0

Today I stripped down and rebuilt my PC (to get rid of surplus cabling, and to give it a clean since it was pretty dusty in there), and everything is running fine.

Well, except my SSD (apparently).

I run from time-to-time, and I ran it after the rebuild to check that the temperatures were OK. This is what I saw in the S.M.A.R.T tab when I checked my SSD's health:

enter image description here

Meanwhile, the data for my two hard-drives shows this:

enter image description here

My two hard-drives are apparently in much better health than my SSD is, but maybe one thing to note is that the SSD runs on a PCI-E SATA3 controller, and the two hard-drives run on my motherboard's own SATA2 ports.

Should I be concerned, or is SpeedFan reporting a false-reading (aka, my SSD is completely fine)? My SSD is still under warranty, is it advisable to RMA it (since maybe this is a sign of imminent total failure)?

According to this forum thread:

Using your OCZ Toolbox app, your SSD's life curve SMART attribute is what you should watch; "0" is perfect

The life curve on my SSD is reporting as 100, so does this mean that my SSD is on pre-failure:

230 Life Curve Status                     0x0013  100 100 000 Always      100

The S.M.A.R.T data from OCZ Toolbox:

SMART Data

Model Number:   OCZ-AGILITY3                            
Serial Number:  OCZ-KNL379LKBSM8E68A
WWN:        5e83a97e4438904c

 ID ATTRIBUTE                           STATUS  VALUE   WORST   THRESHOLD   UPDATED RAW
  1 Raw Read Error Rate                 0x000f  108 108 050 Always      0/18925461
  5 Retired Block Count                 0x0033  100 100 003 Always      0
  9 Power-On Hours                      0x0032  098 098 000 Always      2415h+13m+56.580s
 12 Device Power Cycle Count            0x0032  100 100 000 Always      244
171 Program Fail Count                  0x0032  000 000 000 Always      0
172 Erase Fail Count                    0x0032  000 000 000 Always      0
174 Unexpected Power Loss               0x0030  000 000 000 Offline     707
177 Wear Range Delta                    0x0000  000 000 000 Offline     3
181 Program Fail Count                  0x0032  000 000 000 Always      0
182 Erase Fail Count                    0x0032  000 000 000 Always      0
187 Reported Uncorrectable              0x0032  100 100 000 Always      0
194 Temperature Celsius                 0x0022  030 030 000 Always      30 (Min/Max 30/30)
195 ECC On-the-Fly Error Count          0x001c  120 120 000 Offline     0/18925461
196 Reallocation Event Count            0x0033  100 100 003 Always      0
201 Uncorrectable Soft Read Error Rate  0x001c  120 120 000 Offline     0/18925461
204 Soft ECC Correction Rate            0x001c  120 120 000 Offline     0/18925461
230 Life Curve Status                   0x0013  100 100 000 Always      100
231 SSD Life Left                       0x0013  100 100 010 Always      0
241 Lifetime Writes from Host           0x0032  000 000 000 Always      3731
242 Lifetime Reads to Host              0x0032  000 000 000 Always      6430
5
  • Its to be expected that the rellocation value would be better on a SSD since data is continously being moved around within an SSD.
    – Ramhound
    Sep 2 '14 at 19:53
  • @Ramhound Don't you mean 'worse' or 'higher'?
    – AStopher
    Sep 2 '14 at 19:54
  • I mean better. I am saying that its to be epxected that the reallocation sector would be higher ( i.e. better ) because the firmware is continously doing write protection mechanics. This means until the end of a specific life of a cell the data is unlikely to have trouble being read/written which often happens with a mechanical disk which has physical platters which become unreable in certain sections of the platter itself. I say exactly what i mean by the way.
    – Ramhound
    Sep 2 '14 at 19:59
  • @Ramhound Gotcha.
    – AStopher
    Sep 2 '14 at 20:00
  • But as pointed out, unless you also monitor the amount of data SSD has written, the S.M.A.R.T data hold little value by itself. The data itself simply indicates firmware believes the drive is healthy. Without using SSD specific tools S.M.A.R.T data by itself hold little value.
    – Ramhound
    Sep 2 '14 at 20:03
5

There is nothing wrong with that SSD, the self-test even says "no error" and the SMART attributes look fine. This is just the case of the tool (SpeedFan) having no idea about the relevant SMART attributes on OCZ's SSDs, thus the bars displayed at the bottom are worthless.

Actually, your SSD is in much better shape than the two harddrives you tested. I'll explain why:

In general, SMART counts down: The values are initially at 100 (or more), and go down to 0 when there are errors. The "worst" value is thus the lowest one. If any values are below the "Warn" value (PREFAIL), then that is an indication for (future) disk failure. Note however, that not all attributes have an impact on warranty: In the case of an RMA, only very few SMART attributes are considered when determining if the drive warrants replacement. Usually they are "Reported Uncorrectable Errors" and "Reallocated Sector Count", but it varies by manufacturer (some won't RMA until the drive actually stops working).

You need to use the OCZ Toolbox to read out the SMART values if you want them interpreted for you in more detail. Also, if any values are indeed PREFAIL you need to contact OCZ to check if that particular attribute/issue is covered by the warranty.

Note that "Temperature", "Power On Hours" and similar attributes don't have a "Warn" value (or it's set to 0), as they do not indicate failure.

A word on "raw" values: These are the actual values measured (e.g. number of sectors reallocated). These values are then normalized to the 100-0 range as explained above. In order to understand what the raw numbers actually count, you need to look at the manual for your harddrive/SSD as it varies wildly between manufacturers.

EDIT: In particular

                              VAL WOR THR        RAW
230 Life Curve Status  0x0013 100 100 000 Always 100

means everything is fine. As the normalized value (100) is greater than the threshold (0), everything is OK.

4
  • Added the S.M.A.R.T info from OCZ Toolbox (I actually downloaded + ran it just as you posted your answer).
    – AStopher
    Sep 2 '14 at 19:41
  • @zyboxenterprises Yes, still the same result. These values say that the SSD is in mint condition. However, from personal experience I can say that SMART isn't very reliable when it comes to SSDs: They can fail from one day to the other with no warning, even if SMART says they're fine.
    – jmiserez
    Sep 2 '14 at 19:43
  • I updated the question with more info.
    – AStopher
    Sep 2 '14 at 19:45
  • 1
    To put it bluntly, that comment in the forum thread is wrong. SMART is standardized, the normalized (not raw!) values are always bigger = better. Also, if you read the thread correctly, you'll see that they also have "100" as the value and not the stated "0", which makes the statement most likely a typo.
    – jmiserez
    Sep 2 '14 at 19:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.