[German]In Windows 10 20H2 with installed cumulative update KB4592438, chkdsk causes massive issues. It destroys the file system during a disk check on SSDs, so Windows 10 can't start after a reboot. Here is some information about the problem and the affected Windows 10 build. Addenum: Microsoft has fixed that bug.
I've been contacted by several German blog readers by email and via comments within my German blog (thanks for that). They all pointing to a strange bug, obviously brought to systems with cumulative update KB4592438.
The problem with ChkDsk
The problem was initially described by Nero24 in the German planet3dnow.de forum (the affected person contacted me via email). In a school, several systems with Windows 10 20H2 already installed should be manually updates to the current update status (Windows and Office updates). The administrator, who did the manual update install, decided to execute the following command after finishing the update installation:
chkdsk c: /f
The intension is to check the Windows drive for damage and (with the /f switch) immediately perform a repair. In the scenario described above, things went very badly wrong. The affected person had seven systems, that can no longer boot after the file system check including repair. The guy was able to prevent the re-boot from the other computers, after the seventh machine failed.
Stop error NTFS File System, Source: planet3dnow.de-Forum
The blue screen shown in the screenshot above appears with the stop code NTFS File System (it shows a NTFS FILE SYSTEM stop code, see here for an English version of the BSOD). The affected person was able to observe this on seven computers with Windows 10 20H2. When analyzing the affected disks (SSDs) on a working system, it could be determined that the logical Windows drive on the disk was detected only as a RAW partition. The /f option of chkdsk probably destroyed the NTFS file system.
Further analysis of the RAW partition with chkdsk in offline mode revealed errors with a corrupted 'file 9' and an error in the BITMAP attribute of the Master File Table. These could be corrected with chkdsk in offline mode. After that, the systems booted again after installing the SSD as a Windows disk in the original PCs.
Error analysis and condition
In the planet3dnow.de forum, the affected person gives some characteristics of the affected devices. Here is the relevant excerpt:
They are all Kaveri systems with ASUS A68HM-PLUS motherboard and Kingston SATA SSD running in AHCI mode with the Windows standard AHCI driver. At first I suspected a faulty TRIM behavior in combination with chkdsk and the fact that the system reboots immediately after completion.
But they are not identical SSDs. Most are Kingston A400s with Phison S11 controllers and 2D TLC NANDs, but there was also a Kingston V300 with Sandforce controller and MLC NANDs, which certainly has a completely different TRIM behavior than the Phison-fired ones.
Whether the hardware has an influence is currently unclear. Within my German blog a user reported, that the error didn't occur on two systems with hard disks, but a 3rd system with a M2.SSD caused the BSOD as expected. The case was summarized on German site planet3dnow.de in this article. There, the affected person writes that it affects build 19042.685 of Windows 10 20H2. The cumulative update KB4592438 should be the culprit. The blog reader wrote to me:
Have you heard anything similar from your community in this regard? In the forum thread above, we were able to narrow it down quite well where it occurs and where not, but I would still be interested in external observations.
After publishing the German edition of this blog post, several blog readers confirmed this bug. If you have further insights, please drop a comment. I will try to forward a link to this blog post to Microsoft's developers.
Addendum: Microsoft has fixed the bug – details may be found within my blog post Windows 10 2004/20H2: Microsoft fixes chkdsk issue in update KB4592438.
Cookies helps to fund this blog: Cookie settings
I've experienced a (potentially) related problem that may be of interest, which I've already reported to Microsoft.
Following the 20H2 update, I was experiencing a 0x77 (KERNEL_STACK_INPAGE_ERROR) BSOD with a frequency of ~2 times per week. After running through a series of failed fixes and problem identification tests (not fixed on a reboot, virus/malware scan finds nothing, memtest finds nothing wrong with the RAM, sfc /scannow fails to correct, and perhaps something I've forgotten?) I ran a chkdsk /f /r. On completion of this, I entered a 0x0000001 boot loop which no automatic repair could correct.
