Virtual Machine Disaster Recovery

One of the single greatest features of virtual machines is how easy it is to use for disaster recovery. If you can imagine that an entire server is nothing more than a file on a hard drive, you can easily see how easy it could be to backup and restore a virtual machine. Backup can be as simple as turning off the virtual machine, copying it to another hard drive, then turning the virtual machine back on. That’s it. No magic, no smoke and mirrors, just copy it and go just like a Word or Excel file. If you need to restore it, just copy it back, find the file with the Hypervisor software, and start it right up!

Ok, so maybe it isn’t quite that easy, because depending on how much data is packed into that file it could still take hours to make the copy of the file. What if you can’t afford to turn the server off to back it up, what then? Well, making a copy of the drive is not the only way to back up your virtual server. Backing up the server is also not the only way to recover from a disaster when you are talking about virtual machines. Virtual machines can be moved between physical servers over a network with no downtime. They can also be replicated on a regular basis to another location, and those copies of the virtual server be brought online quickly in the event of a major disaster at your office. You can force a virtual server or group of servers to move to another machine while you take a physical server down for maintenance like upgrading the memory or doing an update to the hypervisor without turning the virtual servers off and interrupting service, then move them back when you finish.

Another feature of virtual machines that helps reduce the risk of a disaster is the snapshot feature. A snapshot is a picture of everything on the virtual machine as it was when the snapshot was taken that can then be rolled back to if a problem comes up. An example would be you are installing a new version of your accounting software, you reboot the server, and before Windows loads it crashes every time you try to restart. Don’t worry. All is not lost, because you took a snapshot just before you started installing the software. Just choose the last snapshot you took in the hypervisor and roll back to it. Now it is like it never happened. If you have already experienced a bad software installation, you might consider making a copy of your virtual server and testing the installation on a copy inside a hypervisor on your laptop disconnected from your network to see if there are going to be any problems, before you do the installation live. For that matter, you could just copy the virtual machine files you tested the installation to back to your live environment and not have to go through the process of installing the software again. In this way, you have completely avoided any risk of a disaster caused by the software installation.

As much as I want you to be excited about how easy and how powerful virtual machines can be in a disaster recovery plan, it is important to note that every network is different and has different needs. The more complicated your data and applications, the more complex the disaster recovery plan will be. The plan may require all your files to still be backed up daily using standard backup software from within the guest server, or SQL databases require special scripts to be run to prepare data for backup prior to the virtual machine being backed up. There are dozens of variables that could complicate the backup and recovery process implemented for a virtual machine, and that is why it is important to work with your IT staff or a knowledgeable consultant you trust to develop and test your backup plan.

Leave a Reply