Acronis Backup for Windows Server Essentials - User Guide - Manual (Page 312)

312 Copyright © Acronis International GmbH, 2002-2014
If the selected database is in use, Acronis Backup forcibly disconnects all users from the database
and then unmounts it.
12.6 Protecting clustered SQL Server instances and AAG
SQL Server high-availability solutions
The Windows Server Failover Clustering (WSFC) functionality enables you to configure a highly
available SQL Server through redundance at the instance level (Failover Cluster Instance, FCI) or at
the database level (AlwaysOn Availability Group, AAG). You can also combine both methods.
In a Failover Cluster Instance, SQL databases are located on a shared storage. Because this storage
can only be accessed from the active node, SQL Server data is backed up only when the active node is
backed up. For the same reason, SQL databases can only be recovered onto an active node. If the
active node fails, a failover occurs and a different node becomes active.
In an availability group, each database replica resides on a different node. If the primary replica
becomes not available, a secondary replica residing on a different node is assigned the primary role.
Other solutions include database mirroring and log shipping. For more information about SQL Server
high-availability solutions refer to the Microsoft documentation:
http://msdn.microsoft.com/en-us/library/ms190202.aspx.
Backing up SQL Server configured for high availability
In both the FCI and AAG cases, backing up only one node is not sufficient. If this node fails, the SQL
Server will continue functioning, but its databases will not be backed up. If you want the SQL Server
data to be backed up uninterruptedly, regardless of how many nodes are up and operational,
consider the following approach.
1. Install Agent for SQL on all of the WSFC nodes.
2. On each of the nodes, create a backup plan with identical settings. Or create a single centralized
backup plan for all of the nodes.
The settings are as follows:
In What to back up, click Items to back up and select the check box next to the machine (not
next to individual disks). This ensures that the shared storages will be included in the backup
when the node becomes active.
In Where to back up, specify a single location for all of the nodes. This may be a centralized vault
or just a network share. This ensures that all the backed-up data will be stored in one place.
Single-pass disk and application backup - Enabled.
3. Specify other settings of the backup plan as appropriate.
With these settings, if a failover occurs, the SQL databases will continue to be backed up on a
different node. When it comes to recovery, you will find the databases by expanding the cluster in
the vault Data view or Archive view.
Recovery of databases configured for high availability
A database configured for mirroring or included in an AlwaysOn Availability Group cannot be
overwritten during a recovery because Microsoft SQL Server prohibits this. You need to remove the
target database mirroring or exclude the target database from the AAG before the recovery. Or, just
recover the database as a new non-AAG one. When the recovery is completed, you can reconstruct
the original mirroring/AAG configuration.