• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Samsung 870 EVO - Beware, certain batches prone to failure!

Joined
Jun 30, 2013
Messages
89 (0.02/day)
Certain 870 EVO 4TB and 2TB drives are affected by early failures where they develop uncorrectable errors and some data just cannot be read from them anymore.
This seems to primarily affect drives produced in January/February 2021. For example, i have three 870 EVO 4TB, only one is affected (so far), serial number S6BCNG0R207xxxx from February '21.

This is not an isolated issue, see https://www.reddit.com/r/synology/comments/p15wwq That's just one example of such a thread, there's many more on different forums, just google for "870 evo failing" or the like.

Just today, i really started to notice the issue, some files just couldn't be properly read/copied off the drive, no matter what. Explorer aborts with a failure message, other tools the same.

On my affected drive, CrystalDiskmark reports Health 99%, Wear Leveling Count 5, around 12 TB written. Power on count is 273, "Reallocated Sector Count" is on 329, this last one is already a bit worrying. But that's not the main thing. Each time a file cannot be read, "Uncorrectable Error Count" and "ECC Error Rate" keep rising by a couple dozen or even couple hundred!

I started a while ago with "Uncorrectable Error Count" and "ECC Error Rate" already at over 600, now i tried to copy my files to a HDD, and i'm at 2000 for each value and counting. Each time i try to read a damaged file, it rises. Some files i'm completely unable to copy, it just can't read a certain part of them from the SSD anymore. For some other files, the count seems to increase a little bit, but they still copy over, suggesting they are partly corrupted now. It's hard to give the total stats of how many percent of files are affected, but it's a serious issue already even if it's less than 10% of the files affected.

I saw one or two files failing to be read from this SSD before, but i thought, maybe they were corrupted in some way and didn't pay it much mind, since the SMART still didn't look too terrible. I know from other SSDs that there can be bad blocks even if they are new, so i didn't worry too much. But now this looks very different, i can't even copy some files anymore.

Check your 870 EVO SSDs for these things:
Elevated "Reallocated Sector Count", "Used Reserve Block" and "Runtime Bad Block" count - first warning sign (my two other 870 EVOs have none).
Non-zero "Uncorrectable Error Count" and "ECC Error Rate", and especially if those two keep rising when you read/write files. Definitely affected then!


You can read out the self-monitoring SMART values of your drives with CrystalDiskInfo.
In there under "Function" -> Advanced Feature -> Raw Values, select "10 [DEC]" to have human-readable values. You can then upload screenshots of the data, press CTRL-S to screenshot in there.

CrystalDiskInfo_20220131234924.png
 
Last edited:

eidairaman1

The Exiled Airman
Joined
Jul 2, 2007
Messages
42,639 (6.68/day)
Location
Republic of Texas (True Patriot)
System Name PCGOD
Processor AMD FX 8350@ 5.0GHz
Motherboard Asus TUF 990FX Sabertooth R2 2901 Bios
Cooling Scythe Ashura, 2×BitFenix 230mm Spectre Pro LED (Blue,Green), 2x BitFenix 140mm Spectre Pro LED
Memory 16 GB Gskill Ripjaws X 2133 (2400 OC, 10-10-12-20-20, 1T, 1.65V)
Video Card(s) AMD Radeon 290 Sapphire Vapor-X
Storage Samsung 840 Pro 256GB, WD Velociraptor 1TB
Display(s) NEC Multisync LCD 1700V (Display Port Adapter)
Case AeroCool Xpredator Evil Blue Edition
Audio Device(s) Creative Labs Sound Blaster ZxR
Power Supply Seasonic 1250 XM2 Series (XP3)
Mouse Roccat Kone XTD
Keyboard Roccat Ryos MK Pro
Software Windows 7 Pro 64
Joined
Jun 30, 2013
Messages
89 (0.02/day)
Yep, i will. First i have to copy everything off it, as much as i can anyway. The data is not super important (i have backups of super important data), but it's still very annoying.

Especially since i spent extra to get 870 EVOs over, say, an MX500. The 870 EVO is one of the most expensive SATA drives for the consumer market, and then this...