Rolling back Windows Update provided a temporary solution, though a Windows reset was eventually required. In the intervening period, a S.M.A.R.T analysis showed nothing of note regarding the HD. Unfortunately, I cannot provide an event log as they were lost with the reset.
Device affected is a laptop with ASUS S500CA motherboard, boot disk SanDisk U100 24GB running standard Windows AHCI. For whatever it's worth, data drive is Hitachi HTS545050A7E380.
Feel free to contact me at the posting email if additional information would be helpful to your analysis and consolidation.
Seems my computer just did an automatic update and now i have a cycling blue screen. Windows will not load.
Also running the 20H2 release and as soon as KB4592438 I started having issues. First steam starts crashing to desktop, then after the 3rd I got my first BSOD (page error non paging area). I ran the same chkdisk command and it seemed to find a lot of errors, going through stages 1-5 on the nvme. Then back into the desktop there's a notification from security & maintenance that my disc needs checking. It reboots, checking goes much faster just stages 1-2 and very fast which then leads me to a 'pc had an issue/windowd setup kind of screen' it rebooted, checked disk but presented no issues this time and back into desktop the warning was gone. I have uninstalled this KB and hopefully all is well again. No changes to hardware, all new pc hardware which went though 72hr burn-ins as well as further tests.
It I would have know that …. It would have saved me from a full reinstallation. Luckily I'm doing weekly backups.
It is like a fire extinguisher (chkdsk) filled with fuel.
Sounds like more german bs. lol
I HAVE HAD THE SAME ISSUE AND IT COST ME ALL OF MY APPS & FILES. VERY UPSET WITH MIROSOFT.
Connect it as an external or secondary drive to another computer and see if you can still access the files. With mine I could.
It did happen tp me, but thought it was my pc (I do lots of Multiboot Linus & Windows).
I was able to fix it by booting an older version of Windows 10 on USB, then ran Chkdsk /f on Damaged Partition, and all worked out fine.
I had the same thing happen with chkdsk /f. Running chkdsk again (from recovery, since Windows was unbootable) fixed it.
I believe that I have been affected by the error also.
A not too old Asus with a SSD, that my kid use, and I keenly update.
About 10 days ago it went into a bsod cycle upon boot. I can see that the symptoms are right. A complete reinstall was nessecarry.
I have not seen the error since, but I remember to have installed optional driver updates just before the loop started.
Best, Jesper Lorentsen
My computer went into BSOD and, being unaware of the bug, I treated it as a possible SSD malfunction and somehow this got my computer working again. This is what I did:
– Restart the computer. Autorepair Windows did not work. Bios listed the drive.
– Connect the SSD it externally to another computer and check content – readable.
– Downloaded Macrium reflect (free version) to make an image of the SSD.
– The program reported "MFT corrupt" and suggested using Chkdsk /f which I did.
– I made the image and put the SSD back in the original computer.
– Now the Windows autorepair started automatically but reported not being able to repair the problems. I went into the advanced options but decided to do another restart before trying a rebuild of the mbr.
– Windows autorepair started again and remained running – probably fixing stuff.
– Then Windows restarted correctly.
At this moment the computer is still a little buggy, but I may also have a problematic SATA cable I need to replace. I will remove the KB4592438 update and hope that helps.
Sorry for that, but Linux is an answer!
My old Dell with a Crucial BX500 SSD running 20H2 has been virtually unusable since kb4592438 update.
Very slow to boot, then when it does, clicking on icons nothing happens for about 10 seconds.
Uninstalled the update and it's all OK again.
But my Asus with a Samsung SSD is fine. Weird.
Yes, I've had the same thing happen on two laptops. After upgrading to 20H2 I tried to run my backup s/w to copy image backup to external HD. I got read error on C-drive (SSD), ran chkdsk (I'm an old-timer, retired, so know chkdsk). On reboot got the BSOD and had to recover from USB install back to 1903 and update from there.
This happened to me! I had windows on a small, old SSD. It's like an ancient SSD from when they were first hitting the market so I just thought its time was up.
