Back to all scenarios
Scenario #319
Storage
Kubernetes v1.24, Velero + CSI Snapshots
Volume Snapshot Restore Includes Corrupt Data
Snapshot-based restore brought back corrupted state due to hot snapshot timing.
Find this helpful?
What Happened
Velero snapshot was taken during active write burst. Filesystem was inconsistent at time of snapshot.
Diagnosis Steps
- 1App logs showed corrupted files after restore.
- 2Snapshot logs showed no quiescing.
- 3Restore replayed same state.
Root Cause
No pre-freeze or app-level quiescing before snapshot.
Fix/Workaround
• Paused writes before snapshot.
• Enabled filesystem freeze hook in Velero plugin.Lessons Learned
Snapshots must be coordinated with app state.
How to Avoid
- 1Use pre/post hooks for consistent snapshotting.