Automatic Backup for Home Server — restic vs borgbackup, Which is Right for You?

Some time ago, I panicked because one of the containers on the home server suddenly couldn't be started. The cause was simple: the disk was full, and I only realized it after the service went down. Fortunately, important data is already stored elsewhere. But that incident made me think: backup is not just about keeping a backup, but a small insurance that makes you sleep better at night.

In this article, I want to discuss two backup tools that come up frequently in the home server community: restic and borgbackup. Both are free, open source, support deduplication, encryption, and can be run from the terminal. But the way it works and the philosophy is different, such as choosing between a light backpack suitcase or a more structured hard suitcase.

Why Backups Need a Plan, Not Just an Intention

In the cloud era, many people think that backup is simply saving files on Google Drive, Dropbox, or similar services. In fact, a good backup follows the 3-2-1 principle: three copies of data, on two different media, one in a physically different location. Cloud services are helpful, but we also need local backups that can be accessed quickly without relying on an internet connection.

Imagine you have a filing cabinet. The cloud is a storage warehouse outside the city, while local backup is a drawer in the work desk. If you need documents suddenly, of course it's quicker to take them from the drawer. But if the house catches fire, then come to the warehouse outside the city. The two complement each other, not replace each other.

What is Deduplication, and Why is it Important

One of the reasons I like restic and borgbackup is the deduplication feature. Without deduplication, every time you run a backup, all files are copied again even though the contents are the same. Deduplation ensures that only distinct blocks of data are stored. So, daily backups will not take up much space, it's similar to writing a diary in the same book: new pages are added only if there is a new story.

This feature is very useful for home servers, where we often store many small files, container configurations, or databases that change slightly every day. In addition to deduplication, both also offer encryption, compression, and backup integrity verification.

restic: Light, Fast, and Cloud-Native

restic is written in Go and designed with a modern spirit: fast, parallel, and supports multiple storage backends. You can save backups to local directories, SFTP, Amazon S3, MinIO, Backblaze B2, Azure Blob, Google Cloud Storage, even dedicated REST servers.

Installation on Debian or Ubuntu is very easy:

sudo apt update
sudo apt install restic

Create restic repository:

export RESTIC_REPOSITORY=/mnt/backup/restic
export RESTIC_PASSWORD=password-yang-kuat-dan-unik
restic init

Running backup:

restic backup /home/adam/data /var/www --exclude-file=/home/adam/exclude.txt

Displays a list of snapshots:

restic snapshots

Recover specific files:

restic restore latest --target /tmp/restore --include /home/adam/data/file.txt

One of the advantages of Restic is that the backup process can run in parallel. If there are a lot of small files, restic is generally faster than borgbackup. Additionally, the syntax tends to be more consistent and easier to remember.

borgbackup: Solid, Structured, and Resource Efficient

borgbackup, or Borg as it is often called, is written in Python and C. Borg is older than restic, and its approach is a little more traditional. Borg stores backups in one repository with a neatly structured archive format. Deduplication is excellent, even for large files that change little.

Borgbackup installation:

sudo apt install borgbackup

Create repository:

export BORG_REPO=/mnt/backup/borg
export BORG_PASSPHRASE=password-yang-kuat-dan-unik
borg init --encryption=repokey-blake2

Running backup:

borg create ::{now} /home/adam/data /var/www --exclude-from /home/adam/exclude.txt

Displays a list of archives:

borg list

Recover files:

borg extract ::2025-07-25T10-00-00 home/adam/data/file.txt

One of Borg's unique features is borg mount, which allows us to mount archives as a read-only filesystem. This is very useful for browsing old backups without having to extract the entire file first.

Quick Comparison: restic vs borgbackup

The choice between the two often comes down to needs. If you want to backup to cloud object storage with fast synchronization, restic is a more natural choice. If you are more comfortable with local or SSH backups, and want the stability of an archive format that has been tested for years, borgbackup is still very suitable.

Backup Strategy on Home Server

On my server, I use a combined approach. Important data such as Nginx configuration, databases, and personal documents are backed up daily to an external disk using borgbackup. Once a week, the finished Borg snapshot is synced to cloud storage using rclone. This way, I get restore speed from local and remote redundancy from the cloud.

If you are just starting out, start simple. Choose the most important directory, create a local repository, then schedule the backup with cron or systemd timer. Don't forget to add exclude for files that don't need to be backed up, such as logs, cache, or temporary files.

Exclude file example for home server:

/var/log
/tmp
/home/*/.cache
/var/cache
*.log

Don't Forget to Test Restore

The most dangerous backups are backups that have never been tested. Many people feel safe just because they have a backup file, even though they have never tried to restore data. At least once a month, do a test restore of a small file from backup. This process is like a fire drill: tedious when there is no problem, but invaluable during an emergency.

For restic, the restore test can be done by extracting to a temporary directory. For borgbackup, you can install the archive and browse it like a normal folder. Make sure the recovered file can actually be opened and is not corrupt.

Conclusion

Backup is insurance that is most often ignored until it is needed. restic and borgbackup both offer powerful, secure, and efficient solutions for home servers. Choose the one that best suits your workflow and storage goals. The most important thing is not the tool, but the consistency to run it and the courage to occasionally test the restore process.

If you have experience with one of these tools, or have your own unique backup strategy, write it in the comments. I like learning from other people's stories, because usually practical ideas emerge from there that are not in the official documentation.