Block-Level Incremental Backups

When backing up from a frozen image, backup applications need to back up all of the files that have changed since the last backup. In environments where there are large files, this causes an overhead if only a small percentage of blocks within these files has changed. This is particularly true of database environments. By only backing up changed blocks, the backup window, the time to take a backup, can be reduced significantly. As the amount of data increases, incremental backups are becoming more critical, particularly in large enterprise environments, because the amount of time to perform a backup increases as the amount of data increases.

An incremental backup does not remove the need for a full backup. It does however avoid the need to produce a full backup on a daily basis. One scenario is to make a full backup once a week followed by an incremental backup on a daily basis. In order to fully restore a filesystem, the full backup needs to be restored, followed by the incremental backups.

VxFS storage checkpoints can be used to enable block level incremental backups. There are two different types of checkpoints that are used to enable this feature, namely:

Datafull checkpoints. The type of checkpoints described in previous sections are all datafull checkpoints. When a block is to be changed in a file in the primary fileset, the block is first copied to the file in the downstream checkpoint.

Dataless checkpoints. When a change occurs to a file ...

Get UNIX Filesystems: Evolution, Design, and Implementation now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.