The 3 Types of Backup (and How to Use Them Effectively)
When we talk about backup, we often mention three types: file-based, database, and server image backups. But which kind is best?
Well, like most things, “best” depends on the situation. Each backup type has its own benefits and drawbacks, and each excels in a particular scenario. Let’s dive into what those are.
File Backup: The Precision Method
File-based backup is exactly what it sounds like – backup of an individual file or set of files. This is less finicky than it sounds, as you can also configure backup of entire directories. In Zetta, this is as simple as selecting C:\Users\JohnDoe as your backup location. Everything in John Doe’s user directory would then be captured in the backup – documents, downloads, contacts, and whatever else.
What it’s good for: If you only care about some of the data on a server, then file-based backups are ideal. For instance, the documents in your user shares definitely need backing up. That server’s instance of Windows? Not so much. You can save valuable space in backups by just protecting the data that actually matters.
Another big benefit is small-scale restores. Besides just plain hardware failure, the most common data loss scenario actually isn’t the catastrophic site failure most people envision. It’s the fat-fingers, “but I thought emptying the Recycle Bin would make the file come back!” human error scenario, usually resulting in the loss of a single crucial file or directory. The file backup approach makes it fast and easy to get those individual files back.
What it’s not so good for: If your server is scrap metal (and possibly on fire), restoring from file-based backups isn’t exactly ideal. You’d have to first reinstall and reconfigure the operating system and necessary applications before beginning to restore your actual files, which can be time consuming and occasionally fraught. Another potential downside is that being able to specifically select your backup targets means that you can FORGET to specifically select something – not a good thing to find out at restore time.
Database Backup: The Special Case
Another term for database backups is application-aware backups. Pretty much every organization that calls itself an enterprise is going to have databases like SQL and Exchange – and probably a decent number of them. While said databases obviously need to be backed up, they behave differently than run-of-the-mill files like PDFs. Because databases by nature are constantly changing, everything from retention to maintenance needs special attention.
What it’s good for: Database backups are good for backing up applications/databases (the name was probably the giveaway here). They usually include features like independent retention schedules (so you can set database retention differently than the rest of your dataset), and automated maintenance tasks like log truncation (this is important to keep your storage free from unnecessary clutter – both saving money and getting your data back to you faster).
What it’s not so good for: Surprise – if it’s not a database, then application-aware backup isn’t the way to go.
Image Backup: The Whole Enchilada
Server image backups capture everything on a server’s disk – files, programs, operating system, everything. There’s no risk of forgetting something with image backups; if it was on the server, it’ll be on the backup.
What it’s good for: Besides just the peace of mind of knowing that yes, ALL your data is being backed up, server image backups really shine in disasters. If you need to restore an entire server, the easiest way to do it is through a server image.
What it’s not so good for: The traditional problem with server image backups has been restoring individual files, especially for online backups. Backing up a whole server image frequently means having to restore a whole server image, which adds a lot of time vs. just recovering the file that’s actually needed. We actually addressed this in a recent Zetta release to allow you to browse server images in the cloud and select individual files to restore, but vendor capabilities vary and that might not be an option with all solutions.
Additionally, server images tend to be very large, and take a lot of space to store. In situations where there’s a physical restriction on backup storage this can limit retention (i.e. if you’re using a 1TB backup appliance and each server image is 200GB, you can only retain the last 5 backups before you run out of space).
D. All of the Above
So which is the best for your organization? You don’t have to pick just one – many companies use multiple types to ensure they’re ready for anything. The key is to pick whichever type will make your particular dataset easiest to restore, if and when you need it.