I tried to repair the old drive and after a day of effort via some magic combination of recovery options and manual booting from the BIOS I got it running again. Then, the next day, I got another freeze and had to restart. That kicked off the auto chkdsk again and broke windows again.
I bought a whole new drive to replace it with and did a fresh reinstall since I didn't have hours every day to keep repairing windows. I've been going fine since but it never crossed my mind chkdsk was the actual thing causing the problem.
I still don't know precisely what the issue was but all I know is I installed the windows update, restarted the computer, then a it froze a few minutes after reboot and the whole thing was toast.
I guess I jumped the gun on buying parts but the behaviour it was exhibiting was a lot like stuff you would see when I drive is starting to die.
o meu notebook deu o mesmo problema mas consegui recupera lo e ja desistalei a ultima atualização.Pensei que era virus pois quando desliguei normalmente o notebook ele ligou sozinho ja com a tela azul e a parte onde fica o coole estava muito quente tive que forçar o desligamento precionando e segurando o botão de desligar.O meu nootebook nunca apresentou esse tipo de tela
Just a little counter anecdata… I'm on 19042.685 with KB4592438 installed, using a Samsung M.2 SSD, and haven't had any problems. I will refrain from running chkdsk though!
Question – Why are we not all on Linux or Macs? I ask this while typing on my Win7 because my Win10 is in the shop for hardware issues and blue screen issues. I did my due diligence and banned Win10 Crapdates. However, I then reallowed them due to the stupid cracking bug, but not sure if I installed that particular update or not. I don't remember.
Seriously, how many people are considering suing Macroscam? Vista didn't like Open Software, Win10 seems to be the same. I've had more blue screen while runing GIMP2 then I have just using browser alone. Win 10 seems to hate multitasking, wants KISS routines only (aka only checking email, or only doing this 1 thing or that 1 thing.) My old and worn out laptop with Win7 does better. :P
Win 10 2004 with that KB4592438 installed on 10/12. Today 21/12 I noticed my backup disk was no longer being detected. It's actually a disk created from 2 HDD's using Storage Spaces, so no SSD involved.
Launching Disk Management pops up the 'initialize disks' window but when I try to, it displays an error that the device is not present. Will try some other stuff later.
The OS itself is on an SSD and I have two more HDD's combined as a second Storage Space (mirrored) with my data on it. Now running a backup to external storage before these also break. Microsoft, WTF?!?
Thanks for the post, guenni, to help warn others–and for the suggestions so far in the comments, to help others who may have been tripped up already by this.
[And tahnks for the updated post, with news of the MS fix that arrived literally as I was typing this.]
As for geeker's last comments, well I know the signs are everywhere: "Don't feed the trolls"…but…I… can't…help…myself….
Geeker, no, not everyone suffers those problems or finds things as woeful (running Win10, and WELL beyond such basic operations). You've had a bad experience. Got it. So have others. But it does NOT mean everyone using Windows has, or would want to dump it.
Move on if you prefer, sure. But really, don't look back on we who choose to remain with disdain or pity. Some of us are really quite happy with Windows 10–even if we may see that, yes, odd things like the issue in this post may happen.
We just don't all see it as another example of evil hegemony on MS's part, but rather as normal stuff that happens with about any OS–especially any used by hundreds of mullions of users with widely disparate machines, components, use cases, etc. Stuff happens.
And there are tragic tails of woe from Mac and Linux users as well. But really, let's not get into a *-measuring contest, if you may feel there are none or far fewer.
To each their own. Seriously.
(I know, I know, folks. I'm literally dropping troll food out of my basket here. But this post caught my attention, and it will catch that of others. I could just walk away, but bullying and fake news both need to be called out, right?)
In fact, literally as I was typing this (and before I hit submit), I happened to see Günter (guenni) just offered a new post and updated this one to report that a fix has been identified. Whaddya know. Even evil bastards who make mistakes can right them:
I experienced this as well on two different hosts with Crucial MX500 SSDs.
Asus Z390-A + TOSHIBA-TR200 hit >_<