I have published many posts on backup and on virtualization, this post will bring both areas together for “the win”. And what is the win, you may ask? That’s simple, it is Disaster Recovery that even small businesses can afford! This is the first of a multi-post series on DR for smaller organizations. In this post I’ll touch on the various bits and pieces and concepts. In subsequent posts I’ll demo the process.
Disaster Recovery (DR) in a nutshell is the ability to quickly bring your IT infrastructure back online after some form of disaster. A disaster might be something as simple as a failed server or something much worse like a catastrophic fire. Whatever the disaster, the goal is not necessarily to have the same performance level as your normal production systems but at least be able to function on a basic level. This DR ability could be the difference between surviving a disaster or going under because your business simply cannot function.
From my point of view, the enabling tools to providing this ability are easily within the reach of any organization: Virtualized environments provided by hypervisors such as Microsoft Hyper-V and/or VMware ESXi and hypervisor aware backup tools such as Altaro VM Backup and/or Veeam Backup and Recovery. What these products do is give you the tools to build a DR framework and process that are maintainable and affordable for the average small organization.
Both Altaro and Veeam provide the tools for backing up and recovering virtualized environments. In my view, Altaro is better suited to smaller shops as it is priced much more attractively than Veeam and is far better suited to a one or two host-server installation. Veeam is much more feature-rich and is priced accordingly; it is suited to larger installations although many smaller shops also use it. I really like both products and both have their individual strengths; the subject of this post, DR for smaller organizations, is product “agnostic”; either backup product will work for you.
So, what are the basic elements required to be able to build a DR framework and process? There are many things that can come into play but I think the must haves are as follows:
– Virtualized servers on one of the previously mentioned platforms
– Properly configured and tested backup using Altaro or Veeam
– Backup copies sent to offsite location using the remote copy process in Altaro or Veeam
– Offsite “endpoint” for Altaro or Veeam (see notes below)
– Offsite virtualization host for VM recovery
Each element provides a building block for your DR framework.
First off, virtualization is the basic building block because it divorces your servers (the VM’s) from the underlying hardware which makes it “easy” to recover to other hardware if required. If you have multiple hosts you have the ability to teleport VM’s around between your hosts which in and of itself is very useful. There are many other advantages to virtualizing that I won’t go into here but one main advantage is that backup of VM’s can be very fast when using a virtualization specific backup tool such as Altaro or Veeam. Backup can be many times faster than with file-based (traditional) backup tools and the same can be said for recovery. So even if you don’t plan on building an offsite DR you are already ahead of the game if you use these technologies.
Second, Altaro and Veeam are both designed specifically to backup and recover virtualized environments in Hyper-V or VMware. Both leverage the baked in capabilities of the underlying hypervisor and both can backup and recover VM data very quickly. Both are capable of recovering whole VM’s or individual files. And both offer a critical piece that is a major enabler for a DR framework and that is the ability to perform “backup copies” to an offsite location. This is critical because people are unreliable, relying on George or Georgina to schlep USB drives or tape cartridges offsite with copies of backup files is just setting yourself up for failure!
Third, note that I have referred to “backup copies” in the preceding paragraph, it is an important distinction. Altaro and Veeam both provide the ability to make a copy of an existing backup within your current backup repository and have that copy go offsite. This distinction is important because it highlights the fact that the process of copying your backups to the offsite location is not tied directly to the process of backing up any given VM. The reason this is important is twofold: First, it means there is no impact on your production VM’s while the backup copy is running as you have already performed the backup of the VM’s. Second it means a failure of the backup copy will not impact your ability to recover VM’s from the main, production backups. This is pretty critical as it means a backup copy across a slow WAN connection is perfectly acceptable from a time window point of view. A live production backup across a slow WAN connection is another thing entirely and is something to be avoided.
There is another reason why the notion of a backup copy is important – the George or Georgina factor! Many small organizations fail in their attempts to manage backup because they rely on a George or Georgina to take copies of backup data offsite. Then a disaster hits and they discover the offsite copies are not available, or the media has been damaged or, worse, George or Georgina forgot to take the media offsite! Whatever the case, the end result is not good. Keep this in mind, people are a lousy substitute for good automation!
Finally, there is the offsite virtualization host. Both Altaro and Veeam will require you to have a machine as an “endpoint” for their remote tools to accept and channel the backup copy “stream” to the storage destination (repository). Altaro will work with a Windows desktop OS (7, 8, 8.1, 10) or a Windows Server as the endpoint regardless of what your chosen virtualization technology happens to be. Veeam requires you to have a Windows Server host as the endpoint if you are in the Hyper-V world while a desktop OS is fine if you are in the ESXi world. You will also require a virtualization host so that you have somewhere to restore VM’s. If you are a Hyper-V user your endpoint can also be the Windows Server host that will be used to house the VM’s; if you are an ESXi user you will need an ESXi box. Keep in mind that you do NOT need to have monster resources, you just need enough to run your critical server VM’s in a reduced performance mode.
OK, that’s the very basics. My next post will focus on setting up the whole DR infrastructure and I’ll demo it using Altaro as the tool.