Has one of the disks in your RAID 5 configuration died, or is this really a RAID 1 or RAID 0 configuration? If a RAID 5 member has died,, then I wouldn't use the old server for testing until either a third disk becomes available, or you reconfigure the the RAID for an approriate config for two drives.
Please refer to this post which I commented on earlier:
https://www-secure.symantec.com/connect/forums/test-restore-group-policy-aborted
You can store your server images on a network location, but only boot your target computer in the production environment from the System Recovery Disk (SRD) CD. When the recovery process is finished, disconnect it fro the production enviroment, and connect it to your lab environment. You can then boot the recovered system without affecting you production environment
You should always test recovered systems separate from your production network, especially if the system is either a domain controller or mail server. If you must run a recovered, non domain controller or mail server in a production environment for testing purposes, then sysprep the unit and give it another name. It will be less troublesome.
An inexpensive switch or hub works well enough for a lab environment.. It is good that you have a server class unit to test with, but you can also use Workstation class computers for lab systems. The restores will not be as complicated, but you'll get practice with the mechanics of the recovery process. I also tend to buy my replacement servers unconfigured, so that I can test a recovery to disimilar server class hardware from time to time.
