Consider the following scenario: You install a Microsoft SQL Server 2008 failover cluster in disk that contains a mounted volume. After the installation is complete, the mounted volume and the disk are added to the SQL Server cluster resource group. However, no dependencies are set between the mounted volume and the disk.
This behavior is by design.
If no dependency exists between the mounted volume and the disk, data corruption may occur on failover. To prevent this kind of corruption, you must manually create a dependency between the mounted volume and the disk. To help further automate the above process for multiple machines, visit the following Microsoft Technet Web site: Customize the 2007 Office system to work with a non-Microsoft e-mail applicationhttp://technet.microsoft.com/en-us/library/dd630735.aspxAdditionally, each mounted volume must also be added as a SQL Server dependency. If only the root physical disks dependency is added and the mounted volumes are not added, database corruption will occur on failover.