, you ensure that your servers stay up 24/7 without needing manual intervention to clear disk space.

Next Lesson: Final checks: Docker security checklist for production.