Plus it's not just about me, there's the much bigger issue at hand here. By the looks of it, we have a large batch of 870 EVOs failing early, from Samsung, king of SSD reliability...
 

eidairaman1

The Exiled Airman
Joined
Jul 2, 2007
Messages
42,639 (6.68/day)
Location
Republic of Texas (True Patriot)
System Name PCGOD
Processor AMD FX 8350@ 5.0GHz
Motherboard Asus TUF 990FX Sabertooth R2 2901 Bios
Cooling Scythe Ashura, 2×BitFenix 230mm Spectre Pro LED (Blue,Green), 2x BitFenix 140mm Spectre Pro LED
Memory 16 GB Gskill Ripjaws X 2133 (2400 OC, 10-10-12-20-20, 1T, 1.65V)
Video Card(s) AMD Radeon 290 Sapphire Vapor-X
Storage Samsung 840 Pro 256GB, WD Velociraptor 1TB
Display(s) NEC Multisync LCD 1700V (Display Port Adapter)
Case AeroCool Xpredator Evil Blue Edition
Audio Device(s) Creative Labs Sound Blaster ZxR
Power Supply Seasonic 1250 XM2 Series (XP3)
Mouse Roccat Kone XTD
Keyboard Roccat Ryos MK Pro
Software Windows 7 Pro 64
Yep, i will. First i have to copy everything off it, as much as i can anyway. The data is not super important (i have backups of super important data), but it's still very annoying.

Especially since i spent extra to get 870 EVOs over, say, an MX500. The 870 EVO is one of the most expensive SATA drives for the consumer market, and then this...

Plus it's not just about me, there's the much bigger issue at hand here. By the looks of it, we have a large batch of 870 EVOs failing early, from Samsung, king of SSD reliability...
I go pros myself. But im glad you do what most dont, back up!
 
Joined
Dec 11, 2019
Messages
286 (0.16/day)
Check your 870 EVO SSDs for these things:
Elevated "Reallocated Sector Count", "Used Reserve Block" and "Runtime Bad Block" count - first warning sign (my two other 870 EVOs have none).
Non-zero "Uncorrectable Error Count" and "ECC Error Rate", and especially if those two keep rising when you read/write files. Definitely affected then!


You can read out the self-monitoring SMART values of your drives with CrystalDiskInfo.
In there under "Function" -> Advanced Feature -> Raw Values, select "10 [DEC]" to have human-readable values. You can then upload screenshots of the data, press CTRL-S to screenshot in there.

Hmm, my drive has more power-on hours than yours but way less written and already 10 uncorrectable errors and it's using reserved blocks. Got it in March so nearly a year old, not sure when it was manufactured.

My other SSDs are showing no errors or reserved blocks used.

870 Evo - 2022-02-01.png
 
Joined
Jun 30, 2013
Messages
89 (0.02/day)
Hmm, my drive has more power-on hours than yours but way less written and already 10 uncorrectable errors and it's using reserved blocks. Got it in March so nearly a year old, not sure when it was manufactured.

My other SSDs are showing no errors or reserved blocks used.

Looks like the very early stages of one of these failures, yes. You wanna keep a close eye on the aforementioned values, and as soon as there's a file you can't copy/read anymore, backup all your data (in fact, you might wanna do that already) and RMA it. By the way, you could've used the integrated screenshot function (CTRL-S) and under "Function" there's also a setting to hide the serial number. Which i'd actually like to know (you can leave out the last four digits), cause it tells me roughly when this was produced. Looks like an early drive, exactly like most affected ones. The exact day it was produced is printed on the side of the packaging.
 
Joined
Dec 11, 2019
Messages
286 (0.16/day)
It's actually only an OS drive (ironically I chose the Samsung because I thought it would be the most reliable for this purpose :laugh: ), hence the low writes and probably the low read errors (since I don't really store much on the drive).

I found the box, it was manufactured in 2020-12-30. The firmware version: SVT01B6Q and serial number: S62BNJ0NC62.
 
Joined
Dec 11, 2019
Messages
286 (0.16/day)
Ok, it's one of the very first batches, they were only sold from 2021 on IIRC. So yeah, this one is highly suspicious.

Great, but thanks for the heads up. Maybe one of the hardware sites will pick this up and ask Samsung for an update. Would be nice to know if there's any way of heading it off or it's only RMA.
 
