dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
12
share rss forum feed


koitsu
Premium,MVM
join:2002-07-16
Mountain View, CA
kudos:23
reply to robman50

Re: [hard drive] Does anyone know how to read the SMART data?

I'm not sure why we're still discussing this *laugh*.

There is absolutely nothing wrong with the RAW_VALUE field containing a very high number for Attribute 3. It does not indicate there is something wrong with your drive, and any software which tells you otherwise is wrong/broken/designed incorrectly. Again, it's because raw attribute data (RAW_VALUE) can be encoded in any format the vendor chooses; there is no standard. Manufacturer utilities probably know how to decode RAW_VALUE into something sane, but it certainly varies per vendor.

In the above scenario, the only data you can rely on is what's labelled VALUE in said software, and compare it to THRESH. VALUE, WORST, and THRESH are what's known as "adjusted values", based on vendor-specific formulas that are undocumented.

VALUE represents the "current adjusted value" of the attribute, WORST represents the worst VALUE ever seen, and THRESH indicates the point when, if VALUE exceeds it, the overall SMART health status for the disk will trip/fail. Some attributes increment as they get worse, others decrement (see screen shot, example Attribute 3; VALUE=186, THRESH=21) -- again, it varies per vendor and even per drive model. I cannot stress this last point enough.

WORST is often set to something generally absurd; meaning the chance of it being reached (tripping) is extremely low. Vendors appear to do this solely to diminish the number of RMAs they get / gain the highest financial yield on drives in the field. For example, I've seen some disks where Attribute 5's RAW_VALUE is an excessively high number (and isn't encoded), yet VALUE was still far, far away from reaching THRESH, and the disk was seeing more and more blocks go bad.

There's literally nothing you can do about this other than learn how to read attributes correctly; with SMART, a human being familiar with attribute data will almost certainly be more reliable than software trying to achieve the same. I do not use SMART software to "monitor my disk for problems" -- I use smartmontools to show me the SMART data associated with my disks and *I* interpret the results. I advocate strongly that everyone else do the same.

And to reiterate: each SMART attribute is different. To read SMART reliably, you have to become familiar with each manufacturer/vendor and many models of hard disks. WD, Seagate, Maxtor (pre and post-Seagate buyout) Hitachi, IBM (pre and post-Hitachi buyout), Fujitsu, Samsung, Intel SSDs, OCZ SSDs, Crucial SSDs, etc. all differ -- then on top of that, each drive model can differ. So, you just have to get familiar with your disks and know what's what. The lack of ACS/ATA standard for SMART data, formulas, etc. is the core reason for this.

Here's a perfect example, with a Seagate disk, of a user "freaking out" and reading RAW_VALUE literally. Attributes 7, 0xBE, and 0xC3 are all vendor-encoded. None of their VALUEs have exceeded THRESH. Does that mean the drive is healthy? Most likely, but it depends on the attribute. In the users' case, there is nothing wrong with the disk at all (it looks perfect).
--
Making life hard for others since 1977.
I speak for myself and not my employer/affiliates of my employer.


robman50

join:2010-12-14

ok thanks.
i'm sure there is nothing wrong with it.