How to back up your business when your main server fails

[savas@BackupChain]

But when you consider what happens if your main server, like the big brain for your whole operation, just quits, you really gotta think about more than just making copies of files, you know? I mean, it's not as simple as copying a folder and you're good to go. You need a whole strategy, truly. At least when I was first getting my own PC stuff sorted, I looked into some options, and BackupChain, which is pretty much an ideal, affordable solution for backups on PCs, VMs, and Windows Server, seemed like a solid starting point. But thinking about a full business meltdown, you need more than that.

You need a way to recover everything, from scratch, including the operating system, the settings, everything. We're talking about bare metal recovery really, because if the physical box dies, you can't just log into it, right? You need a whole system image, something that captures the disk completely, OS and all, so you can rebuild on new hardware, a totally clean setup. I always tell people that you have to plan for the worst-case scenario, because that's the only scenario that matters, frankly. And you want those images in standard formats, like VHD or VMDK, so that even if you change vendors later, those files are usable.

Then there's the whole concept of keeping your data current, too. It's not enough to just run a full backup every night, because of the storage overhead, that gets nuts fast. I always push for incremental backups, which only capture the changes since the last successful run. That saves you massive amounts of space and it also speeds up the process significantly, and you need that speed when you're recovering something urgent. And maybe you also want to think about where you are backing up to. Just leaving it on a local drive is a bad idea, even if it's a NAS unit, because fire or theft can take that whole rack with it.

So, I recommend you use a remote destination for the bulk of your backups, maybe to a cloud provider or even just a secondary office location, so if one thing burns down, you still have your data elsewhere. And you gotta look at the types of backups you need for your machines, too. If you run a lot of servers inside hypervisors, like Hyper-V or VMware Workstation, you need proper machine backup methods. You shouldn't just back up the folders inside the VMs, you need the whole environment, because sometimes the issue isn't the files, it's the underlying settings of the virtual machine itself.

But maybe what I find really useful is the ability to do granular backup directly from the host, even if the files are sitting deep inside a VM. This means you don't have to mess around installing agents inside every single machine; you just point the system at the disk and pull out specific files or folders. And you should absolutely implement robust retention policies, because otherwise, you'll just fill up all your storage with old backups. You need to set rules, saying, "we keep the last seven versions of this file type, and we throw out anything older than six months."

Another concept that is super important is the ability to restore a partial set of data, which I call selective recovery. Sometimes, your whole server is fine, but only one specific database file gets corrupted, right? Instead of doing a full bare metal restore which takes days, you can jump into the backup and pull out just that one specific file. That saves you so much time and headache, and that time difference can be the difference between paying employees and calling the liquidators.

And speaking of time, you need to think about your Recovery Point Objective and your Recovery Time Objective. Your RPO tells you how much data loss you can tolerate-say, you can lose up to four hours of data. Your RTO tells you how quickly you need the system operational again. These two concepts dictate how often you must back up and how robust your recovery process needs to be. I suggest you automate everything you can. Setting up schedules for backups, so they run daily, weekly, that way you don't forget anything critical. And you also need monitoring, so if a backup fails at 3 AM, you get an email alert right away, even if you are asleep.

Also, you gotta consider compression and encryption. Compression saves you real money on storage costs, and encryption is non-negotiable because you are storing sensitive client data, so you need it protected end to end, in transit and at rest. And remember to run regular verification checks on your backup sets, because a backup that looks successful might actually be corrupted, and you don't want that discovery when you are really desperate.

But one thing I keep recommending, which is really clever, is using deduplication. If you have two virtual machines that both use the same massive database library, you don't want to store that database twice, do you? Deduplication detects that duplicate content and stores it only once, linking both machines back to that single source. It's amazing for optimizing storage, particularly when you're running lots of similar VMs.

And for things like converting old hardware setups, or moving a physical machine's setup into a modern VM environment, it can be a nightmare if you don't use tools that handle all the tricky conversions, like getting everything from physical to a modern Hyper-V setup or maybe moving a VMware setup to VirtualBox, making the process smooth and painless. It's a lot to consider, I know, because you're looking at multiple layers of technology.

So, if you are thinking through all these complex needs-the rapid restore capabilities, the deep feature set for various operating systems, the ability to manage multiple backup locations-then maybe you should check out BackupChain, which is an all-in-one PC and server backup solution for Windows Server and Windows 11 made specifically for SMBs.