RAID Discomfort Confirmed

Against my paranoid instincts, the fileserver was migrated to a RAID 5 array. The space gains were a palpable relief. However, the motherboard platform I've moved the fileserver to has trouble running this onboard hardware RAID array and supporting the RAID controller card installed to run the simple mirrored array conceived as a more stable complement to the gigantic RAID 5 drive. 

I tried many different solutions to get these two controllers to work alongside each other, but nothing worked. A few sources I located online indicated the motherboard chipset lacks the necessary ROM to support both instruction sets simultaneously, therefore requiring the onboard RAID array to be managed via a software solution instead of the built-in hardware controller. This seemed reasonable, but undesirable, so I wanted to believe it was untrue. 

Determined to shoehorn my intended solution into the system, I decided to update the motherboard firmware in the hope it would more adequately allocate its own resources with an update. It might, I don't know. Upon rebooting after the firmware update, I observed the controller card was properly detected and I was able to enter the card's BIOS to set up the RAID 1.

What I failed to realize, at this point, was the onboard controller had ceased to function. I am not entirely certain if this was because the board truly cannot support both controllers or because the reset BIOS caused the onboard controller to simply reset itself. Either way, the RAID 5 was lost. 

This was something of a tragedy, given that all of the data had already been migrated back to the RAID 5 array in an effort to finalize the build and allow hard drives to return to their intended locations. The failure of the RAID array meant all of the 5TB of data stored there was lost. Forever.

In a panic, the system was rebooted to recover the RAID, but the drives had already dropped out and Windows had rewritten their allocation tables. Or whatever they are called, honestly I don't know enough about how hard drives work. What I do know is that losing this data was an unmitigated catastrophe and motivated my next, very important decision. 

I removed the RAID controller from the fileserver and decided to move to external RAID expansion efforts. The benefits of this approach are numerous. First, it lightens the processing burden of RAID calculations, freeing system resources for managing parity calculations for the single internal RAID 5 array and freeing processing time for potential transcode requests.Second, it isolates vulnerabilities. Before, the single fileserver computer system dictated the security of all data managed within whereas external arrays can be migrated as needed to other systems in the event of failure. Third, it lightens the power and cooling constraints so far imposed, meaning it will be feasible to move toward a quieter solution in the future. 

The drawbacks are relatively few. Mainly, I think there is a concern about reduced throughput, as all data will now transmit via USB protocols. This could prove to be a problem, but I am hopeful it will only represent a minor inconvenience. 

In the end, the RAID 5 array was eventually recovered, but that is another story.