Your cron shows "backup OK" for 400 days. Then ransomware encrypts prod on Tuesday. You launch restore: rsnapshot did not include /var/lib/docker/volumes; Restic fails because repo password was on compromised server. Two "reliable" tools, zero data recovered.
Right question is not "which is more modern?" It is: which have you already restored, on what, in how long?
rsnapshot: readable local snapshots
Based on rsync plus hard links, daily/weekly/monthly rotation.
Strengths: simple, files readable in clear (no repo), trivial file restore over SSH, low learning curve.
Limits: poorly suited to native S3, no built-in encryption, risky snapshots on same filesystem, frequently forgotten paths.
Hosting: VPS with separate backup disk, NAS, second server.
Restic: encrypted repo, cloud-ready
Deduplicated block backup, AES encryption, S3/B2/SSH backends.
Strengths: native offsite, verified integrity (restic check), one command for partial restore.
Limits: opaque repo without Restic, slower full restore, critical key management, learning curve.
| Criterion | rsnapshot | Restic |
|---|---|---|
| File restore | Very simple local | Dedicated command |
| Offsite | Manual rsync | Native S3 |
| Encryption | External required | Built-in |
| Ransomware resilience | Weak if same host | Better if external repo + keys off prod |
What makes restores fail
Paths not included: Docker volumes, databases without coherent dump, configs outside /home. Lost keys: Restic password on dead server. Snapshots on same disk as prod — ransomware wipes all. Never practiced restore: incompatible Restic version, insufficient disk, unknown RTO.
See Restore proof and backup strategy.
Hybrid possible: local rsnapshot for fast file restore plus Restic to S3 for encrypted offsite. What matters is quarterly drill on both paths — not the tool stack displayed in README.
Record measured RTO in a document shared with leadership: without dated number, next incident repeats same panic.
The summit: tool matters less than drill
Decide and move forward without blind spots
List what must survive: files, DB dumps, configs, secrets (outside repo). Pick rsnapshot if local plus simplicity; Restic if mandatory S3 offsite. Store Restic keys off prod (vault, second admin). Run restore drill this quarter and time RTO. Hosts with object storage: directory, comparator.
Frequently asked questions
Main difference?
Restic: encrypted blocks plus native cloud. rsnapshot: simple local rsync snapshots — poorly suited offsite without manual work.
Single file restore?
rsnapshot: direct SSH copy. Restic: restic restore command — more flexible, sometimes slower.
Encryption?
Restic native. rsnapshot: add LUKS or gpg — mandatory once backups leave datacenter.
Restore test frequency?
Quarterly on one file; annual full to isolated environment — documented and timed.
Best backup is the one you already restored — not the one whose cron shows OK for six months.
