DR used to be the one thing every administrator treated like cod liver oil, horrible tasting but with many upsides. I recently wrote about it here for SearchVMware.
Essentially there are a number of ways to think about DR. In that article I give some essential bits of advice as to what works and what doesn’t. Essentially it comes down to cutting your cloth accordingly and planning it properly.
Sure it aint cheap at high several hundred dollars a node but it comes down to a couple of things to bear in mind:
- Administrators don’t need to replicate everything. Think core apps and essential services (DNS, AD etc)
- Virtual DR only works with virtual machines. No physical database servers allowed!
- Hardware and link capacity. The DR site should have sufficient hardware capacity to run the applications as speedily as the live site. The last thing you want in a DR scenario is slow machines slowing things down further.
- Have a DR plan and test it. An untested plan is more or less guarenteed to have gotchas.
If you are looking for a good intro to the do’s and don’ts of virtual DR you may want to take a look.