This seemingly odd activity of storing copies of pages before they are changed ensures that the unmodified pages are available in case the database server fails or the backup procedure needs them to provide an accurate snapshot of the database server data. These snapshots are potentially used in two activities: fast recovery and the database server backup.
After a failure, the database server uses the before-images of pages modified by non-fuzzy operations in the physical log to restore these pages on the disk to their state at the last checkpoint. Then the database server uses the logical-log records to return all data to physical and logical consistency, up to the point of the most-recently completed transaction. Checkpoints and Fast Recovery, explains this procedure in more detail.
When you perform a storage-space backup, the database server performs a full checkpoint and checks disk pages to see which should be backed up. If a backup is active, the database server sends physically-logged pages to the backup program. For more details, see the IBM Informix: Backup and Restore Guide if you use ON-Bar.
Home | [ Top of Page | Previous Page | Next Page | Contents | Index ]