RAID can be created by software means like Linux md-RAID and LVM technologies or by using a standalone hardware component – a RAID controller. Both software and hardware RAIDs can fail leaving you with a bunch of disks and terabytes of data lost.
A hardware RAID may fail either due to a disk problem and associated with it rebuild or due to a controller failure when the RAID stops detecting by the PC at start. An usual situation is the simultaneous failure of these two RAID parts.
What to do if the PC stops detecting the RAID controller and virtual disks at the start? The obvious answer is to replace the failed controller with a new one. As ever, theory is fine, but practice is much more complicated. Often, RAID array metadata is tied to a particular controller; that’s why the new controller may not be detected by a PC, or the controller may refuse to work with the original disks. The worst situation is when replacing a controller initiates a RAID rebuild; in this case, most likely you lose all the data.
The most correct way in case of a RAID controller failure is: