Understanding Snapshots, Backups, and Mirroring
Not all data protection strategies are created equal. Here's a breakdown of three commonly used methods, what they're good for, and when to use them.
Snapshot — "The Undo Button"
A snapshot is an instant, space-efficient copy of your data at a specific point in time.
- Stored on the same disk
- Useful for quick recovery from accidental changes or deletions
- Fast and lightweight
Example Use Cases
- You deleted a file and want it back
- You messed up a configuration and want to revert
- You want to view your files "as they were yesterday"
In many self-hosted setups, snapshots are scheduled hourly, daily, or weekly and stored inside a private .snapshots/ folder.
Note: Snapshots are not safe against disk failure. If the disk dies, the snapshots go with it.
Backup — "Fire Insurance"
A backup is a secure, offline copy of your data stored on separate storage.
- Stored on a different disk or server
- Designed for disaster recovery
- Slower, but more resilient
Example Use Cases
- Ransomware wiped or encrypted your files
- The server was physically damaged or lost
- You want to restore your entire volume to a new location
Backups are often created with tools like Restic or Borg, using clean snapshots as the source so you avoid copying in-progress changes.
Backups survive total disk failure or data loss.
Mirroring — "Availability, Not Safety"
Mirroring creates an exact, real-time duplicate of your data — often across two disks or devices.
- Changes are instantly reflected in the mirror
- Common in RAID setups (e.g., RAID 1)
- Not a backup — it copies deletions and corruption too
Common Example: RAID 1
RAID 1 uses two identical hard drives. If one fails, the other keeps running without downtime.
But if:
- A file is accidentally deleted? It's gone from both.
- The system is hit by ransomware? Both disks are encrypted.
Mirroring is about uptime, not recovery.
Summary Table
| Feature | Snapshot | Backup | Mirroring |
|---|---|---|---|
| Where stored | Same disk | Separate disk/server/site | Another disk or node |
| Primary benefit | Quick rollbacks | Disaster recovery | High availability |
| Time to restore | Seconds or minutes | Minutes to hours | Immediate |
| Protection scope | Human error, corruption | Hardware loss, ransomware | Drive failure |
| Caveats | Lost if disk fails | Slower and needs monitoring | Copies deletions and corruption |
The 3-2-1 Rule of Backups
A well-known best practice for serious backup hygiene:
| Rule Component | Why It Matters |
|---|---|
| 3 copies of your data | One primary + two backups = peace of mind |
| 2 different storage types | e.g., SSD + external HDD, or disk + cloud |
| 1 offsite | Protects against fire, theft, flood, etc. |
Following this principle means keeping multiple encrypted copies on different storage types, with at least one stored offsite (or in a different region) for resilience.
Pulling It Together
Snapshots, backups, and mirrors complement one another when you use each for its intended job:
- Snapshots handle quick rollbacks from accidents or configuration mistakes.
- Backups protect against bigger disasters and let you restore to new hardware.
- Mirroring gives you availability but needs to be paired with backups so corruption or deletion doesn’t spread.
Audit the trio from time to time—make sure schedules still match your risk tolerance, that offsite copies are restorable, and that mirrors aren’t the only safety net. A little maintenance up front saves a lot of panic later.
How it works at SelfHostSam
If you're curious what this looks like in practice, SelfHostSam runs all three layers:
- Live mirroring keeps storage online even if a drive fails.
- Snapshots run on a frequent schedule so you can roll back to last hour, day, or week.
- Encrypted backups replicate data to a separate site, with cadence that depends on your plan.
Want the details tailored to your setup? Let us know via the interest form.