Blog Post

Setup a Database Snapshot


I have been tinkering lately with database snapshots and I thought the subject warranted a short post on the subject. This post will details a few things of note about snapshots and how to create a snapshot on your database.

Database snapshot provide a read-only point in time view of the source database. Database snapshot utilise sparse files and “copy on write” technology to take a snapshot of your database at a point in time. Database snapshots work at the page level, Before a page is changed  in the source database, the page as it exists before any change is made  is copied to the snapshot thus preserving the values on that page in the snapshot.

Because of the dependency on the source database, the snapshot must reside on the same SQL Server instance as the source database.

In terms of size, the largest a snapshot can get is the exact same size of the source database when the snapshot was taken. Thus if all the data in your database changes frequently you will need enough space to hold two copies of the source database.

Pages will only be copied to the snapshot once.

You cannot currently backup a snapshot, I think Steve Jones of SQL Server Central has posted a connect requested asking for snapshot backups and he has received a response, he has a post on the subject here

Snapshots are only available in Enterprise edition of SQL Server.

The write on copy (write before copy) procedure can be a performance overhead because of the additional IO needed to maintain the snapshot

Snapshots can be useful for reporting against the source database and also provide a way to recover the source database in certain scenarios.

OK so how do you create a database snapshot?

You need to use TSQL, you can’t create a snapshot through management studio. I have a bankingDB database on my instance of SQL Server (developer edition) I will create a snapshot of that database for the purposes of this example.

You create a snapshot using the CREATE DATABASE statement specifying AS SNAPSHOT OF clause. You also need to specify all the logical file names of the data files in the source database.

To get all the logical filenames of the data files in the source database you run this TSQL:

USE BankingDB
select * from sysfiles

When you have logical name of the database files you can form your TSQL script. This is script i used to create a snapshot of my BankingDB database:

( NAME = BankingDB, FILENAME =
'C:\DATA\SQLServer2008\' )

When you have created you snapshot you should have an entry under the database snapshots folder in SSMS


You can reference this snapshot as you would a normal database with the USE statement and you can run select statements against it too.


You rated this post out of 5. Change rating




You rated this post out of 5. Change rating