Teadmistebaas

Keeping every backup ever taken forever isn't practical — disk space is finite, and most old backups have little ongoing value. This guide covers implementing a sensible, automated retention policy.

Why Retention Policy Matters

• Unlimited retention eventually fills any disk, regardless of size
• Most incidents are discovered within days, not months — very old backups rarely get used
• A structured policy balances recovery flexibility against storage cost

A Common Retention Pattern: Grandfather-Father-Son (GFS)

This gives fine-grained recent recovery options while keeping longer-term history at a manageable storage cost.

Simple Retention with find (Time-Based Only)

find /var/backups -type f -mtime +7 -delete

Simple, but doesn't implement tiered retention — just deletes everything older than N days uniformly.

Implementing GFS Retention with a Script

sudo nano /usr/local/bin/backup-with-rotation.sh

#!/bin/bash
BACKUP_DIR="/var/backups"
DAY_OF_WEEK=$(date +%u)
DAY_OF_MONTH=$(date +%d)
TIMESTAMP=$(date +%F)

# Always take the daily backup
mysqldump -u root --all-databases | gzip > "$BACKUP_DIR/daily/backup-$TIMESTAMP.sql.gz"

# Copy to weekly on Sundays
if [ "$DAY_OF_WEEK" -eq 7 ]; then
    cp "$BACKUP_DIR/daily/backup-$TIMESTAMP.sql.gz" "$BACKUP_DIR/weekly/"
fi

# Copy to monthly on the 1st
if [ "$DAY_OF_MONTH" -eq 01 ]; then
    cp "$BACKUP_DIR/daily/backup-$TIMESTAMP.sql.gz" "$BACKUP_DIR/monthly/"
fi

# Retention cleanup
find "$BACKUP_DIR/daily" -type f -mtime +7 -delete
find "$BACKUP_DIR/weekly" -type f -mtime +28 -delete
find "$BACKUP_DIR/monthly" -type f -mtime +365 -delete

Retention for Object Storage (S3-Compatible)

Rather than a script, use your provider's native lifecycle policy feature — configure rules directly in the bucket settings to automatically transition or delete objects based on age, avoiding the need for scripted cleanup on the remote end entirely.

Monitoring Backup Storage Usage

du -sh /var/backups/*

Track this over time to confirm your retention policy is actually keeping storage growth bounded as expected.

Choosing Retention Periods for Your Situation

Common Mistakes

• No retention policy at all, letting backups accumulate until disk is full
• Retention so aggressive that a needed older backup is already deleted when you discover you need it
• Cleanup script silently failing, with no monitoring to catch it

Verifying Retention Is Actually Working

ls -la /var/backups/daily/ | wc -l

Periodically confirm the count of retained backups roughly matches your intended policy — catching a broken cleanup script before it fills the disk.

Related Articles

• Backup Strategy 101: The 3-2-1 Rule Explained
• How to Back Up to Object Storage (S3-Compatible)
• How to Check and Manage Disk Usage on a Linux VPS