VERITAS Volume Manager 3.1 Administrator's Guide

Volume Manager Operations
VxSmartSync Recovery Accelerator
Chapter 3126
VxSmartSync Recovery Accelerator
The VxSmartSync™ Recovery Accelerator is available for some systems.
VxSmartSync for Mirrored Oracle© Databases is a collection of features
that speed up the resynchronization process (known as resilvering) for
volumes used in with the Oracle Universal Database™. These features
use an extended interface between Volume Manager volumes and the
database software so they can avoid unnecessary work during mirror
resynchronization. These extensions can result in an order of magnitude
improvement in volume recovery times.
Oracle automatically takes advantage of SmartSync when it is available.
The system administrator must configure the volumes correctly to use
VxSmartSync. For Volume Manager, there are two types of volumes used
by the database:
• redo log volumes contain redo logs of the database.
• data volumes are all other volumes used by the database (control
files and tablespace files).
VxSmartSync works with these two types of volumes differently, and
they must be configured correctly to take full advantage of the extended
interfaces. The only difference between the two types of volumes is that
redo log volumes should have dirty region logs, while data volumes
should not.
Data Volume Configuration
The improvement in recovery time for data volumes is achieved by
letting the database software decide which portions of the volume
require recovery. The database keeps logs of changes to the data in the
database and can determine which portions of the volume require
recovery. By reducing the amount of space that requires recovery and
allowing the database to control the recovery process, the overall
recovery time is reduced.
Also, the recovery takes place when the database software is started, not
at system startup. This reduces the overall impact of recovery when the
system reboots. Because the recovery is controlled by the database, the
recovery time for the volume is the resilvering time for the database
(that is, the time required to replay the redo logs).