Joined
Jun 30, 2013
Messages
89 (0.02/day)
I would think the only solution is an RMA, because from the cases i've read about so far, Samsung are very quick to offer an RMA once you describe these characteristical symptoms.

I haven't actually tried to RMA mine, still busy with getting the files off there and then i'm gonna do a full scan with Samsung Magician etc., so i have more proof of the failure for the RMA.
Also, i kinda want to understand the issue more, especially since i have two more such drives.
 
Joined
Dec 11, 2019
Messages
286 (0.16/day)
I would think the only solution is an RMA, because from the cases i've read about so far, Samsung are very quick to offer an RMA once you describe these characteristical symptoms.

I haven't actually tried to RMA mine, still busy with getting the files off there and then i'm gonna do a full scan with Samsung Magician etc., so i have more proof of the failure for the RMA.
Also, i kinda want to understand the issue more, especially since i have two more such drives.

So, I installed Magician. The smart scans didn't complete and both had the same error "Failing LBA". The images are the results of the full scan, smart data and Crystal Disk, after running all of magician's tests.

SM - full scan - 2022-02-01 no SN.pngSM - smart - 2022-02-01.pngCDI - after magician - 2022-02-01.png
 
Joined
Jun 30, 2013
Messages
89 (0.02/day)
Yup there we have it, it's a full-blown failure. Get any important data off of it and RMA it.

I wonder how many ticking timebombs are out there. Most users don't know of this issue yet and the drives have only been sold for a year.

I also worry about my two other 870 EVOs. I don't know yet if only the early production dates are affected, or if this is a more widespread problem that only manifests after around half a year of use, but with almost all drives. Time will tell, i guess.

If it's more widespread, this will be a complete disaster not only for Samsung, but of course also for all the affected users if they don't have a backup. This problem will randomly corrupt some of the files and they can't be read anymore (at least using FreeFileSync, it will show a log afterwards of which files couldn't be copied). I think you're eventually gonna read everywhere about it if it's more widespread. Samsung's reputation for SSD reliability is on the line. I also worry about the replacement drive, on reddit i saw someone get (slightly) used drives as replacements from Samsung. Makes me want to write the vendor first and not Samsung, i don't want to get a used drive back. Also, even with a new drive, it's kinda hard to trust this model right now. Let's hope it really is just a problem of the early batches.
 
Last edited:
Joined
Dec 11, 2019
Messages
286 (0.16/day)
Yup there we have it, it's a full-blown failure. Get any important data off of it and RMA it.

I wonder how many ticking timebombs are out there. Most users don't know of this issue yet and the drives have only been sold for a year.

I also worry about my two other 870 EVOs. I don't know yet if only the early production dates are affected, or if this is a more widespread problem that only manifests after around half a year of use, but with almost all drives. Time will tell, i guess.

If it's more widespread, this will be a complete disaster not only for Samsung, but of course also for all the affected users if they don't have a backup. This problem will randomly corrupt some of the files and they can't be read anymore (at least using FreeFileSync, it will show a log afterwards of which files couldn't be copied). I think you're eventually gonna read everywhere about it if it's more widespread. Samsung's reputation for SSD reliability is on the line. I also worry about the replacement drive, on reddit i saw someone get (slightly) used drives as replacements from Samsung. Makes me want to write the vendor first and not Samsung, i don't want to get a used drive back. Also, even with a new drive, it's kinda hard to trust this model right now. Let's hope it really is just a problem of the early batches.

I'm surprised it's not being talked about already, if it was a widespread issue with the flash I'd think all Samsung drives using this flash would be failing (980 and 980 Pro are 128L iirc), which suggests it is a limited batch, or an issue with the controller? I noticed some people seemed to resolve the issue by formatting the drive and updating the firmware, but I wonder if they're just hiding it because a drive with no data doesn't produce errors when reading data.
 
Joined
Jun 30, 2013
Messages
89 (0.02/day)
Yes, i've read about that "trick" too (zero-filling the drive and suddenly the error count doesn't increase anymore when you do a full scan). It might be that the drive controller knows that it's just zeroes in the cells and doesn't bother to actually read them, or something to that effect. I wouldn't trust this method at all. The only thing that counts is what the drive does with the data that's on there, and evidently, not all files can be read out successfully anymore. I also have the newest firmware on my 870 EVOs.

