Do SMART values have any sense at all?

1

Is it possible to understand SMART values somehow?

Here is an example of values shown for one of my drives:

enter image description here

Following questions arise:

  1. If Reallocated Sectors Count is 2 and threshold is 36, then why is marked with red? Threshold is not exceeded yet!
  2. What is FB3? From hex it is 4019, i.e. neither 36 nor 2. Where it took the values then?
  3. Current Pending Sector Count is 100 and threshold is 0. Threshold is exceeded. Why is it marked with yellow then, not red?
  4. Why are there so much 100 values? Looks like it is stub value. How it can know status then, if the value is stubbed?
  5. What is B? From hex it is 11. Neither 0, nor 100. Where it too value then?

So, the only understandable value is the color of knob at left, while all other values are filled with delirium. Is it possible to understand here something?

Note that by some reason, this delirium is repeated from one SMART application to another.

Dims

Posted 2016-06-29T08:58:03.600

Reputation: 8 464

Interesting read: https://www.backblaze.com/blog/hard-drive-smart-stats --- according to that statistic your drive has an annual failure rate of 150%. Replace/Backup it ASAP if you care about the data!

– mpy – 2016-06-29T20:06:18.000

Answers

1

  1. Because the actual value of reallocated sectors is 4019 (it is the raw value that represents the count of the bad sectors that have been found and remapped), which exceeds flag value 2 (which is 36) by a lot.

  2. FB3 is 4019 in HEX.

  3. Because the actual value (hex:B = 11) is not close to 100, more specifically, in your case, 100-11 is not close to 0, yet. See 5. (related).

  4. They are chosen by the manufacturer.

  5. Some values are set as max by default and actually decrement or are based on differences from 100 (or other values). See 3. (related).

Basically, it's the raw values that in many cases should interest you a lot.

The conclusion is that your HDD has way too many reallocated sectors and a few uncorrectable ones. you should consider replacement.

Here's some good info on the subject.

As for values:

  • Current/Normalized: This is the above value normalized so a higher value is always better for most of the values (total R/W is an exception example). So 95 in spin-up time is better than 94. This also can vary from one manufacturer to another, which can introduce confusion.

  • Worst: The worst normalized value that your HDD had in the past (values like 99 or 199 are common default settings).

  • Threshold: When the normalized value is lower than this value the HDD is likely to fail. Note your 36 value compared to the Raw. That's why it's red.

  • Actual/Raw Value: This is the raw value that the controller reports. It's most of the time an easy to understand value, but it can also vary from one manufacturer to another.

Overmind

Posted 2016-06-29T08:58:03.600

Reputation: 8 562

I suppose that your explanation of point 3 is correct. There are two types of SMART parameters: pre-fail and old_age (see this question for example). My guess is red indicates a problem with pre-fail parameter, yellow is a problem with old_age parameter.

– gronostaj – 2016-06-29T10:14:46.800

The most of the confusion comes from how they are counted. Some values increase to a max red-line value (N), some decrease to 0 by N-V (N - reference value, V - current value). In the current 'yellow' case, it is not important when the errors occurred because they are pending errors (sectors failed to be reallocated). – Overmind – 2016-06-29T10:19:45.770

@Overmind consider your p.1 Why are you comparing "raw" value with "current" value? Looks like you just compared whatever matched mostly, ignoring column titles. It doesn't make any sense. Also try to apply this approach to other lines -- and you will get even more nonsense. – Dims – 2016-06-29T16:44:49.223

Do not be confused. Raw values can be totally different from the rest (some are even composite/2n1 like F0), they rarely can be directly compared. Many Raws are just counters for various stuff that does not even affect SMART status. – Overmind – 2016-06-30T05:40:38.307

@Overmind but why do you compare 0xFB3 with 2 then? You said "value of reallocated sectors is 4019 ... exceeds 2 by a lot", i.e. you did compare. – Dims – 2016-06-30T08:03:00.263

I don't compare as pure numbers. 0, 1 and 2 may be for example values for 0 (good), 25(yellow alert) and 100(red alert) reallocated sectors. This is highly dependent on manufacturer and drive model. In the current case, 2 is the equivalent of the 36 value, which is a lot exceeded by FB3. So 2 is the flag, 36 is the value of red alert limit which is exceeded significantly, as I stated at 'thershold'. I'll add a note to initial post. – Overmind – 2016-06-30T08:26:34.203

Also note that the raw value of 05 attribute shows the total number of ATTEMPTS to transfer data from reallocated sectors to spare sectors, not just of successful transfers. – Overmind – 2016-06-30T08:39:01.120

1

Current, Worst, Threshold are normalized values as opposed to Raw Values.

  1. HDD has 0xFB3 (4019) reallocated sectors, which yields a normalized current score of 2, way below the threshold of 36.
  2. Raw Values as returned by internal sensors/counters.
  3. HDD has 0xB (11) sectors pending reallocation.
  4. Good Question! My guess is that these attributes might not be critical so manufacturer does not bother to normalize them properly.

guest-vm

Posted 2016-06-29T08:58:03.600

Reputation: 2 984

Asked: 2016-06-29T08:58:03.600

Viewed: 1 013 times

Active: 2016-06-30T08:27:36.973