Schedule restic backups with systemd timers on a Linux VPS
Create a restic backup script, a oneshot systemd service and a timer unit so your Linux VPS backs itself up on a fixed schedule, with logs you can read in journalctl.
- Intermediate
- Linux VPS
- 3 min read
To schedule restic backups with systemd timers you need three files: a shell script that runs restic, a oneshot service that calls the script, and a timer that starts the service on a calendar. This guide builds all three on a Linux VPS, then shows how to confirm the schedule and read the output of each run.
Systemd timers are a good fit for backups because they log to the journal, survive reboots, can catch up on missed runs, and never start a second copy of a job that is still running.
Before you start
You need:
- A Linux VPS running Ubuntu 22.04/24.04, Debian 12, or AlmaLinux/Rocky 8–9, with root or
sudoaccess. - restic installed. On Debian and Ubuntu use
sudo apt install restic; on AlmaLinux and Rocky enable EPEL and runsudo dnf install restic. Check the restic documentation for the current version and release notes. - A backup target. This guide uses a repository on a mounted disk at
/srv/backup/restic-repo. You can pointRESTIC_REPOSITORYat an SFTP host, S3-compatible bucket or any other backend restic supports — the timer setup is identical. - A rough idea of what to back up.
/etc,/homeand your web roots are the usual starting points.
The restic repository password is the only key to your backups. If you lose it, the data is unreadable. Store a copy somewhere outside the VPS before you go further.
Install restic and create the repository:
sudo install -d -m 700 /etc/restic
sudo sh -c 'openssl rand -base64 32 > /etc/restic/password'
sudo chmod 600 /etc/restic/password
sudo mkdir -p /srv/backup/restic-repo
sudo RESTIC_PASSWORD_FILE=/etc/restic/password restic init --repo /srv/backup/restic-repoYou should see created restic repository and a repository ID. Copy that ID into your notes.
Write the restic backup script
Keep the logic in a script so the service unit stays short and you can run the same job by hand when debugging.
Create /usr/local/bin/restic-backup.sh:
#!/usr/bin/env bash
set -euo pipefail
export RESTIC_REPOSITORY="/srv/backup/restic-repo"
export RESTIC_PASSWORD_FILE="/etc/restic/password"
export RESTIC_CACHE_DIR="/var/cache/restic"
restic backup /etc /home /var/www \
--exclude=/home/*/.cache \
--exclude=/var/www/*/node_modules \
--exclude-caches \
--tag scheduled
restic forget \
--keep-daily 7 --keep-weekly 4 --keep-monthly 6 \
--tag scheduled --pruneMake it executable and test it once:
sudo chmod 750 /usr/local/bin/restic-backup.sh
sudo /usr/local/bin/restic-backup.shThe first run uploads everything and can take a while. Later runs only send changed data, so they are much faster.
--exclude-caches skips directories that contain a CACHEDIR.TAG file, which covers most build and package caches without you listing them by hand.
Create the oneshot service
The service describes what to run. It is Type=oneshot because the script does its work and exits.
Create /etc/systemd/system/restic-backup.service:
[Unit]
Description=Restic backup to /srv/backup/restic-repo
Wants=network-online.target
After=network-online.target
[Service]
Type=oneshot
User=root
ExecStart=/usr/local/bin/restic-backup.sh
Nice=10
IOSchedulingClass=idle
TimeoutStartSec=6hNice and IOSchedulingClass keep the backup from competing with your live services. TimeoutStartSec gives a large first backup room to finish.
Write the systemd timer
The timer describes when to run the service. Create /etc/systemd/system/restic-backup.timer:
[Unit]
Description=Run the restic backup daily at 02:30
[Timer]
OnCalendar=*-*-* 02:30:00
RandomizedDelaySec=15m
Persistent=true
Unit=restic-backup.service
[Install]
WantedBy=timers.targetThe three settings that matter:
OnCalendaris the schedule.*-*-* 02:30:00means every day at 02:30 server time. UseMon *-*-* 03:00:00for weekly runs.RandomizedDelaySecspreads load if you run several servers against one repository.Persistent=trueruns the job at the next boot if the VPS was down when the timer was due.
Check that systemd parses your calendar expression the way you expect:
systemd-analyze calendar "*-*-* 02:30:00"Enable the timer
Reload systemd, then enable and start the timer:
sudo systemctl daemon-reload
sudo systemctl enable --now restic-backup.timerConfirm the
Was this guide helpful?
Still stuck?
Ask RunRacks AI for a quick answer, or open a ticket and our team will take a look at your server.
Related guides
Docker & DevOps
Self-host Jellyfin with Docker and hardware transcoding
Install Jellyfin in Docker on a Linux VPS, mount your media library, publish it over HTTPS and switch on hardware transcoding with VAAPI or Intel Quick Sync.
- Intermediate
- Linux VPS
- 6 min read
Docker & DevOps
Deploy a Self-Hosted App with Docker Compose from Scratch
Write a compose.yaml, keep your secrets in .env, start the stack with docker compose up -d, then verify it and update it later without losing data.
- Beginner
- Linux VPS
- 7 min read
Docker & DevOps
Manage secrets and environment variables in Docker Compose
Keep API keys and database passwords out of your compose file using .env files, environment keys and Docker secrets, then confirm containers read them correctly.
- Intermediate
- Linux VPS
- 7 min read