Question SSD uncorrecttable errors

[Juforq]

I'm not sure how bad it is, but I checked my SSD and found some uncorrectable errors. The raw value is 82. I still have a few weeks of warranty left (bought it 3 years ago), should I try to get the disk replaced or is it ok to continue using it?

 

[Colif]

I wonder if thats Reported Uncorrectable errors or Uncorrectable Sector count which seems unlikely on an ssd, or uncorrectable soft read error rate. I blame Kingston - https://media.kingston.com/support/downloads/MKP_306_SMART_attribute.pdf
they could have been more obvious.

I assume it (based on this)
187: (SSD Reported Uncorrectable Errors)
100 is best, 1 is worse, drive isn't that bad at 82. You should be fine

 

[SkyNetRising]

I checked my SSD and found some uncorrectable errors. The raw value is 82.

What model drive?
Can you show screenshot from HDtune health?

 

[Juforq]

What model drive?
Can you show screenshot from HDtune health?

It's Kingston A400 [SA400S37/240G]
Edit: I should also probably mention that this value remained unchanged in the last few months.

 

[SkyNetRising]

It's Kingston A400 [SA400S37/240G]
Edit: I should also probably mention that this value remained unchanged in the last few months.

It is still considered fine. I don't think warranty replacement is applicable here.
SSD life parameter has decreased significantly. But you still have almost 2 years of ssd life left.

Just be ready, you'll have to replace it eventually. Have backups for your important data in case of a sudden failure.

 

[Juforq]

100 is best, 1 is worse, drive isn't that bad at 82. You should be fine

Thanks for the answer. I'm a bit confused, isn't the normalized current value of 1 out of 100 means that SMART thinks the parameter is close to 0, which is a threshold?
I also posted HD Tune results above, and threshold = 0 basically means that SMART test will always return an "OK" status no matter how bad the situation is, unless the drive actually fails completely. Am I correct?

 

[SkyNetRising]

When current value reaches threshold value, drive is considered failed.

 

[Colif]

100 is max score, 0 is lowest and a Fail. 82 is not a fail as you can probably sustain a few more of those but I wouldn't use drive for anything crucial if it got too many more.

its possible there is a score system and if you were below 50 it might say Warning instead of OK. It shouldn't say OK right up until it stops working, it should warn you before then.

HD Sentinal for instance ranks any drive over 50 as excellent, and anything below 50 and above 25 as fair, and anything below 25 as Bad, so I assume HD Tune is similar, buiit thats just drive health. I have a feeling it be same for individual ratings

 

[Colif]

a) Check the Health screen
Here you will see S.M.A.R.T information which shows the condition of the hard disk.
The status of each parameter should be Ok.
If the parameter is highlighted in yellow it means that the parameter almost reached a critical value. At this point it is highly recommended to check this parameter regularly to see if the value becomes worse.
If the parameter is highlighted in red it means that the parameter has reached a critical value. In this case you should make a backup of your data and replace the hard disk.

http://www.hdtune.com/faq_1.html

since its not highlighted Yellow it is fine

since its been same for a few months, I think you okay to use it normally, Just check it every so often to see if its changed.

 

[Juforq]

since its been same for a few months, I think you okay to use it normally, Just check it every so often to see if its changed.

Actually I double checked now and it seems I was mistaken, the number of Raw Read Error has gone up from 80 to 82 in about four months.

 

[Colif]

well, you aware of it now so just watch it. I use hd sentinel and it has a icon in taskbar that always shows health of my ssd

I believe HD Tune does similar - that way you don't have to do anything to see it.

Raw read error count generally relates to hdd but can be seen on ssd. As this doesn't apply to ssd

RAW Read Error Rate, a SMART disk error, indicates problems with the disk surface (platter that stores the data), the actuator arm, and the read/write head.

But

RAW Read Error Rate on an SSD is an indication that there are some bad connections between disk and the drive controller or the problem is within the SSD. Hence, the SSD may fail after a period of time.

link

from what i can tell the threshold for that score on HDTune is 6

if that score is Raw read error rate, this is what it says in the Kingston ssd manual I linked above

Raw error rate relative to the number of sectors read this power cycle. For the SF-2000, this attribute includes both
Uncorrectable ECC (UECC) errors, and Uncorrectable RAISE (URAISE)errors.
Normalized Equation: 10log10[BitsRead/(ReadErrors + 1)]
SectorsRead= Number of sectors read
SectorsToBits= 512*8
BitsRead= SectorsRead*SectorsToBits

Normalized Value Range:
Best = 120
Worst = 38

Raw Usage:
[3-0] : Number of sectors read this power cycle
[6-4]: Read errors (UECC+URAISE)

So its a score between 120 & 38.

its not very clear

 

[fzabkar]

Kingston's SMART spec doesn't appear to apply to the OP's SSD. It applies to "Kingston SF-2000 Based SSDs" which use the SandForce SF-2000 controller. The OP's SSD uses Phison's PS3111-S11 controller.

This review shows that the Current attribute values start at 100, not 120:

https://goughlui.com/2020/08/02/quick-review-kingston-a400-960gb-sata-ssd-sa400s37-960gb/

In any case there appears to be a bug in the SMART reporting. The Worst value should always be less than or equal to the Current value. The fact that the Current value for the Raw Read Error Rate is at 1 suggests that it has bottomed out. Since it needs to hit a Threshold of 0 before a SMART failure is triggered, this would suggest that the manufacturer has decided that this is not a critical attribute. I would argue otherwise.

The raw value cannot be interpreted in the same way as the normalised values (Current, Worst, Threshold). Instead, think of the raw values like blood pressures (eg 82) and the normalised values as health scores (100%, 1%, 0%).

Kingston's SSD Manager reports attribute 0xAA as Bad Block Count. Therefore it would appear that the drive has replaced 9 bad blocks with spares.