Reliability of S.M.A.R.T.?

Question

I've been using ActiveSmart to monitor my hard-drives health for a few weeks now, and its telling me my brand new 1.5 TB hard-drive is half-dead already. About on-par with one of my hard-drives which I know is at least half dead because I've been having some read errors and heard ticking noises. Now I haven't actually noticed any problems with my 1.5 TB drive; should I be concerned that it's going to crap out on me too? Or could ActiveSmart be giving a mis-diagnosis because I use it a lot or something (I've used up 795 GB in the 2 and a half weeks I've had it). The events that ActiveSmart has been catching is "Hardware ECC recovered". Maybe these new fangled super big hard-drives somehow rely on ECCs to squeeze out the extra space, but this isn't actually a cause for concern?

Answer 1

Modern hard drives have massive levels of ECC. This allows them to handle the massive amount of noise in the read heads. The 40 byte ECC from a 2004 vintage drive can recover from a 50 consecutive bit error. The size of the ECC block is steadily increasing with drive density so modern drives have even bigger ones (because the higher densities have a higher error probability).

http://arstechnica.com/microsoft/news/2010/03/why-new-hard-disks-might-not-be-much-fun-for-xp-users.ars

As a side note, of the half dozen HD's I've seen fail only 1 ever generated SMART errors before bricking itself, and it didn't start until suffering drop damage sufficient to render the OS unbootable. I wouldn't put much faith in it's ability to detect problems in advance.

Answer 2

The high density of bits on that drive will lead to more reading errors which are corrected by the drive's firmware. I don't think it's a huge concern. Wikipedia doesn't list the value as critical either (source: http://en.wikipedia.org/wiki/S.M.A.R.T).