Posted 16 January 2010 - 04:22 PM
I was also really hoping as I'm running out of rack space for the 1TB drives, however if they are still following a 1:10^15 bit uncorrectable error rate that's still insane at 12% per full drive read. For enterprise use (at least using their own label) even in nearline storage where these would be used you would do full checks at least weekly, plus whatever production I/O is going to/from the subsystems.
I know they know how to do it (their tier-1 drives all have 10^16 already). Sure w/ 512byte blocks it takes up more space on the platter for the ECC functions, but I would be more than willing to shell out the same $$ for a 1.5TB drive at 10^16 rating than for a 2TB at 10^15. (heck, would love this on even 1TB drives).