Docker System Maintenance

1# See overall Docker disk usage
2docker system df
3
4# Detailed breakdown
5docker system df -v
6
7# Example output:
8# TYPE            TOTAL     ACTIVE    SIZE      RECLAIMABLE
9# Images          45        12        8.5GB     5.2GB (61%)
10# Containers      15        8         2.1GB     1.8GB (85%)
11# Local Volumes   23        8         15GB      12GB (80%)
12# Build Cache     150       0         3.2GB     3.2GB
13
14# The RECLAIMABLE column shows what can be cleaned up

1# Remove unused containers, networks, images, and cache
2docker system prune
3
4# Also remove unused volumes (DANGEROUS - data loss possible!)
5docker system prune --volumes
6
7# Remove ALL unused images, not just dangling ones
8docker system prune -a
9
10# Individual prune commands (more control):
11docker container prune    # Remove stopped containers
12docker image prune        # Remove dangling images
13docker image prune -a     # Remove all unused images
14docker volume prune       # Remove unused volumes (CAREFUL!)
15docker network prune      # Remove unused networks
16docker builder prune      # Remove build cache
17
18# Prune with filters
19docker image prune -a --filter "until=168h"  # Older than 1 week
20docker container prune --filter "until=24h"  # Older than 1 day

1# Create cleanup script
2cat > /usr/local/bin/docker-cleanup.sh << 'EOF'
3#!/bin/bash
4set -e
5
6echo "Docker cleanup starting at $(date)"
7
8# Remove containers stopped more than 24 hours ago
9docker container prune -f --filter "until=24h"
10
11# Remove images not used for 7 days
12docker image prune -a -f --filter "until=168h"
13
14# Remove build cache older than 7 days
15docker builder prune -f --filter "until=168h"
16
17# Remove unused networks
18docker network prune -f
19
20# DO NOT auto-prune volumes - too risky
21
22echo "Cleanup complete at $(date)"
23docker system df
24EOF
25
26chmod +x /usr/local/bin/docker-cleanup.sh
27
28# Add to cron (runs daily at 3 AM)
29echo "0 3 * * * /usr/local/bin/docker-cleanup.sh >> /var/log/docker-cleanup.log 2>&1" | crontab -

1# Check log sizes
2sudo du -sh /var/lib/docker/containers/*/*-json.log | sort -rh | head -10
3
4# Set daemon-wide log defaults
5# /etc/docker/daemon.json
6{
7  "log-driver": "json-file",
8  "log-opts": {
9    "max-size": "10m",
10    "max-file": "3"
11  }
12}
13
14# Restart Docker to apply
15sudo systemctl restart docker
16
17# Note: Existing containers keep their old log config
18# You need to recreate containers for new defaults
19
20# Emergency: Truncate a specific log (use with caution)
21sudo truncate -s 0 /var/lib/docker/containers/<container-id>/*-json.log

1# List all volumes
2docker volume ls
3
4# Find orphaned/unused volumes
5docker volume ls -f dangling=true
6
7# Inspect a volume
8docker volume inspect volume_name
9
10# Find what's using a volume
11docker ps -a --filter volume=volume_name
12
13# Remove specific volumes (after verification!)
14docker volume rm volume_name
15
16# NEVER run this without verification
17# docker volume prune  # This deletes ALL unused volumes!
18
19# Backup a volume before removing
20docker run --rm -v volume_name:/data -v $(pwd):/backup alpine tar czf /backup/volume_backup.tar.gz -C /data .

1# Stop Docker
2sudo systemctl stop docker
3
4# Move existing data
5sudo mv /var/lib/docker /new/path/docker
6
7# Create symlink (simple method)
8sudo ln -s /new/path/docker /var/lib/docker
9
10# OR configure in daemon.json (cleaner)
11# /etc/docker/daemon.json
12{
13  "data-root": "/new/path/docker"
14}
15
16# Start Docker
17sudo systemctl start docker
18
19# Verify
20docker info | grep "Docker Root Dir"