To me, once this problem appears, the drive is dead to me and i will RMA it, there is no other way. Unless Samsung want to acknowledge the problem and deliver an official fix with a new firmware update. But i think that is still some way off, first they will silently replace the affected drives and hope it doesn't make the rounds on the big tech websites.

More reading material: https://blog.yuo.be/2021/12/23/the-ten-month-old-ssd-that-ate-my-data/
 

stinger608

Dedicated TPU Cruncher & Folder
Joined
Nov 11, 2008
Messages
11,232 (1.91/day)
Location
Wyoming
System Name Dean Machine/2020 Ryzenfall
Processor Intel 4790K/AMD Ryzen 5800X
Motherboard MSI 1150 Gaming mATX/Asus ROG STRIX B550-F Gaming Wifi
Cooling Cooler Master Hyper 212 LED/SilverStone AH240 AIO
Memory 16 gigs Crucial Ballistix Tactical Tracer/32 gigs G.Skill TridentZ NEO DDR4
Video Card(s) Gigabyte 1660 Super/EVGA GTX 1080 FTW Hybrid Gaming
Storage Crucial SSD 256 and 2TB spinner/Dual Samsung 980 Pro M2 NVME 4.0
Display(s) Overlord 27" 2560 x 1440
Case Corsair Air 540
Audio Device(s) On board
Power Supply Seasonic modular 850 watt Platinum/EVGA T2-850 Titanium
Software Windows 10 Pro/Windows 10 Pro
Holy crap, I may have to check my 980 pro. It's almost new and it was showing 97% last week. I've firing up that system today and I'm going to check.
 
Joined
Jun 30, 2013
Messages
89 (0.02/day)
I have no evidence of the 980 Pro being affected (and i also use one, from 2021-03). Plus it would be difficult to find out on the 980 Pro, unless you're already getting read errors in normal operation. Because the 980 Pro has a completely different set of SMART values, it doesn't report "Uncorrectable Error Count" or "ECC Error Rate", the main early warning signals for this failure. The health percentage, on the other hand, is not a good indicator for this failure. My affected 870 EVO still reports 99% health, despite multiple files having been unable to read from it, 329 reallocated sectors, and almost 5000 Uncorrectable Errors.

So instead of looking at the health percentage (which relies mostly on a conservative estimate from the manufacurer how many times the cells can be overwritten, i.e. the TBW value, and the real endurance is often 10x higher), i would suggest a full drive scan for bad sectors with Samsung Magician. This will make sure all cells can be properly read out. If that's the case, then there's no need to worry.
 
Joined
Dec 5, 2013
Messages
640 (0.16/day)
Location
UK
Is this just an 870 thing or are the 970 EVO / 970 EVO Plus's affected too?
 
Joined
Jun 30, 2013
Messages
89 (0.02/day)
I would say more of a 870-EVO-only thing, since it uses the more modern NAND flash that is also used on the 980 Pro. The 970s are on a bit older NAND, plus they've also been around for much longer. So if there was a similar problem with them, you would've definitely heard about it already.
 
Joined
Aug 20, 2007
Messages
21,546 (3.40/day)
System Name Pioneer
Processor Ryzen R9 9950X
Motherboard GIGABYTE Aorus Elite X670 AX
Cooling Noctua NH-D15 + A whole lotta Sunon and Corsair Maglev blower fans...
Memory 64GB (4x 16GB) G.Skill Flare X5 @ DDR5-6000 CL30
Video Card(s) XFX RX 7900 XTX Speedster Merc 310
Storage Intel 5800X Optane 800GB boot, +2x Crucial P5 Plus 2TB PCIe 4.0 NVMe SSDs
Display(s) 55" LG 55" B9 OLED 4K Display
Case Thermaltake Core X31
Audio Device(s) TOSLINK->Schiit Modi MB->Asgard 2 DAC Amp->AKG Pro K712 Headphones or HDMI->B9 OLED
Power Supply FSP Hydro Ti Pro 850W
Mouse Logitech G305 Lightspeed Wireless
Keyboard WASD Code v3 with Cherry Green keyswitches + PBT DS keycaps
Software Gentoo Linux x64 / Windows 11 Enterprise IoT 2024
Hmm, my drive has more power-on hours than yours but way less written and already 10 uncorrectable errors and it's using reserved blocks. Got it in March so nearly a year old, not sure when it was manufactured.

