Frank_100It is hard to believe it is an accident with the MS"s resources.

Well, when you fire all the people who should be doing the internal dogfooding, this is what happens.

At this point, nothing short of Microsoft losing a MASSIVE lawsuit over companies (and users) losing terabytes of data will make them change their mind when it comes to outsourcing their testing onto their own users.

Someone... go find the most sleaziest lawyer you can find.

YouseethisWell, rather than saying that the SSD manufacturer is to blame, it seems like this mysterious explosion is the result of Windows messing around with the cache specifications.

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

Does someone here knows if it makes any sense?

Lol, no.
The values are implemented as the NVME spec defines, if vendors deviate from the spec in ways that cause hardware issues its on the vendors heads.

Frank_100If the hardware works under windows 10 and Linux it most certainly is a MS issue.

The hardware fails on Linux too, each of these drives have workarounds in the linux kernel to avoid faults in their firmware/controllers.

It's the vendors fault for playing loose with the nvme spec.

Several of the listed nvme's could be protect mode crashed using their own toolkits to perform tests from day one.

reiterating, there hasn't been a driver change for the nvme stack in the latest update, its dated current (for packaging regions) but its still based on 26100.4484 from June.

I hope this is a tipping point for a lot of people to switch to Linux, or failing that, Mac.

Are these the only two microsoft issues? Not more?

'KB5063878' and 'KB5062660'

Is it only microsoft related problem? yes?

How comes I read it for microsoft only? How comes I did not read anything else for the other use cases?

seems this company related ucontroller for the flash
www.techpowerup.com/ssd-specs/corsair-mp600-pro-1-tb.d366
i did not found my 2tb drive
www.techpowerup.com/ssd-specs/kingston-kc3000-1-tb.d259

What are for those low quality drives where manufacturer did not provide any updates last time i checked in windows? Where the low quality manufacturer only offers windows only upgrade software?
e.g.
www.techpowerup.com/ssd-specs/corsair-mp600-pro-1-tb.d366

to make it clear other use cases in my point of view may be the linux kernel with any userspace. That includes also embedded linux. Freebsd... What else?

I think it is a heavily windows only issue with FLASH based drives.
There are for sure quite more problematic Storage based, especially flash storage based issues.

'KB5063878' and 'KB5062660'

SoraThe hardware fails on Linux too, each of these drives have workarounds in the linux kernel to avoid faults in their firmware/controllers.

Maybe the mp 600 pro 1TB is not affected.
AFAIK what i remember when i bought it by a computerbase.de forum recommendation topic it uses phision controller

I wrote all that data in one rush from a backup around may 2025. I changed mainboard and therefore I wanted to rule out an issue with filesystem compression and size of /
I wrote over 100GB in one go from the backup.
+ w11 pro installation before which i aborted as i realised it is not worth it to reinstall.

Sienna_Cichlid /home/roman # df -h /
Filesystem Size Used Avail Use% Mounted on
/dev/dm-3 127G 103G 23G 82% /

Feel free to provide data and proof that the linux kernel is affected by this topic.

you can look up my mainboard MSI x670E topic. that was the point when i moved my / from KC3000 2TB to mp 600 pro 1TB. should be around may 2025.

i need to move my data back anway and reduce / on my kc3000. 350GB is far too big. 127GB was a bit too small. ... It does not matter as I need to verify regular if backups can be restored. backups need to be tested. backup instructions need to be improved.

Hecate91I have a 980 pro 2TB

I assume it is a samsung 980 pro drive. Samsung had a lot of data loss cases in the past. Are you sure you are worried about your data?

This update also has killed WD red, and seagate exos ENTERPRISE HARD DRIVES!!

Jayztwocents youtube vid about this comments section had 2 comments saying hard drives were even effected

Thaaaa plot thickinz
wccftech.com/phison-statment-false-documents-sent-to-customers-claiming-windows-11-ssd-breaking-update-specific-to-their-controllers-legal-action/?fbclid=IwY2xjawMSlydleHRuA2FlbQIxMQABHh5fj7_Z-UaV0ulmBEmezPq1vBeIcDh6GD59WriAsZI7eHpDcxnc0dqSXdSl_aem_wLAZc4Vuhs3PjJ8MKs0Ujw

Oh boi and I hope phison sues tf outta Microsoft!!!

