Cloning the hard drive offline on a schedule is not something that I'm comfortable with. With the hard drives offline it is very vulnerable to an untoward change that could render the server unbootable or data damaged. This is true no matter how careful a person is. That is not to say that that sort of thing happens a lot, but taking a live server offline and rendering the operating system's checks and balances of permissions and process ownership completely dead can lead to some scary possibilities.
I recommend two possibilities:
Option One
Use a product that is intended to clone one server to another server while the source server is alive and running. This is not a backup and recovery product, but a high availability product. For example, Marathon everRun or Vision Solution's DoubleTake product.
There's no hacky "take it offline once in a while, back it up and hope we can restore the server if something bad happens - oh and I hope the data is fresh enough to keep us from getting fired" happening here. You can be up and running on the recovery server within seconds using data that is itself potentially only seconds old (or not old at all, depending on how you set it up).
Option Two
Use backup tools that are intended to perform image-based backups of live systems, particularly products that use CDP. The image will likely be block based and have point in time recovery options. You can then recover to different hardware, and it often doesn't even have to be the exact same hardware. JoeQwerty's solution is good enough for many shops - Windows Backup has come a long way.
Whatever you do, I advise you to not take the server down to make the backups. That's flirting with disaster, inefficient and way to much manual labor for a SysAdmin. There are three rules to effective systems administration:
- Automate
- Automate
- Automate
Anything else is just busywork.