Client asks "do you have backups?" — yes, daily TransIP snapshots. On D-day, restore takes 45 minutes, VM reboots without secondary disk mount, nobody knows step order. Copy existed; testable restore did not.
TransIP offers simple VPS snapshots. Ops value comes from quarterly executed restore runbook, not panel checkbox.
Snapshot vs durable backup
| Aspect | TransIP snapshot | External backup |
|---|---|---|
| Rollback speed | Fast same VPS | Variable |
| Retention | Limited | Long possible |
| Off infra | No | Yes if exported |
| Restore test | VPS clone | Isolated restore |
| Logical corruption | Reproduces bug | Older version |
Plan: frequent snapshots + periodic off-TransIP export if data critical.
Minimal restore runbook
- Create test VPS or use staging.
- Restore snapshot (or clone) — note start/end time.
- Service checklist: nginx, php-fpm, mysql, cron, firewall.
- Validate data: DB sample checksum, upload file.
- Document gaps vs prod (IP, DNS, secrets).
- Archive test log for audit.
Align with proof of restoration.
Common mistakes
Snapshots without schedule — vague RPO.
Full prod restore without prior test — DNS panic.
Forgot licences / secrets off disk snapshot.
Confuse block snapshot and DB consistency — quiesce or SQL dump too.
The crux: photo has no RTO until printed
That is the difference between having a copy and knowing how to go back.
Decide and move forward without blind spots
- Snapshot calendar + written RPO.
- Quarterly documented restore test.
- Off-infra export if compliance needs.
- Runbook accessible off TransIP panel.
- TransIP profile + compare tool.
Frequently asked questions
TransIP snapshot or backup?
Snapshot short-term rollback; durable backup complementary.
Test restore?
Test VPS clone, boot, services, documented timing.
Realistic RPO/RTO?
Snapshot interval = RPO; measured restore = RTO.
GDPR?
Technical continuity; not location/DPA substitute.
Schedule the next restore test before the next snapshot — one without the other reassures nobody.