Reclaiming disk without losing data

Docker · lesson 25 of 31 · 3 min read

Find where the tens of gigabytes went and delete the right things, not your local databases.

Key points

docker system df splits usage into images, containers, volumes and build cache. Read it before deleting anything.
Build cache is usually the biggest number. docker builder prune --filter until=168h keeps the recent, useful part.
docker system prune removes stopped containers, unused networks, dangling images and build cache. Volumes are kept.
docker system prune -a --volumes also removes every unused image and every unused named volume - your local database.
On CI runners prune on a schedule with a time filter. A cold cache on every build costs more than the disk it saved.

Example

# where did it go?
docker system df            # summary
docker system df -v         # per image, container, volume and cache record

# safe reclaim: stopped containers, dangling images, build cache
docker system prune

# keep the last week of build cache, drop the rest
docker builder prune --filter until=168h

# only unused images older than 30 days
docker image prune -a --filter "until=720h"

# this one deletes unused named volumes. no undo, no bin.
# docker system prune -a --volumes

Measure with df, prune with a filter, and think twice before --volumes.

This is a reading copy. The full lesson — with the visual explainer, the interactive lab and a Run button for the code — lives in the Docker course, and every lesson in it is listed on the Docker contents page.