Defragmenting a hard drive that has shadow copies enabled on it can result in loss of shadow copy information. That means users who have files stored on the server may not be able to recover previous versions when they need to. The reason this problem occurs is that when defragmentation is underway, the Volume Shadow Copy Service thinks that new files are being written to the volume and tries to create shadow copies of them as defragmentation occurs. The result is corrupted shadow copy information.
There are two ways you can work around this problem:
- You can format the NTFS volume using 16kB or larger clusters. Since the Volume Shadow Copy service works with file units this size, it understands how to deal with them appropriately. Unfortunately this can mean wasting lots of space if your file server is used to save lots of small files.
- Create a separate volume and store the shadow copy information for all your existing volumes on this new volume, and never defragment this new volume. Having a dedicated shadow copy volume like this is a good workaround to the problem, and if users only occasionally need to restore old versions of files then even if this new volume is never defragmented it should perform its function reasonable well, especially if the new volume is on its own separate physical drive.