Backups are crucial for making sure that important data does not get lost but when you have to remember to connect an external drive everytime you need a backup it becomes a chore and is liable to be forgotten. In order to force myself to take regular backups, I setup a backup server where I could store Borg Backup repos, then I used systemd-timers to automatically take backups everyday. Here are the instructions for how I set everything up for my systems. The instructions assume that you have a system to store backups but most of the information could be adjusted to work in other contexts as well, for example if you wanted to use an external drive for storage. At the end I’ll include links to the articles where I got most of this setup from.


Make sure that /etc/hosts includes a line for the IP address of your backup server.     dv-data.localdomain          dv-data

You’ll also need to ensure that the client system you are backing up has ssh access to the computer where the backups are being stored. I like this explainer from Digital Ocean on how to set up ssh keys.

Install borgbackups on both the client and server and borgmatic on the client:

sudo pacman -S borg borgmatic

If you are not using archlinux you will need a different command to install borg. Check the respective borgbackups and borgmatic documentation for installation instructions.

Next is initializing the backup repository on dv-data from the client system, dv-arch. First make sure that the folder you want to store backups in is setup, in my case I created ~/borg on the server. The following command will ask for an encryption passphrase used to access the repo, select something secure and store it somewhere safe, losing this password means losing access to the backups. The second command is for backing up the repo key, store this somewhere safe as well.

borg init --encryption=repokey-blake2 dv-data:/home/dexmexter/borg/dv-arch
borg key export dv-data:~/borg/dv-arch ./dv-arch_borg.key

Backup Script

Now that the backup repository is ready, the next step is to create the borgmatic yaml file that contains the configuration for the backups. Backups can always be triggered manually but using borgmatic makes everything a lot easier to manage.


        - /home
        - /etc

        - dexmexter@dv-data:~/borg/dv-arch

    one_file_system: true

        - '*.pyc'
        - **/.cache

    exclude_caches: true

        - .nobackup

    encryption_passphrase: "super_secure_password"
    compression: auto,zstd
    ssh_command: ssh -i /home/dexmexter/.ssh/backups_ed25519 -o ServerAliveInterval=30 -o ServerAliveCountMax=3
    relocated_repo_access_is_ok: true

    keep_daily: 7
    keep_weekly: 4
    keep_monthly: 6

        - echo "Error during prune/create/check."

For details on how to make changes to this file and what the different options mean check the borgmatic documentation.

Automating with systemd-timers

For automating the backups I chose to use systemd-timers. To make a timer work there are two parts needed, a file with a .timer which defines how often the job is run, and a file with a .service which defines what commands are going to be executed. In my case the .service file is executing the autoborg script from above. It is important to note that both the .timer and .service files should have the same prefix name and should be stored in the same location. This is what mine look like:


Description=borgmatic backup


# Lower CPU and I/O priority.



ExecStart=systemd-inhibit --who="borgmatic" --why="Prevents interrupting scheduled backup" /usr/bin/borgmatic --syslog-verbosity 1


Description=Run borgmatic backup

OnCalendar=*-*-* 20:00


This timer is set to run everyday at 8pm. If you need something different, you’ll need to look up instructions. The archlinux wiki has some great examples on this page.

Enable/Start the timer, then verify that it is loaded and active with:

sudo systemctl enable autoborg.timer
sudo systemctl start autoborg.timer
systemctl status autoborg.timer

Verify that it has been started by checking if it appears in the list of timers:

systemctl list-timers

When you want to see if the backups have been running correctly you can check the most recent log with:

systemctl status autoborg

Or view all the logs with:

sudo journalctl -u autoborg


That’s it, now you’ve got proper backups set up!

The next important thing to know is how to restore files when needed. I’m not going to explain that here, maybe in another post. The Borg Backups documentation has instructions so if you need to restore any documents that have been backed up, or if I didn’t explain something very well, it would be best to rtfm and see if you can figure it out yourself.

In the future I want to use a proper network monitoring system like Zabbix or Nagios or Icinga that would monitor the backups and notify me when something has gone wrong. In the meantime I’ve just been periodically checking the logs.