Wait, so when you click the "Mark drive as normal" link next to a failed drive in the Intel RSTe application, all of the array information is destroyed and it is immediately impossible to recover from the problem without leaving the failed drive in until the rebuild process completes? So having a drive that could just as well be blank and impossible to read or write information from/to in place for the rebuild process somehow makes it possible for the rebuild process to magically divine the information that used to be stored on the drives? Why shouldn't any blank drive of the same size be ok? What if I just put in a drive that has the same model and serial number (somehow)?
I have two of the original three drives in my computer in perfectly-working condition, which is all that you should need to rebuild a RAID5 array. If I had known that attempting to recover a failed drive would cause the Intel RSTe program to destroy my array configuration, I would have passed on the idea. That kind of option shouldn't even be allowed to exist! Why wouldn't marking a drive as normal, starting the rebuild process, simply consist of reading the stripes from the two good drives and attempting to reassemble the real or parity information on the third drive? Why would anything have to be destroyed on the two good drives?
