Microsoft SQL Server 2012 Bible
by Adam Jorgensen, Jorge Segarra, Patrick LeBlanc, Jose Chinchilla, Aaron Nelson
How Do Database Snapshots Work?
As discussed earlier, a database snapshot is a point-in-time, read-only, static view of the source database as it existed at the time of the database snapshot creation. Understanding the mechanics and technology behind database snapshots can help you understand how it all works.
Mechanics of Snapshots
When you create a database snapshot, SQL Server runs recovery on the database snapshot and rolls back uncommitted transactions to make the database snapshot transactionally consistent. The transactions in the source database are not affected.
A database snapshot is not the same as the source database. It is a different database and has the same number of data files as the source database, but it does not have any transaction log file. When it is initially created, the data files in the snapshot database do not contain any user data and are almost empty. That is why creating a database snapshot does not take a long time.
Copy on First Write
Database snapshots use a copy-on-first-write method for each source database page updated for the first time after you create the database snapshot . For every database snapshot, SQL Server creates an in-memory bitmap. It has a bit for each data page indicating if the page is copied to the snapshot.
Every time you make an update to the source database, SQL Server checks the bitmap to see if it has been copied to the snapshot. If it is not copied, SQL Server copies the data page from the source database to the database ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access