Restoring a Droplet from Snapshot | Generated by AI
Question: What does “restore droplet” mean?
Answer:
“Restore droplet” means restoring a Droplet to an older point in time using a snapshot, which can be useful for restoring files that were accidentally deleted or to revert to a known-good Droplet to recover from accidental configuration changes.
Key differences:
| Operation | What happens | When to use |
|---|---|---|
| Restore Droplet | Replace the running Droplet entirely with its previous state from a backup/snapshot. The Droplet turns off during restoration and turns back on automatically. | You made a mistake and want to undo everything since the snapshot |
| Create Droplet | Spin up a new separate Droplet from a snapshot | You want a copy of an old state without losing current state |
⚠️ Critical caveat:
You currently cannot undo restoring a Droplet from a backup. Once you restore, everything after that snapshot is gone. So if you restore to a snapshot from yesterday, any work you did today is permanently lost.
For your AMD DevCloud workflow:
You’d use restore only if:
- You accidentally broke ROCm/PyTorch setup and want to go back
- You want to revert a bad config change
Normally you’d just destroy the droplet (stop paying) and create a fresh one from your snapshot later. That’s cleaner because you keep your current state intact.
References: