Something that used to plague IT Administrators in the early 1990s, RAID controllers that came as a card that plugged into a computer were not entirely reliable, and when we built servers, we would often use multiple RAID controllers for additional redundancy.
Nowadays, the failure of a RAID controller is fairly rare. Unless, of course, we are talking about a NAS box, or any storage array with a built-in RAID controller. For reasons unknown to use, these do fail. Not on a regular basis, but enough to leave us wondering why this is the case. Everything is manufactured in China these days – so why embedded RAID controllers are more susceptible than discrete controllers remains a bit of a mystery.
That said, the failure of a RAID controller will immediately bring your RAID to an abrupt halt. To an experienced data recovery company, this is not a major issue. But to the IT administrator, it can cause headaches. One reason is because the RAID parameters are stored on the RAID controller itself (we’re talking hardware RAID here – not software). So replacing the controller with an identical model is not immediately going to solve the problem – and it could make things an awful lot worse.
The main issue with simply installing a new RAID controller, is that it is going to want to reinitialise the hard drives attached to it – otherwise they simply won’t be recognised. Many users will naively accept the controller’s recommendations, and proceed with the reinitialisation. And why not? What good are hard disks if the controller can’t see them?
However, the next step sees the user, encouraged by the suggestions put forward by the controller’s configuration page, will then tend to rebuild the array with the original disks. And he will typically use the default settings. Which is great – unless the previous RAID was configured for data-specific performance. (Large stripe size for large files such as video, smaller stripe for normal user data such as email, word processing, spreadsheets, etc.)
We suddenly arrive at the situation where the original RAID parameters have been trashed and, accordingly, the RAID and data availability itself. At this stage, the user has no option but to admit defeat, power the system down, and seek professional data recovery advice.
If your controller fails, don’t adopt a gung-ho attitude unless you know precisely what the previous configuration was. Even then, it’s best to have secure backups of either all your data, or of each hard drive in the array before trying to install a replacement RAID controller.
