Docker Disk Space Management Guide for Dokploy/Ubuntu Servers

Table of Contents

• Problem Overview
• Diagnosis Steps
  • 1. Check System Disk Usage
  • 2. List Docker Artifacts
• Cleanup Procedures
  • 1. Remove Stopped Containers
  • 2. Clean Unused Images
  • 3. Clean Unused Volumes
  • 4. Additional Cleanup Options
• Prevention Strategies
  • 1. Docker Daemon Configuration
  • 2. Deployment Best Practices
• Automation Setup
  • 1. Cron Job for Regular Cleanup
  • 2. Post-Deployment Hooks
  • 3. Alternative Tools
• Monitoring
• Troubleshooting
• FAQs

---

Problem Overview

🚨 Urgent: VPS Disk Space Nearly Full – Immediate Action Required

Your VPS plan xxxxxxxxxx.hstgr.cloud has almost reached its disk limit.

⚠️ Exceeding this limit may cause your website or application to stop working.

Please take immediate action to:

• Free up disk space by deleting unnecessary files.
• Upgrade your VPS plan to a higher storage capacity.
• Monitor your server usage to prevent service interruptions.

Why ?

Disk space gradually increases with each Docker deployment due to:

• Dangling Docker Images: Untagged layers from previous builds (10-100MB+ each)
• Orphaned Containers: Stopped containers with lingering writable layers (5-50MB each)
• Unused Volumes: Persistent storage from old deployments (100MB-10GB+)
• Build Cache: Accumulated intermediate layers (500MB-5GB+)
• Unrotated Logs: Growing container log files (1MB-1GB+ per container)

---

Diagnosis Steps

Connect to your VPS account

1. Check System Disk Usage

# Overall disk usage
df -h

# Docker-specific usage
docker system df

• Sample output analysis:

TYPE            TOTAL     ACTIVE    SIZE      RECLAIMABLE
Images          15        3         8.5GB     7.2GB (84%)  # 7.2GB can be cleaned!
Containers      10        2         500MB     500MB (100%)
Local Volumes   5         2         1.2GB     1.2GB (100%)

2. List Docker Artifacts

# All images (look for <none> tags)
docker images -a

# All containers (look for "Exited" status)
docker ps -a

# Dangling volumes
docker volume ls -f dangling=true

# Check log sizes
sudo du -sh /var/lib/docker/containers/*/*-json.log | sort -h

Cleanup Procedures

1. Remove Stopped Containers

# Remove all stopped containers
docker container prune -f

# Remove specific container
docker rm [container_id]

2. Clean Unused Images

# Remove dangling images only (safe)
docker image prune -f

# Remove ALL unused images (more aggressive)
docker image prune -a -f

# Remove images older than 24h
docker image prune -a -f --filter "until=24h"

3. Clean Unused Volumes

# Remove all unused volumes
docker volume prune -f

# Remove specific volume
docker volume rm [volume_name]

4. Additional Cleanup Options

# Build cache
docker builder prune -f

# Networks
docker network prune -f

# Complete system cleanup (careful!)
docker system prune -a -f --volumes

Prevention Strategies

1. Docker Daemon Configuration /etc/docker/daemon.json:

{
  "log-driver": "json-file",
  "log-opts": {
    "max-size": "10m",
    "max-file": "3",
    "compress": "true"
  },
  "storage-driver": "overlay2",
  "storage-opts": [
    "overlay2.override_kernel_check=true"
  ]
}

Apply changes:

sudo systemctl restart docker

2. Deployment Best Practices

• Image Management:

# Use consistent tags
docker build -t myapp:latest .

# Multi-stage builds
FROM node:18 as builder
WORKDIR /app
COPY . .
RUN npm build

FROM nginx:alpine
COPY --from=builder /app/dist /usr/share/nginx/html

• Container Management:

# Auto-remove containers
docker run --rm myapp:latest

# Stop and remove old containers
docker stop old_container && docker rm old_container

Automation Setup

1. Cron Job for Regular Cleanup

# Edit crontab
crontab -e

# Add daily cleanup at 2AM
0 2 * * * /usr/bin/docker system prune -f --filter "until=24h"

2. Post-Deployment Hooks In your deployment script:

#!/bin/bash
# Deployment commands...
docker-compose up -d --build

# Cleanup
docker image prune -f
docker builder prune -f

3. Alternative Tools

• Docuum (LRU-based cleanup):

cargo install docuum
docuum --threshold 10GB

• docker-gc (Spotify's solution):

curl -L https://raw.githubusercontent.com/spotify/docker-gc/master/docker-gc > /usr/local/bin/docker-gc
chmod +x /usr/local/bin/docker-gc
docker-gc

Monitoring

# Quick status check
docker system df && df -h /var/lib/docker

# Set up alerts (example for Nagios)
command[check_docker_disk]=/usr/lib/nagios/plugins/check_disk -w 20% -c 10% -p /var/lib/docker

Troubleshooting

Problem: Cleanup doesn't free expected space Solution:

Check for mounted volumes: mount | grep docker

Verify storage driver: docker info | grep "Storage Driver"

Check for systemd mount leaks: systemctl list-units --type mount

Problem: Docker daemon won't start after config changes Solution:

sudo journalctl -u docker.service -n 50 --no-pager
sudo rm /etc/docker/daemon.json && systemctl restart docker

FAQs

Q: How often should I run cleanup?

A: For active deployment servers, run cleanup daily. For development machines, run it weekly.

Q: Will this affect running containers?

A: No, prune commands only remove unused resources. Running containers will not be affected.