For your workstations the solution should be quite simple. You just create images/mirrors/ghosts of your standard installations. If anything goes wrong that image can be installed in a matter of minutes

For a server with a RAID5 array, you can afford to lose one drive; with RAID6 this is two drives. Both systems "stripe" data across all the drives in the array, so once it has been in use you cannot selectively replace individual drives as you have no idea where the data physically are.

If you lose more than the allowed number of drives you will have to rebuild the entire array from scratch.

An alternative would be to use a RAID1+5 or similar setup. Here you mirror the operating system, applications, and static data and use striping for the transaction data. This allows for business continuity, although the data would have to be subsequently input to the system once it was restored.

If you lose a pair of mirrored drives then once again you have to rebuild from scratch. The operating system and applications should be easily applied from your latest mirror/ghost image, and the standing data (price lists, customer files, metadata etc.) from your backups.

My main concern regarding cloning an active RAID array is that it sets a restore point. You will have processed data since that point so would need an incremental backup or journalling system to bring your system back up to date.

A lot of cruder backup mechanisms will backup all the data, so you cannot apply it to a cloned system without getting a mass of duplicates etc.