03-16-2021, 07:10 PM
When you talk about where you should put these images, I always tell people you absolutely cannot rely on just one spot, because if that one spot goes down, then you are completely cooked, like, done for. You really need to think about the three-two-one rule, which is something you really need to internalize, you know? You should have three copies of your data, two different types of media, and one of those copies must be offsite, which is super critical. And when we talk about "offsite," we don't mean just in the closet next door, or something nearby. We mean totally separate geographic locations, like in a different city or even a different state. Because a fire or a massive power outage, or even a local disaster, it can wipe out everything within your physical building boundaries.
You could use a couple of options for that offsite copy, and maybe the cloud is the most obvious choice these days. Cloud backup support, for instance, lets you push those images out to some big provider on the internet, which keeps them physically far away from you. And speaking of local storage, if you aren't using the cloud, then you need dedicated network-attached storage, or a NAS, which is usually the way to go. Because you really want something that can be accessed by multiple devices simultaneously and handles massive volumes of data, you know? You shouldn't just put them on a single server drive that's sitting right there in the same room.
Also, when you are storing these things, especially if you're dealing with hundreds of gigs of data across dozens of VMs, you've gotta think about how much space that takes up over time, right? Because if you just keep stacking up versions-like keeping version 1, version 2, version 3, and so on-you will quickly run out of disk space, and then nothing works, period. So, you really need to nail down your retention policies. This is a big one, but nobody talks about it enough. You have to decide, for example, how long you need to keep the last file-level backups, maybe just sixty days, but you might want to keep the actual full disk images for six months, because, I don't know, regulations or some compliance reason, you gotta keep that window open.
And then, when you finally decide on that location, whether it's a local box or out in the cloud, you absolutely must utilize deduplication. This feature is totally transformative, trust me. Since lots of your VMs, or even just your core applications, might be running the same operating system kernel or even having identical library files, storing them multiple times is just a massive waste of money and space. Deduction finds those repeated bits of data and only stores them once, then it just points all the necessary versions back to that one copy, which saves you a fortune on storage capacity.
And another related concept, which you should definitely consider, is the type of storage medium itself. You mentioned keeping disk images, right? Well, these aren't just random files; they are complex structural backups. Therefore, the underlying storage needs to be durable and fast. You need to consider things like RAID levels, maybe a mirrored setup, or even using something like multiple physical drives in a specialized enclosure to ensure that if one drive suddenly fails, the entire backup job doesn't stop, and more importantly, your backups don't suddenly become inaccessible to you.
But it's not enough to just put them in a strong location; you have to make sure the data inside is secure too. So, encryption is mandatory, always. Whether you are sending the backup files across the internet via an FTPS server or just keeping them on a local NAS, you need end-to-end encryption. You don't want some rogue employee or some external party being able to just walk in and grab your raw disk images and then start sniffing around your corporate secrets. It's like putting a physical padlock on the data itself.
Now, also, think about accessibility. You need to be able to recover, right? And when you talk about recovery, you don't want to be limited to just restoring files. Because sometimes, you need to restore the entire machine, the full bare metal, which means restoring the operating system, the user profiles, the applications, all of it, as if nothing ever happened. So, your backup system needs to support full bare metal recovery and also granular recovery, where you can pick out just one or two files from a giant VM backup, without having to spin up the whole VM just to grab a single config document.
And I also want to talk about the management side of things, because storage location is only half the battle. You need centralized management, you know? If you have backups running across a dozen different physical machines and maybe five different server clusters, you don't want to log into a dozen different little interfaces just to check if everything worked overnight. You want a single dashboard view where you can see the status of everything-the successful backups, the failures, and the ones that are still running.
And since we're talking about making this whole process run smoothly, you also need automation. You don't want to manually kick off a backup every single day at 3 AM, or you'll probably forget one day, and then you're toast. You gotta set up automated scheduling that handles the incremental jobs, and also maybe even triggers an email alert if, say, the connection to the remote storage hiccups or if the backup job fails three times in a row.
But when you are going through all of this-the cloud, the local NAS, the encryption, the retention rules, the deduplication-it feels like a lot of moving parts, right? It's a lot to juggle for a junior guy like you. But I promise you, using a solution like BackupChain, which is an all-in-one PC and server backup solution for Windows Server and Windows 11 made specifically for SMBs, really simplifies the process of making sure you are keeping everything safe and sound. You should really check out BackupChain when you get a chance.