My other SSDs are showing no errors or reserved blocks used.

View attachment 234812
oof. SSDs should not be showing uncorrectable errors until late in their life stages... thats not a good sign.
 

eidairaman1

The Exiled Airman
Joined
Jul 2, 2007
Messages
42,639 (6.68/day)
Location
Republic of Texas (True Patriot)
System Name PCGOD
Processor AMD FX 8350@ 5.0GHz
Motherboard Asus TUF 990FX Sabertooth R2 2901 Bios
Cooling Scythe Ashura, 2×BitFenix 230mm Spectre Pro LED (Blue,Green), 2x BitFenix 140mm Spectre Pro LED
Memory 16 GB Gskill Ripjaws X 2133 (2400 OC, 10-10-12-20-20, 1T, 1.65V)
Video Card(s) AMD Radeon 290 Sapphire Vapor-X
Storage Samsung 840 Pro 256GB, WD Velociraptor 1TB
Display(s) NEC Multisync LCD 1700V (Display Port Adapter)
Case AeroCool Xpredator Evil Blue Edition
Audio Device(s) Creative Labs Sound Blaster ZxR
Power Supply Seasonic 1250 XM2 Series (XP3)
Mouse Roccat Kone XTD
Keyboard Roccat Ryos MK Pro
Software Windows 7 Pro 64
Joined
Oct 15, 2011
Messages
2,485 (0.52/day)
Location
Springfield, Vermont
System Name KHR-1
Processor Ryzen 9 5900X
Motherboard ASRock B550 PG Velocita (UEFI-BIOS P3.40)
Memory 32 GB G.Skill RipJawsV F4-3200C16D-32GVR
Video Card(s) Sparkle Titan Arc A770 16 GB
Storage Western Digital Black SN850 1 TB NVMe SSD
Display(s) Alienware AW3423DWF OLED-ASRock PG27Q15R2A (backup)
Case Corsair 275R
Audio Device(s) Technics SA-EX140 receiver with Polk VT60 speakers
Power Supply eVGA Supernova G3 750W
Mouse Logitech G Pro (Hero)
Software Windows 11 Pro x64 23H2
This reminds me of Maxtor Diamond Max 8 HDDs, (the slim models, the ones that look skinnier than the usual 3.5-inch sizing) which are failure prone. (The ones that are mainly model code 6Exxxxx)

While Diamond Max 9, IIRC, before Seagate bought Maxtor, were reliable. (ones with model code 6Yxxxxx were solid, OTOH)
 
Joined
Feb 1, 2019
Messages
3,667 (1.70/day)
Location
UK, Midlands
System Name Main PC
Processor 13700k
Motherboard Asrock Z690 Steel Legend D4 - Bios 13.02
Cooling Noctua NH-D15S
Memory 32 Gig 3200CL14
Video Card(s) 4080 RTX SUPER FE 16G
Storage 1TB 980 PRO, 2TB SN850X, 2TB DC P4600, 1TB 860 EVO, 2x 3TB WD Red, 2x 4TB WD Red
Display(s) LG 27GL850
Case Fractal Define R4
Audio Device(s) Soundblaster AE-9
Power Supply Antec HCG 750 Gold
Software Windows 10 21H2 LTSC
I am cursed, the drive I brought to replace the dodgy MX500 in my laptop? You guessed it, is a 870 EVO.

Why did the guy on reddit delete himself after posting this though?
 
Last edited:
Joined
Jun 30, 2013
Messages
89 (0.02/day)
Hmm, you're right, he did. Weird. But i don't think it has anything to do with this 870 EVO issue.

About that, it's enough to keep a close eye on your Samsung (and make backups of your data), no need to replace it out of panic or anything. It seems that primarily (or only?) drives from a certain batch are affected, early in the production. If you have a more recent drive, they are probably ok. I mean, i have three (yes three) 870 EVO 4TB, and only the drive with an early production date gives me problems, i will RMA it eventually. The other two drives are completely normal, no read errors and SMART data looks good.
 
Top