Backup and Restore Operations - Microsoft SQL Server 2008 Analysis Services - page 791

Now that you know the relational database and table names, you can back them up. You
can create a full backup of the FoodMart relational database or make sure that content of
WriteTable_Sale s is stored along with the Analysis Server backup.
Backup Strategies
Let's look at some common backup scenarios—most, but not all, of which will use either
the Backup Database dialog box or a DDL file.
Typical Backup Scenario
In a typical backup scenario, you build a system of scheduled backups. You can use SQL
Server Integration Services to schedule them to run automatically without human interac-
tion.
NOTE
SQL Server Integration Services enables you to execute and schedule any Analysis
Services command.
It's a good idea to schedule backup operations immediately after your database is incre-
mentally updated so that you always have the most recent version available for a restore
operation, if one is needed.
High-Availability System Backup Scenario
When Analysis Server is part of a highly available application, there is, unavoidably, a
stretch of time that users will find the system unavailable or will experience slowdowns
while the files are being backed up. In this situation, even though the backup functional-
ity operation is designed to be minimally intrusive, you can't avoid a little intrusiveness.
Therefore, you need to do what you can to your backup operation to make the perfor-
mance hit as small as possible.
Standalone Backup Server
One backup strategy for highly available systems is to maintain a standalone backup
server on a separate computer. You set up a procedure to synchronize the state of the
backup server with the state of the main production server. You can use the Synchronize
command to set up this procedure. You can find detailed information about the synchro-
nize functionality in Chapter 40, “Deployment Strategies.”
One advantage of this strategy is that it results in a lighter load on the production server
computer. After it synchronizes the state of the production computer with the state of the
backup server, the Synchronize command is very efficient at determining which portions
of the database have changed. It then sends to the backup server only the files that differ
from the database on the main server.