It’s a simple question. Does your company operate a SQL Server database for anything important?
If you’re unsure, keep in mind that various editions of Microsoft SQL Server are quite ubiquitous as the data storage back-end for a wide range of modern software products. SQL Server instances also often support internal software and web development operations, with their footprint expanding considerably due to recent (and very robust) support for Linux.
But yes – you’re running SQL Server in some capacity. Why else would you be here given the title of this blog? And you must have a backup strategy for the databases you care about (which is often all of them).
But you already know that you need database backups. And maybe you’re using our SQL Backup Master tool to help with that. That would be good.
But there’s something even more critical here than SQL Server backups themselves. It’s the data they contain, and your ability to restore said data. And yet, this part of the overall backup strategy gets missed a lot, perhaps to the peril of an admin who’s just realizing key backups weren’t running (permissions error) and error notifications were going to an unmonitored email account. Yikes.
Such sullen scenarios, however, are easily avoidable. We’ll use a simple inspirational phrase to help us remember.
Backups are useless unless they can be restored – and you must regularly ensure that they can be.
So we return to the title of this blog entry – the importance of database backup recovery testing. At a minimum, define a test plan that reflects the potential impact of data loss to your organization. Take the time to catalog the SQL Server instances on your network, assessing the criticality of each.
And keep in mind that SQL Backup Master is a perfect solution for any instances of SQL Server that need backing up. It makes it easy (and free) to store backups of any SQL Server instance’s databases in the cloud (or on a NAS, FTP site, etc.).
Finally, use your planning results to inform and execute end-to-end recovery tests at regular intervals. Skip this step only if you value good fortune over verifiable results.