 | ||
| Additional information about COW snapshots | ||
|---|---|---|
| Prev | Copy on Write (COW) | Next |
Since COW snapshots rely on a Shadow Storage area, COW snapshots have the following constraints:
COW snapshots are developed by keeping track of the blocks that have changed. If the original volume that the COW snapshot is based upon becomes unavailable or corrupted, then the COW snapshot cannot be used to restore the snapshotted volume.
After restoring the original volume with a COW snapshot, any subsequent COW snapshots that were taken after that snapshot are automatically deleted by Windows.
The size of the allocated shadow storage area limits the number of COW snapshots for a volume.When the shadow storage area has reached its capacity, then older snapshots are deleted to make room for new snapshots. In addition, 512 COW snapshots is the maximum number allowed per volume.
During a restore operation, a forced dismount of the original volume or the volume containing the shadow storage area may result in I/O corruption and the loss of COW snapshots.
Due to the incremental nature of COW snapshots, COW snapshots can only be used to restore the snapshotted volume to a point in time.
For every write operation on a new block of the snapshotted volume, a read operation is performed to allow the COW snapshot to store the block in the shadow storage area. As a result, a COW snapshot may affect the I/O performance of the snapshotted volume.
SFW only supports COW snapshots of dynamic volumes.
Both the shadow storage area and the snapshotted volume must reside in the same dynamic disk group.
If a volume containing a shadow storage area is a part of a snapshot set and the shadow storage area volume is restored, then all snapshots that were stored on that volume are deleted automatically by Windows. Make sure that volumes containing shadow storage areas for other volumes are not part of the same snapshot set. Otherwise COW snapshots may automatically get deleted during a restore operation, resulting in a snapshot not found error.
If a VMDg resource for a disk group on which the volume is created is present in a Failover Cluster, then before initiating the restore operation you must take the resource offline. If you fail to do so, the restore operation fails. After the snapshot is restored, you must bring the resource online.
After performing operations for COW snapshots or operations on the Shadow Storage area, the VEA GUI may not be updated to reflect the results of the operation. Use the VSS Refresh command to update the VEA GUI.