VPS Full with Docker

1) Typical Symptoms

Docker no longer starts

apt fails

sort, journalctl, systemctl fail

Errors such as:

No space left on device
cannot create temporary file in /tmp
Cannot connect to the Docker daemon

2) Quick diagnosis: confirm that the disk is full

Check total disk space


df -h

👀

Watch for:

/ at 95-100%

/var very full

3) If `/tmp` is full: avoid commands that use `sort`

When the disk is too full, commands such as:


du ... | sort ...

may fail because sort uses /tmp.

Example of an error:


sort: cannot create temporary file in '/tmp': No space left on device

4) Identify the largest folders (without scanning the entire system)

(Cleaning up Docker first can help)

Scan only `/var` (often the culprit)


sudo du -h --max-depth=1 /var 2>/dev/null

Then drill down into the largest folder

Example if /var/lib is huge:


sudo du -h --max-depth=1 /var/lib 2>/dev/null

Example if /var/log is huge:


sudo du -h --max-depth=1 /var/log 2>/dev/null

Simple sort (if there’s enough space)


sudo du -h --max-depth=1 /var 2>/dev/null | sort -hr

Top 20


sudo du -h --max-depth=1 /var 2>/dev/null | sort -hr | head -n 20

5) Common issue: systemd-journald logs are too large

If you see a large directory such as:


/var/log/journal/<id_machine>/

Check the log size


sudo du -sh /var/log/journal/*

Clean it up, keeping only 100MB


sudo journalctl --vacuum-size=100M

Or keep only the logs from the last 24 hours:


sudo journalctl --vacuum-time=1d

6) Common issue: huge Docker logs (containers/*.log)

Check the size


sudo du -sh /var/lib/docker/containers

Clear the logs without deleting the containers


sudo find /var/lib/docker/containers/ -name "*.log" -exec truncate -s 0 {} \;

It’s safe: it clears the logs but keeps the files.

7) Restart Docker

After freeing up space:


sudo systemctl restart docker

If it fails:


sudo systemctl status docker
journalctl -xeu docker.service

8) Docker diagnostics: understanding what’s taking up space

Once Docker is running:


docker system df

Example output:

Huge images

A lot of reclaimable space

9) Docker cleanup (with verification)

View stopped containers


docker ps -a -f "status=exited"

View dangling images


docker images -f "dangling=true"

View unused volumes


docker volume ls -f "dangling=true"

10) Docker cleanup (actions)

Delete stopped containers


docker container prune -f

Delete unused images (dangling only)


docker image prune -f

Delete all unused images (frees up a lot of space)

⚠️ May force re-pulls later.


docker image prune -a -f

Delete unused volumes


docker volume prune -f

Complete (aggressive) cleanup

⚠️ Deletes unused images/volumes.


docker system prune -a -f --volumes

11) Verify that everything is OK after cleanup

Check disk space


df -h

Check Docker


docker system df
docker ps

Prevention (recommended)

A) Limit systemd-journald logs

Edit:


sudo nano /etc/systemd/journald.conf

Add or modify:


SystemMaxUse=1G

Then apply:


sudo systemctl restart systemd-journald

B) Automate a weekly Docker cleanup

Create a script


sudo nano /usr/local/bin/docker-prune.sh

Content:


#!/bin/bash
docker image prune -a -f
docker container prune -f
docker volume prune -f
docker builder prune -f

Make executable:


sudo chmod +x /usr/local/bin/docker-prune.sh

Add a cron job (Sunday at 3 a.m.)


sudo crontab -e

Add:


0 3 * * 0 /usr/local/bin/docker-prune.sh

C) Limit the size of Docker logs (recommended)

Configure Docker to limit log size.

Create / edit:


sudo nano /etc/docker/daemon.json