30.06.2013 Views

SQL Server Backup and Restore - Simple Talk

SQL Server Backup and Restore - Simple Talk

SQL Server Backup and Restore - Simple Talk

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Why Partial <strong>Backup</strong>s?<br />

Chapter 10: Partial <strong>Backup</strong> <strong>and</strong> <strong>Restore</strong><br />

Partial backups allow us to back up, on a regular schedule, only the read-write objects<br />

<strong>and</strong> data in our databases. Any data <strong>and</strong> objects stored in read-only filegroups will not,<br />

by default, be captured in a partial backup. From what we've discussed in Chapter 9, it<br />

should be clear that we can achieve the same effect by capturing separate file backups of<br />

each of the read-write filegroups. However, suppose we have a database that does not<br />

require support for point-in-time restores; transaction log backups are required when<br />

performing file-based backups, regardless, so this can represent an unnecessarily complex<br />

backup <strong>and</strong> restore process.<br />

With partial backups, we only need to take transaction log backups if they are needed<br />

for point-in-time recovery, <strong>and</strong> so they are ideal for use with SIMPLE recovery model<br />

databases, as a means to trim database backups down to a more manageable size, without<br />

introducing the added administrative overhead of log backups. Note that partial backups<br />

are valid for databases operating in any recovery model; it's just that they were specifically<br />

designed to simplify back up for very large, SIMPLE recovery model databases that<br />

contain a large portion of read-only data.<br />

So, in what situations might we want to adopt partial backups on a database in our<br />

infrastructure? Let's say we have a <strong>SQL</strong> <strong>Server</strong>-backed application for a community<br />

center <strong>and</strong> its public classes. The database holds student information, class information,<br />

grades, payment data <strong>and</strong> other data about the courses. At the end of each quarter,<br />

one set of courses completes, <strong>and</strong> a new set begins. Once all of the information for a<br />

current quarter's courses is entered into the system, the data is only lightly manipulated<br />

throughout the course lifetime. The instructor may update grades <strong>and</strong> attendance a few<br />

times per week, but the database is not highly active during the day. Once the current set<br />

of courses completes, at quarter's end, the information is archived, <strong>and</strong> kept for historical<br />

<strong>and</strong> auditing purposes, but will not be subject to any further changes.<br />

This sort of database is a good c<strong>and</strong>idate for partial backups. The "live" course data will<br />

be stored in a read-write filegroup. Every three months, this data can be appended to a<br />

349

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!