January 30, 2006 at 9:43 am
Hi,
I'm in the process of documenting various procedures for the admin of some of our SQL Servers. I would like your thoughts on best practices on TESTING backups, such as how often. Should the backups be restored to a seperate testing server for testing? Should all databases be tested for each server, or is testing just 1 indicative of the rest? etc.
The Scenario:
Many SQL Servers
Each SQL Server has many databases
All Tran logs are backed up as regularly as every 10 minutes.
Thanks
Drew
January 30, 2006 at 11:16 am
Backup how often:
Depends on requirement (how much of the data may be gone, how critical is the database)
Depends on the time limit to restore a database. If your timewindow is smaller you may want to make more differential backups so you don't have to apply all transactionlogs)
Restore:
I would want to restore all database to have an estimate on the time required, a verification the backup isn't corrupt and as a practice.
Restoration time may depend heavily on available hardware.
January 30, 2006 at 11:51 am
On top of what Jo is saying, not only will verifying your backups tell you if you have corrupt backup files, it will also help identify media problems.
January 31, 2006 at 7:20 am
Along the same vein as testing backups is "Disaster Recovery" planning. Therefore, I think it is best to test the procedures for restoring an entire server (including all of its databases). If your server doesn't go through many changes then you can probably do the DR test once in a while. However, changes to your server can alter your DR plan and it is important to keep the plan up-to-date.
Hope that helps.
Norene Malaney
Viewing 4 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply