Question

Complicated NVMe SSD/possible motherboard issue

Jan 11, 2020 6:50PM PST

I built a custom gaming PC approximately 2 months ago (specs below). Everything had been working peachy this whole time. The setup I had uses two NVMe SSDs, the C: drive containing the OS and all programs, and the D: drive containing all documents, music, pictures, and downloads--as in, I had re-routed the default paths of those Windows shortcuts to the D: drive. Then one day, on Thursday 1/2/2020, all this happened...strap in:

1. In the middle of usage (as in, the computer had been running fine for hours), I tried to open some files on the D: drive, but Windows explorer crashed and stopped responding.
2. I told the computer to restart, since it had been a day or so since I had done so. Windows hangs up on the restart screen. After waiting about 5 minutes watching it just spin, I forcibly restarted using the power button.
3. Now the computer hangs on the motherboard screen on startup and won't boot to windows. A few attempts of restarting changes this to getting hung up on "preparing automatic repair".
4. I'm able to resolve the no-booting issue by disabling quick boot. PC boots as normal.
5. Upon a quick check in files, I suddenly see that the D: drive is not visible in Windows. It is also not visible in disk management. Just simply not there at all. However, upon restarting and checking BIOS, both drives are detected. Using shift+restart and clicking "use device" ALSO shows both drives.
6. Investigation into Device Manager shows that one of the two NVMe controllers has a Code 10 error and will not start. Drivers are up to date, checked and double checked. I try uninstalling the device and scan for hardware changes. Lo and behold, the drive appears in Windows. HOWEVER, I cannot open it or any files or folders linked to it, and attempting to causes windows explorer to hang and eventually crash. This error is reproduceable every time. Upon restart, the NVMe controller just goes back to the code 10 error.
7. Several other seemingly cascading errors crop up. My PIN code for logging in suddenly stops working, and I'm forced to log in with my MS account password. This seems to screw with many permissions on the PC, which resulted in (among other things) sound services refusing to start.
8. I do a system restore to a restore point 5 days previous to these issues. No changes. Even the PIN error persists.

At this point, after googling for ages, I take it to a local tech. He suggest a total wipe of the C: drive and a reinstall of windows. He does so--AND THE NVMe CODE 10 ERROR PERSISTS. We tried one other thing--my motherboard has 3 NMVe slots; M2M, M2P, and M2A. The D: drive (the one that has issues being detected) is in the M2A slot, and the working C: drive was in the M2P slot. We tried removing the D: drive and placing it in the M2A slot where the D: drive was previously. We then made sure to change the boot order in BIOS and tried to start. After POSTing, we got a black screen with a blinking cursor. No boot. The last thing I did was flash my BIOS. Despite successfully updating to a newer version, there were no changes to the issue. At this point, both the tech and I were unable to do anything further.

Of note is that the PC did not suffer any physical trauma of any kind. It was sitting under my desk, unmoved for 2 months. The case has excellent airflow, and I get rid of external dust fairly regularly.

Our current hypothesis: the M2A slot on the motherboard failed somehow, though for whatever reason, not completely.

Alternate hypothesis: the recent windows update broke the NVMe controllers. I have no idea how to even test this.

Things I am waiting to test: I have an NMVe drive USB reader ordered so I can at least test to see if the SSD itself is working.

System Info:

Motherboard: Z390 Aorus Master (BIOS flashed to version F11c just today)

CPU: Intel Core i9-9900KF

GPU: Nvidia GeForce RTX 2080 Ti

RAM: 32 GB (4 x Cool DDR4 G.Skill Ripjaws V series

SSDs: 2x Kingston 1 TB A2000 M.2 Nvme SSD PCIe

OS: Windows 10 64bit Pro, version 1909

If you need further info, please ask. I am baffled beyond belief by this issue.

Post was last edited on January 11, 2020 6:51 PM PST

Discussion is locked

Answer
Follow
Reply to: Complicated NVMe SSD/possible motherboard issue
PLEASE NOTE: Do not post advertisements, offensive materials, profanity, or personal attacks. Please remember to be considerate of other members. If you are new to the CNET Forums, please read our CNET Forums FAQ. All submitted content is subject to our Terms of Use.
Reporting: Complicated NVMe SSD/possible motherboard issue
This post has been flagged and will be reviewed by our staff. Thank you for helping us maintain CNET's great community.
Sorry, there was a problem flagging this post. Please try again now or at a later time.
If you believe this post is offensive or violates the CNET Forums' Usage policies, you can report it below (this will not automatically remove the post). Once reported, our moderators will be notified and the post will be reviewed.
Comments
- Collapse -
Answer
Time to talk to
Jan 12, 2020 11:30PM PST

The motherboard manufacturer. I agree the problem is involved. That is you have a lot of moving parts and it doesn't tell you what one it is.

HOWEVER. I find techs and most folk unable to install Windows correctly. Even I can't tell you how for this PC since I have yet to work with this motherboard and that drive setup. BUT NOT ALL IS LOST THERE. The motherboard maker must supply you how to not only setup the drives (settings, etc.) but which drivers to install and in what order to avoid CODE 10.

In parting I find a few PCs require us to disable Microsoft's Driver Update feature (how, what, etc is on the web) or they work on driver install and after a little time or a reboot fail. Then you reinstall the driver and it works again until the cycle repeats. Your motherboard maker may know if their board requires this tweak.

- Collapse -
This is very helpful info, however...
Jan 13, 2020 4:33PM PST

...this setup was working fine for over 2 months before one day, it simply decided not to. Is this something that can just happen all of the sudden? A driver just deciding not to work after working for some time? Pardon my ignorance, as all my computer knowledge is just self-taught from google and the like.

But talking with the MB manufacturer does seem like the best idea. After my Nvme reader arrives, it will let me see for sure if the drive itself is at least working, so I can rule that possibility out entirely. And if they don't know how to fix it, or it's not a driver problem of some kind, then I can at least start the inevitable battle to get them to send me a replacement, since a MB shouldn't break after 2 months.

Thank you for the reply! This issue has me at my wits end.

- Collapse -
I noted how this can happen.
Jan 13, 2020 5:02PM PST

But what I noted is just one of dozens of reasons.

I think the story is worth repeating in a different way. Here goes. Johnny Windows User has their PC all working nicely. Microsoft pushes a driver update via Windows 10 and blammo, Johnny is having a bad day.

Microsoft is far from perfect and this change about driver updates for Windows 10 while it works 99.99% of the time seems to blow up on a few PCs. The more off mainstream the PC is, the higher the chance this will affect the PC.

Your PC is such a PC.

Also I don't get the feeling you've installed Windows enough (here I think I'm up to some 10+ thousand installs over the years) to know how badly Microsoft misses the mark on their driver selection and how this issue worked its way into updating working drivers for (Microsoft's choice) of a failing driver (code 10.)

-> Again this is why a chat with the motherboard maker is in order.

As to the NVME READER I would not consider that a definitive "this drive is good" tester. The definitive test is a new drive to install along with the drivers, BIOS, BIOS settings that the motherboard maker calls out. THAT and disabling Windows 10 Driver Update feature during the setup.

- Collapse -
This is really helpful and informative.
Jan 13, 2020 7:47PM PST

Just one question though: what do you mean that the NVME reader is not a definitive test? If I can read the SSD from an NVME reader, it means the drive itself hasn't failed, right? As in, it's not a hardware problem with the SSD itself?

One bit of information I can give about the NVME controller driver: it is the microsoft driver version 10.0.18362.1, which is apparently the latest version of that driver, and it's older than my PC build by at least a few months. I was unable to roll back the driver, because there was never a previous version of it on my machine. It is incorrectly dated 6/21/2006 (I have no idea why, but many driver dates are just wrong in Windows), but I've googled it and seen this exact version mentioned in pages from 6 months ago at least. I'm also not a programmer in any way, so I'm also sure that some completely unrelated driver could have been updated and somehow broke this one which was previously working.

I'm also easily able to wipe the C: drive and reinstall windows again if I need to. Any suggestions as to what settings I could use when trying again? You mentioned disabling the driver update feature during the setup; is that something that is prompted during the setup, or do I have to manually force it to do that in some menu?

- Collapse -
Why it is not definitive.
Jan 13, 2020 9:37PM PST

1. After seeing drives work or not on motherboard ports and reverse on USB adapters my experience is why I call it not definitive. The definitive test is a blank new drive and all the USUAL work to get it working.

2. Sorry but no. As I have yet to see a build like yours and even if I did I would have to have it in the armada between office and home. Since I do not have such a build this is why the setup, settings, drivers and all attendant information has to come from the motherboard maker.

3. I keep mentioning about DISABLING WINDOWS 10 AUTOMATIC DRIVER UPDATES. Let's see if Google helps:
https://www.google.com/search?client=you&q=DISABLING+WINDOWS+10+AUTOMATIC+DRIVER+UPDATES

- Collapse -
Answer
Big update to the original problem
Jan 14, 2020 3:58PM PST

In order to test every aspect of the problem, I ordered a USB NVME reader, which just arrived. After hooking it up the SSD to the reader and plugging it in, I get this message:

"E:\ is not accessible.

The file or directory is corrupted and unreadable."

Checking in disk management shows that the drive is RAW. Now, I believe there are ways to recover this without losing the data; is this correct? And how would I do that? At the very least, it would allow me to back up my stuff to another drive before moving on to fix whatever other issues my PC has. There are over 500 GB of data on that drive that I really don't want to lose.

- Collapse -
Maybe.
Jan 14, 2020 4:48PM PST

But the one thing you must never do is write to that drive. That has to be the number one mistake new techs do when learning about data recovery.

For do it yourself work like this you have another drive to hold what you want to extract then run RECUVA in its deep scan mode. There are pages about this so I don't write more about it except the area about recovery.

In case you missed it, you NEVER write to drives you want to recover data or files from. Never.

- Collapse -
Excellent
Jan 14, 2020 5:55PM PST

Deep Scan was able to find basically everything as far as I could tell (94,000 files and such, so I'm not sure exactly if anything was missing). Recuva is currently running, and it is copying all data to another external HD for backup. At that point, I should be able to reformat the drive, then stick it back in the motherboard and hope and pray everything works again. If it does, then holy moly, this issue that has plagued me for the last two weeks will finally be over.

Thanks for your replies and advice. I sincerely appreciate it, since the two other forums I posted this issue on didn't get a single reply despite hundreds of views.

- Collapse -
It has been said.
Jan 15, 2020 1:48AM PST

"We make our own luck."

What I'm alluding to is that you took to heart about never writing to the failed drive. I've handed that nugget for decades and you can bet folk try to fix the drive anyway resulting is more damage that RECUVA can't sift through.

As to the other forums, your machine is a bit out of the ordinary which may be scary.

CNET Forums

Forum Info