Using Differential Backups

This topic is relevant for all SQL Server databases.

A differential backup is based on the most recent, previous full backup of the data that is included in the differential backup. A differential backup captures only the data that has changed since that full backup. This is known as the base of the differential. A differential backup includes only the data that have changed since the differential base. In SQL Server 2005 and later versions, differential file backups can be very fast because the SQL Server Database Engine tracks the changes that were made since the differential base was created.

Each major type of file backup can serve as the base for a series of differential backups, such as:

• Differential database backups

• Differential partial backups

• Differential file backups

As a best practice, the scope of a differential backup should be the same as the scope of its base. Therefore, a differential file backup should be based on a file backup that covers the same set of files, filegroups, or both.

Usually, differential backups do not require any special attention. However, we recommend that you take some care in the following cases:

• When taking a differential backup of a read-only database. For read-only databases, full backups that are used alone are easier to manage than when they are used with differential backups. For more information, see Backing Up Read-Only Databases.

• When intermixing full and file backups, which creates multibase differentials.

  You can create a multibase differential backup whose scope is larger than a single-base back up. However, multibase differential backups are complex to restore and generally should be avoided, except by expert users. For more information, see Working with Multibase Differential Backups.

• When taking a differential partial backup after changing the IsReadOnly property of a filegroup. For more information, see Differential Partial Backups.