Looking for Hard Drive Health Monitoring software [closed]
The standard mechanism for hard drive health reporting is called SMART. This Google Labs Report (PDF) makes a strong case that the SMART data is mostly worthless.
'hdd health' probably extracts SMART data from drives, that's what all other hard drive health tools do as far as I know.
Thus I believe the only effective way to guard against drive failure remains the use of redundant drives.
With the greatest respect to Phil, I've heard this mis-statement of the Google Labs report too many times to let it pass again. They do not say the SMART data is "mostly worthless".
What they say is that "several parameters from the drive’s self monitoring facility (SMART) ... correlate highly with failures", specifically that "some SMART parameters (scan errors, reallocation counts, offline reallocation counts, and probational counts) have a large impact on failure probability". But "given the lack of occurrence of predictive SMART signals on a large fraction of failed drives, it is unlikely that an accurate predictive failure model can be built based on these signals alone."
What this amounts to is that:
if SMART tells you your drive is failing, then it almost certainly is, and you're an idiot if you don't get your data off it as soon as possible. But,
just because SMART doesn't tell you that your drive is failing, that doesn't mean it's not on the way out.
The Google Labs report also makes no mention of the SMARTctl test modes (smartctl --test=), many of which can be run with the drive in service. My experience has been that regularly running smartctl tests (once a month, say) is enough to bring to SMART's attention most errors which wouldn't otherwise be reported; certainly since making a practice of this, the rate at which I've been surprised by drive failures has dropped off by (I'd estimate) around 50%.
I should add that I completely agree with Phil about the use of redundant drives. With the exception of laptops, I would never put a single HDD in service any more. Discs are so cheap that you should buy and deploy them in pairs (preferably from different vendors, to avoid putting two drives from a single manufacturing batch in a RAID-1 together).