Automation

How to Automate Verified Linux Backups with Bash and systemd Timers

Published October 27, 2026 · 7 min read

On this page

Prerequisites
Step-by-Step Guide
Conclusion

Backups that depend on someone remembering to run them eventually fail. This tutorial creates a nightly compressed backup with a checksum, automatic deletion of old archives and a scheduled run using a systemd timer.

Prerequisites

Ubuntu 24.04 with sudo access
A directory to protect (this guide uses /srv/data)
Free space in /var/backups/data, and a plan to copy archives off the machine

Step-by-Step Guide

Step 1: Create the backup script

Save this as backup.sh in your current directory.

#!/usr/bin/env bash
set -euo pipefail

SRC="/srv/data"
DEST="/var/backups/data"
KEEP_DAYS=14
STAMP="$(date +%F_%H%M%S)"
ARCHIVE="backup-$STAMP.tar.gz"

mkdir -p "$DEST"
tar -czf "$DEST/$ARCHIVE" -C "$SRC" .
( cd "$DEST" && sha256sum "$ARCHIVE" > "$ARCHIVE.sha256" )
find "$DEST" -type f -name 'backup-*' -mtime +"$KEEP_DAYS" -delete

echo "Backup complete: $ARCHIVE"

Step 2: Install and test it

sudo mkdir -p /srv/data
sudo install -m 750 -o root -g root backup.sh /usr/local/bin/backup.sh
sudo /usr/local/bin/backup.sh

Step 3: Create the systemd service

Create /etc/systemd/system/backup.service:

[Unit]
Description=Daily directory backup

[Service]
Type=oneshot
ExecStart=/usr/local/bin/backup.sh

Step 4: Create the timer

Create /etc/systemd/system/backup.timer:

[Unit]
Description=Run the directory backup daily

[Timer]
OnCalendar=*-*-* 03:00:00
Persistent=true

[Install]
WantedBy=timers.target

sudo systemctl daemon-reload
sudo systemctl enable --now backup.timer
systemctl list-timers backup.timer

Persistent=true runs a missed backup after the machine was powered off.

Step 5: Test a restore

cd /var/backups/data
sudo sha256sum -c backup-*.tar.gz.sha256
mkdir -p /tmp/restore-test
sudo tar -xzf "$(ls -1 backup-*.tar.gz | tail -n 1)" -C /tmp/restore-test
ls /tmp/restore-test

Conclusion

You now have scheduled, checksummed backups with 14-day retention. Keep a second copy off-site, check journalctl -u backup.service after changes, repeat the restore test every few months and never keep the only copy on the same disk as the original data.