I'm betting its possible if your drives are sufficiently fragmented and you start a heavy write workload that shotguns sectors all over the disk from multiple T

I'm betting its possible if your drives are sufficiently fragmented and you start a heavy write workload that shotguns sectors all over the disk from multiple TCQs the default write command timeouts on most RAID systems aren't sufficiently long. So the drive may not have failed, the RAID just determined that the command was taking to long, fired a reset at the drive and when it didn't immediately respond marked it bad.