On hardware type used (i.e. a blade which only has 2 drive bays) and on data center security. Is it a secured dark site? Only limited technicians have access rights? How is the SLA and hardware replacement dealing with restrictions?
Do you bitlocker on the DCs? (That is actually another good questions because bitlocker info is stroed in AD and Bitlocker security has been compromised already in part (?!).... the list can be endless.
If you are paranoid someone taking the disk can using the same hardware to attack it with a bunch of Fermi-powered GPUs, don't do it. and then better not use VM technology as well.
You need to weigh risk, cost and administratrative overhead into this...
Other than the security issues mentioned by fghoffmann, I do like the fact raid 1 provides some hardware falut redundancy.
Promoting a new DC when the old one failed is a pain especially when other DCs are over WAN. Depends on the load of the DC, I personally would not use VM or blade.
Blade is just too problemaic and VM is way too risky. If load is high, consider moving the dit file and log to different volume(s), ie. a different raid 1 volume.