So, there has been a response from Phison on this matter. Now, has there been anything specific from Microsoft?

720p lowSo, there has been a response from Phison on this matter. Now, has there been anything specific from Microsoft?

Even more WD drives affected: (and they use their own controllers)

PhilaphlousI'm mildly worried about this... I can't rollback my update and so I'm just sitting here waiting for data corruption... SK Hynix P31 Gold SSD... Guess I should backup my new data? Worried a large read/write is going to screw things up...

Use clonezilla. Does not involve microsoft.

I just increased the number of full disk images that I keep as a backup from three to seven. Every week a full disk image is made of my OS and secondary SSD drives using Macrium Reflect. Hopefully I won't have to fall back to them before the next dreaded Patch Tuesday.

noel_fsinexistent QA, i dont even know what they have insider for

Exactly my thoughts. There are literally millions on Win Insider machines, and no complains so far??

Prima.VeraExactly my thoughts. There are literally millions on Win Insider machines, and no complains so far??

I'm honestly of the opinion that most people don't actually run Insider Preview builds on live hardware or in real-life. I've ran Preview Builds before but only in Hyper-V Virtual Machines.

With that being said, I seriously doubt the Insider Preview program is of any real use to Microsoft because no one in their right mind would run that on a live system. That's just stupid. And since this glaring bug slipped, it lends more credence to my theories.

Honestly I think that Microsoft needs to just scrap the whole idea of the Insider Preview and get back to doing actual internal testing. I don't care if the cost of a license of Windows goes up 200%, I just want a stable version of Windows that we all can depend upon.

Will that happen? *snickers* F*ck no. To go back on all of this would admit that they were wrong and shareholders won't like that.

SoraThe hardware fails on Linux too, each of these drives have workarounds in the linux kernel to avoid faults in their firmware/controllers.

Yep. It's infuriating that I need to throw a bunch of kernel flags at the problem because the 990 Pro's randomly drop from my server until it is powered off and turned back on again, if I recall they are all related to power management. [ICODE]pcie_aspm=off pcie_port_pm=off amd_iommu=fullflush nvme_core.default_ps_max_latency_us=0[/ICODE]
I had the same problem with 980 Pro's too, even with the latest firmware from June it still happened until I put those flags in place.

trparkyI just increased the number of full disk images that I keep as a backup from three to seven. Every week a full disk image is made of my OS and secondary SSD drives using Macrium Reflect. Hopefully I won't have to fall back to them before the next dreaded Patch Tuesday.

I'm gonna get a huge HDD prolly seagate exos or WD ultrastar and have backups hosted on that.

Was planning on getting those drives on my birthday in feb but that's gonna be lookin more like a couple months now.


And have backup ssd drives too

And another cloning dock.

Oyyyyyy

sudothelinuxwizardI hope this is a tipping point for a lot of people to switch to Linux, or failing that, Mac.

I am considering it already, will be dual booting Linux and 11. Steam have brought Linux into the limelight for desktop use. 11 is a mess even without that dodgy update.

ShadowizedYep. It's infuriating that I need to throw a bunch of kernel flags at the problem because the 990 Pro's randomly drop from my server until it is powered off and turned back on again, if I recall they are all related to power management. [ICODE]pcie_aspm=off pcie_port_pm=off amd_iommu=fullflush nvme_core.default_ps_max_latency_us=0[/ICODE]

PCIE ASPM bus resets (WD affected mainly, strangely resolved with Native ASPM used instead of BIOS ASPM)
Thermal throttling that IO locks the drive until rebooted
HBM logical remap corruption when bursting
Nand cell refresh cycling during Active times instead of Idle

Amazing what issues arise when controllers based on draft nvme specs don't match the final spec....

Frank_100If the hardware works under windows 10 and Linux it most certainly is a MS issue.

It is hard to believe it is an accident with the MS"s resources.

It's a Phison issue.

JermelescuIt's a Phison issue.

No it's not. Even wd and samsung drives are affected and they use their own controllers!!!

chrcolukI am considering it already, will be dual booting Linux and 11. Steam have brought Linux into the limelight for desktop use. 11 is a mess even without that dodgy update.

I was going to go Linux too this year too but two things got in my way.
1) I failed to learn how to create a 4 drive striped array that I could easily move between computers in case of motherboard failure. (thanks to my temporary Optane addition)
2) Multi-monitor support for VMWare workstation is just better on Windows.

remixedcatNo it's not. Even wd and samsung drives are affected and they use their own controllers!!!

Are you sure? Vendors have been known to swap parts so it's possible they may have started with their own controller and did a switch during manufacturing to an alternative probably because of cost.

A Computer GuyI was going to go Linux too this year too but two things got in my way.
1) I failed to learn how to create a 4 drive striped array that I could easily move between computers in case of motherboard failure. (thanks to my temporary Optane addition)
2) Multi-monitor support for VMWare workstation is just better on Windows.


Are you sure? Vendors have been known to swap parts so it's possible they may have started with their own controller and did a switch during manufacturing to an alternative probably because of cost.

Samsung and wd have always used their own controllers so nothing has changed.

remixedcatSamsung and wd have always used their own controllers so nothing has changed.

I get what your saying although in the case of WD they do in fact use controllers from other venders for some models anyway.
www.techpowerup.com/ssd-specs/?q=western+digital

Along with this SSD issue I recently found 23H2/22H2 recovery features are broken now too. As a Windows user this really sucks. I already had to repair my recovery partitions from that last botched update that couldn't bother to resize the recovery volume properly.
www.forbes.com/sites/zakdoffman/2025/08/21/microsoft-confirms-windows-is-broken-emergency-update-on-way/

remixedcatNo it's not. Even wd and samsung drives are affected and they use their own controllers!!!

I have not seen any mentions of Samsung being effected and I personally haven't had a single issue myself with those updates installed. Would be curious to see what exactly M$ did to cause such an issue though.

ShadowizedI have not seen any mentions of Samsung being effected and I personally haven't had a single issue myself with those updates installed. Would be curious to see what exactly M$ did to cause such an issue though.

I have seen a comment on Reddit where someone complained that a Samsung 970 Evo Plus, a DRAM-less SSD, suffered this problem.

hardware/comments/1mtcgkg/comment/n9tax09
I am starting to wonder if the thing that is fouled up is whatever Windows uses to manage host memory buffers that DRAM-less SSDs depend on. I understand the need to cut power consumption and costs by not having a DRAM that requires constant refreshes, but that makes the SSD vulnerable to operating system mismanagement. With a hardware DRAM cache, the SSD can be much more self-contained and less vulnerable to operating system or hypervisor bugs.

jnv11I have seen a comment on Reddit where someone complained that a Samsung 970 Evo Plus, a DRAM-less SSD, suffered this problem.

hardware/comments/1mtcgkg/comment/n9tax09
I am starting to wonder if the thing that is fouled up is whatever Windows uses to manage host memory buffers that DRAM-less SSDs depend on. I understand the need to cut power consumption and costs by not having a DRAM that requires constant refreshes, but that makes the SSD vulnerable to operating system mismanagement. With a hardware DRAM cache, the SSD can be much more self-contained and less vulnerable to operating system or hypervisor bugs.

There's plenty of sn850x reports too and those have dram

remixedcatThere's plenty of sn850x reports too and those have dram

Are there reports on other DRAM-based drives falling victim to the Windows patch? I have seen some reports about SN850X and/or other WD/Sandisk DRAM-based drives bugging out in other situations as seen below, so I have to wonder if that could be WD/Sandisk's fault:

hardware/comments/1mtcgkg/comment/n9czehm

jnv11Are there reports on other DRAM-based drives falling victim to the Windows patch? I have seen some reports about SN850X and/or other WD/SanDisk DRAM-based drives bugging out in other situations as seen below:

I'm not keeping track of every single drive and every single report but I know it's affecting some others that's just one example that's popping up a lot in addition to 990 pros which I belive have dram and their own controllers just like wd/SD does. I didn't screenshot every single instance, nor should I be expected to.

All I am saying is that this is absolutely a Microsoft issue, not a "phison issue" "hmb issue" , "drive issue", or anything, the blame is on Microsoft. All those drives were a ok till the update, even those w hmb/DRAMless.

And this update shouldn't have even touched these drives. Microsoft managed to clog the toilet while overcooking dinner.