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

[tabascosauz]

that article seems really weird and written by someone killing their own drive

View attachment 280756


THEY TRIED TO WARRANTY BECAUSE IT DROPPED TO 99%


They also think that the drive would reset to 100% if it was "reset" which... what?
View attachment 280757


The author of this has no idea what they're talking about. For all we know they're runinng XP without TRIM and torrenting to the drive - we have no idea.

In the paragraph immediately following what you quoted, the author says that it's down to 95% a few days later, then a Crystaldiskinfo screenshot showing 94% at 4TB written, at which point they decided to RMA......4TB out of 1200TB is not 6%, even if the life indicator was solely measured on TBW (it isn't)

 

[E.S]

My concern is that the 990 story will hijack the 870evo story and forget about the 870evo issue.

 

[El Broki]

Adding to the list. After almost 2 years my OS drive is causing BSOD and has bad blocks. Recovery always fails. Last time I buy a samsung ssd.

 

[80251]

I read here about a trick to bring SATA SSD's back to life by disconnecting them from their SATA connector then powering up the system and letting it sit for a while. Would this idea make any diff. with the 870 EVO issues?

 

[Eddie58]

I read here about a trick to bring SATA SSD's back to life by disconnecting them from their SATA connector then powering up the system and letting it sit for a while. Would this idea make any diff. with the 870 EVO issues?

No, disconnecting a SATA connector cannot recover bad blocks.

 

[Mussels]

Crystaldiskinfo when it printed a mangled drive name and readings on the 870

Dont forget random errors from the PC itself can mess with these sometimes - something as simple as the SATA driver being updated on that clean install while the program was still running can corrupt some of the results, and it wont refresh them fully til a reboot

 

[SPDIF]

This is my 870 EVO SSD, fingers crossed all good until now. Used every day, FC number does not change and stay like this.

 

[80251]

@SPDIF
What is the date of manufacture for your 870 EVO?

 

[chrcoluk]

Even though my unit is still ok, its bugging me as I feel its about to break.

Not having the accelerated wear levelling that the MX500 has is nice though.

Here is my current stats (only important one's), note my laptop is no longer on 24/7.

2375 power on hours
964GB host writes.
2 Wear levelling count.
0 for ECC/CRC/Runtime bad block/Erase fail count/Reallocated sectors/Program fail count.

Firmware is SVT01B6Q @tpudine my drive shipped with this firmware.

that article seems really weird and written by someone killing their own drive

View attachment 280756


THEY TRIED TO WARRANTY BECAUSE IT DROPPED TO 99%


They also think that the drive would reset to 100% if it was "reset" which... what?
View attachment 280757


The author of this has no idea what they're talking about. For all we know they're runinng XP without TRIM and torrenting to the drive - we have no idea.

Yeah crystal diskinfo needs to dump that XX% health, just keep the good/bad but not the %.

Samsung is currently investigating the 990 issue.
Why do you keep ignoring the 870evo and 980 issue?

Because the 990 got a news article from toms hardware (article seems to not understand properly whats going on, NVME no longer reports erase cycles, so the gradual decline in health wont occur, instead it only goes down when is reallocated blocks which are called available spare on SMART), I have not seen any tech site report on this, I think TPU should consider we have this thread on their site. We know consistently now these companies only care when its a media outlet reporting a problem.

--

Apparently 980 PRO also affected? Wow whats going on samsung, 980 PRO my main drive in my PC. Thankfully still no spare blocks used (this is the metric that is reducing the life reported).

Update -- On the 990 PRO nas article, I found this, will quote below. Luckily I have the 5B2QGXA7 firmware, so should be in the clear.

It is important to note that 980 Pro SSDs running the 4B2QGXA7 or 5B2QGXA7 firmware are not affected by the issue and do not require the update. However, users who are running the 3B2QGXA7 firmware should update to the newest firmware as soon as possible to avoid any potential data loss or failure of the SSD.

 

[SPDIF]

@SPDIF
What is the date of manufacture for your 870 EVO?

Date is; 2020/12

 

[pathtonowhere]

I've got about 12-13 4TB 870 EVOS. I've got two 2021.01 and two 2021.10. One 2021.01 corrupted my files. I've emptied it, fixed the bad sectors (I thought they were logical) then put files on it again (backup though) and now it has bad sectors again. I'm scanning the other 2021.01 right now, it seems fine, but it only has 2 power cycles despite being full. After the scan, I'm going to try to copy the files from it. I hope I'll be able to.

The 2021.10 also passed the bad sector scan, but I haven't tried to copy files from them, which I will do soon.

Are the September 2021 drives also affected? Are there any reports on this?

Also, I'm from Romania, I don't have any papers (warranty or receipt). Will Samsung service accept my damaged drives to warranty?

Honestly, I'm extremely disappointed, I've sold all my HDDs to replace them with expensive EVOs (faster, cooler, quieter, smaller, lighter) but now I regret that I haven't kept the HDDs, too. I have redudancy on my EVOs on older drives, too, but now I can't trust Samsung SSDs anymore, I once thought they were infallible.

 

[ThrashZone]

Yeah crystal diskinfo needs to dump that XX% health, just keep the good/bad but not the %.

Hi,
The life percentage can be compared to other utilities and hwinfo64 is one that is commonly used for other readings.
So ditching a reading is foolish if it matches other utilities findings

Maybe just do what samsung and other bad ssd makers would like and that is stop checking your ssd info and just wait for the ssd to die of natural causes and just rma when it does

 

[chrcoluk]

Hi,
The life percentage can be compared to other utilities and hwinfo64 is one that is commonly used for other readings.
So ditching a reading is foolish if it matches other utilities findings

Maybe just do what samsung and other bad ssd makers would like and that is stop checking your ssd info and just wait for the ssd to die of natural causes and just rma when it does

There is nothing to warrant people should stop checking the health of their drives, however that shouldnt mean a software dev adds a counter that can mislead people is a good idea.

I have looked into crystal diskinfo and have figured out how it works as an example.

It will check erase cycles counter (if available, not available on NVME), and use that to show a lowering of health, this is the first flaw. It will also keep a historical record of mapped reserve blocks, if they get mapped this will also lower the counter, but interestingly if you wipe the installation and reinstall, even with the remapped blocks showing on smart, it will jump back up to 100%. This second one is whats causing the 990 pro reported drop off.

if hwinfo64 also generates a artificial life % left, then I also think that should be ditched as well, but cant say I have noticed it, if it has it, its not big enough to get my attention.

Keep the readings, those who understand them use them, those who dont wait for drive to die or rely on vendor tool to alert you when it thinks the drive is in a critical state.

If the % is kept I would only make it change on reserve block mappings, ignore the erase cycles.

 

[ThrashZone]

There is nothing to warrant people should stop checking the health of their drives, however that shouldnt mean a software dev adds a counter that can mislead people is a good idea.

I have looked into crystal diskinfo and have figured out how it works as an example.

It will check erase cycles counter (if available, not available on NVME), and use that to show a lowering of health, this is the first flaw. It will also keep a historical record of mapped reserve blocks, if they get mapped this will also lower the counter, but interestingly if you wipe the installation and reinstall, even with the remapped blocks showing on smart, it will jump back up to 100%. This second one is whats causing the 990 pro reported drop off.

if hwinfo64 also generates a artificial life % left, then I also think that should be ditched as well, but cant say I have noticed it, if it has it, its not big enough to get my attention.

Keep the readings, those who understand them use them, those who dont wait for drive to die or rely on vendor tool to alert you when it thinks the drive is in a critical state.

If the % is kept I would only make it change on reserve block mappings, ignore the erase cycles.

Hi,
Frankly I believe sammy and the likes should stop screwing up the firmware instead.

 

[pathtonowhere]

One question I have for those who kept better track of how it happened to their drives. Does the problem appear as soon as the drive is full and you try recovering data from it, or does it take some time for it to start acting up? I'm asking because now I only have one clearly defective drive, one of the two 2021.01. The other seems to still work perfectly, but it has been used a lot less than the other and kept mostly unplugged. The two 2021.10 also appear to be both perfect. I so hope that at least the 2021.10 ones will survive. They cost me a lot of money and without a receipt I risk having my warranty refused by Samsung.

 

[uvdu1]

One question I have for those who kept better track of how it happened to their drives. Does the problem appear as soon as the drive is full and you try recovering data from it, or does it take some time for it to start acting up? I'm asking because now I only have one clearly defective drive, one of the two 2021.01. The other seems to still work perfectly, but it has been used a lot less than the other and kept mostly unplugged. The two 2021.10 also appear to be both perfect. I so hope that at least the 2021.10 ones will survive. They cost me a lot of money and without a receipt I risk having my warranty refused by Samsung.

I haven't been checking SMART data, but I had a 870 2TB do a complete failure suddenly, 100% unusable and all data seems to be lost - before that I was using it perfectly normal and never had any write/read errors. Mines from 2021.01.
Definitely don't depend on them - any data you have on them could disappear.

EDIT: Looking at Event viewer, tons of logs about bad blocks on that disk. going a month back.

 

[SPDIF]

Seems like Samsung F*ck*d up things in the beginning of 2021. Maybe all the controllers have Corona it seems.
Now also 980 and 990 Pro, there is an update to stop your SSD come bad, but what is lost stays lost.
It's strange they have the same sickness as the 870 EVo, a rapidly declining health.

Just bought a new 990 Pro 2TB, i will let it stay sit in the box until i know more about what happened. There is a firmware update, but i wait until Samsung open their mouth.
No words about what is fixed, and they still selling them every where.
For anyone here that has a 990 Pro, immediately update your SSD to firmware version 1B2QJXD7 from old version 0B2QJXD7!

Samsung 990 Pro Firmware Update Addresses Failing SSD Health

The update fixes the underlying cause of the rapid health declines, but it's not all good news.

www.tomshardware.com

My 870 EVO does not seems affected, but if it goes bad i have a backup from everything always.

It's time Samsung takes there responsibility and let everyone know what happened there.
They should let it know on there webpage that they are wrong, and give people a choice to RMA it.

But Samsung persists as there noose are bleeding...

 

[80251]

First the 870 EVO and now the 980 Pro and 990 Pro. Maybe Samsung is introducing some planned obsolescence into their parts, after all, if their SSD's die after the warranty has expired they can sell even more!

 

[H3LLGHA5T]

I Have two 870 EVOs 2 TB and both of them are showing this behavior, they do work as of now, but who knows for how long. One of them has 15 TB written, the other just 7. I also have a 980 Pro 2 TB, but I updated the firmware when I bought it a year ago and I couldn't see any unusual degradation thankfully. Sigh now I have to go through support, I just hope it doesn't suck. Samsung definitely lost a customer over this one.

 

[SPDIF]

Yep you are out of luck it seems, they are degrading and soon they will be in an Read-Only state.
Hope Samsung can shed some more light on this. Is it coincidence that the same thing happens with the 890 and 990 PRO series?
This will be an interesting story.

My 870 EVO is from 12/2020 and has no errors at moment. All others from a later date are affected, also 500Gb and 1Tb series.

 

[pathtonowhere]

What is that later date? Are there any defective 870 EVOs from 10.2021 or later?

Is there a firmware update fix for the 870s? Has anyone tried it?

And yes, being firmware related and the damage occuring after some usage, it makes you think they were just testing some kind of planned obsolescence and it went wrong, damaging drives way earlier than they intended. As long as they don't say anything about this, I'm allowed think the worst. Or maybe they were sabotaged from the inside and they're ashamed to admit it.

 

[Bwaze]

How come TechPowerUP is largely ignoring this fiasco in their news coverage? They don't even publish the good news, like firmware fix for 990 Pro SSD:

Samsung 990 Pro Firmware Update Addresses Failing SSD Health

Samsung is the most popular SSD provider, and the products with problems are their most widely sold - it's not like it's about some obscure stuff.

Do we really have to speculate if they are paid to keep the bad news out of sight?

 

[Count von Schwalbe]

Do we really have to speculate if they are paid to keep the bad news out of sight?

Samsung 990 PRO Flagship SSD Has an Endurance Problem, Users Notice Rapid Drive-Health Drops

Samsung 990 PRO is the company's flagship client SSD, which is among the fastest Gen 4 NVMe SSDs you can buy. It also commands a very high price premium, with the 1 TB variant priced at $170, and the 2 TB variant at $290. When you're buying in this segment, you expect the highest endurance...

www.techpowerup.com

 

[Bwaze]

Yeah, I know, but that's one news article about 990 Pro issues, almost a month old. 980 Pro wasn't mentiobed, 870 EVO didn't get it's article, even Samsungs recommendations for 980 Pro users and now for 990 Pro users to update their firmware are not published.

 

[P4-630]

Samsung 990 PRO Flagship SSD Has an Endurance Problem, Users Notice Rapid Drive-Health Drops

Samsung 990 PRO is the company's flagship client SSD, which is among the fastest Gen 4 NVMe SSDs you can buy. It also commands a very high price premium, with the 1 TB variant priced at $170, and the 2 TB variant at $290. When you're buying in this segment, you expect the highest endurance...

www.techpowerup.com

There's a fix out, new firmware for those 990's

Samsung 990 Pro Firmware Update Addresses Failing SSD Health

The update fixes the underlying cause of the rapid health declines, but it's not all good news.

www.tomshardware.com