You can't be like me right? Using Docker and have no clue why you have to do emergency disk usage clean up every so often? Well, I got tired of forgetting, so this repository includes all my hacks to maintain Docker in a way that keeps things organized and not going over board.
- `docker system df`: shows the disk usage just like df but for containers
- `docker system prune -a --volumes`: let's clean it
- `sudo systemctl restart docker`: restart the daemon
Apparently, chatty containers can create large logs. 1TB is not surprising to
see. The script truncates all container logs (*-json.log files in
/var/lib/docker/containers) that quietly eat up space. This is the main
culprit of my multi-terabyte surprise.
Can configure and limit log size and rotation. Edit or create /etc/docker/daemon.json:
{
"log-driver": "json-file",
"log-opts": {
"max-size": "10m",
"max-file": "3"
}
}
So I don’t forget the exact command. Run make clean-docker-logs and move on
with your life.
I might toss in:
- Volume audits and pruning
- Old image cleanup strategies
- Log rotation daemon
- Crontab or systemd integration
make clean-docker-logsThat's it. It runs scripts/docker_log_cleanup.sh with sudo, truncates logs,
and tells you how much space you saved.
If you're also living that "container grind" life and just want Docker to behave, clone this, set up a cron job, and stop thinking about it.
crontab -e
0 3 * * 0 /path/to/docker_log_cleanup.sh >> /var/log/docker_log_cleanup.log 2>&1