Snapshots
Point-in-time copies of VMs and disks for instant rollback.
A snapshot freezes the state of a VM or an individual disk at a moment in time. Creation is near-instant regardless of disk size because the platform records changes made after the snapshot rather than copying data upfront.
VM Snapshots vs Disk Snapshots
| Type | Captures | Typical use |
|---|---|---|
| VM snapshot | All disks, configuration, and optionally memory state | Before OS patches, upgrades, application releases |
| Disk snapshot | A single disk | Before schema migrations on a data disk; creating new disks from a known state |
Taking and Rolling Back
Take a snapshot
Compute → your VM → Snapshots → Take Snapshot. Include memory if you want to return to the exact running state; skip memory for a crash-consistent disk-only point.
Make your change
Patch, upgrade, or reconfigure with confidence.
Roll back if needed
Select the snapshot → Revert. The VM returns to the captured state in seconds. If all went well, delete the snapshot instead to release space.
Scheduled Snapshot Policies
Under Storage → Snapshot Policies, define automatic schedules — for example hourly snapshots kept for 24 hours plus daily snapshots kept for 7 days — and assign the policy to VMs. Retention is enforced automatically.
# Create a snapshot via API
curl -X POST "https://api.daakyicloud.com/v1/vms/vm-0a3f8b2c/snapshots" \
-H "Authorization: Bearer $TOKEN" \
-d '{"name": "pre-upgrade-2026-06-14", "include_memory": false}'Good Practice
- Keep snapshot chains short — long chains slow disk I/O; delete snapshots once a change is validated
- Snapshots share the fate of primary storage — they are rollback points, not backups
- With guest tools installed, snapshots quiesce the file system for application-consistent capture
For protection that survives storage-level incidents and deletion, use backup policies and CDP — see the Backup & Recovery section.