Every Shine Servers Root Server and Cloud VPS includes copy-on-write snapshots. Used well, they turn “risky” maintenance into a reversible experiment.
How copy-on-write works
A snapshot doesn’t copy your disk. It freezes the current block map; subsequent writes go to new blocks. That’s why snapshots complete in seconds regardless of disk size, and why keeping many old snapshots slowly costs write performance — the chain grows.
The pre-upgrade workflow
1. snapshot create "pre-kernel-6.9" # seconds
2. apply the risky change
3. verify (services up? app healthy?)
4a. all good → snapshot delete after 48h
4b. broken → snapshot rollback — back in minutesWhat snapshots are NOT
Snapshots live on the same host as your VM. They protect you from bad changes, not from host failure. For disaster recovery you still want scheduled off-server backups (available as an add-on on all compute plans).
Restoring
From the client portal: Server → Snapshots → Restore. The VM reboots into the frozen state. Anything written after the snapshot is discarded — take a fresh snapshot first if you need to preserve current state.