I think my pc has an m.2 allergy...

[MrMcyeet]

Hi friends, I'm back with another random and obscure issue that I've not been able to find anyone else having.

 

 

A while back, just before starfield launched, I upgraded the PC that I built ~2019. Threw in 32g of corsair memory and a shiny new 2tb m.2 drive.

After about ~70 hours into starfield, I would get seemingly random BSOD issues (see this thread I started here). I managed to figure out that it was related to the SSD. No idea what was wrong, but I think the conclusion that someone (either here or on reddit, I forget) helped me come to is that there was some issue with the windows xbox (or gamebar, I forget which exactly) services was trying to use it wrong, or expected it to be some xbox specific thing.

 

Strange, I know, but whatever. I ended up replacing that SSD with a Samsung 970 evo plus, which produced the same BSOD error until I installed the Samsung NVME drivers from their website.

Once I did that, the issue was seemingly gone, which is great!

 

 

Well, apparently not. Just recently, following an update to cod: mw3, I started getting a really weird crash (which sucks, since that is what I've spent a large portion of my free time playing recently)

Things in the game would just... stop working? then eventually freeze my whole PC, requiring me to hold the power button to reset it. Super weird, and it was contained to mw3, so I figured that the issue was an mw3 issue and not a me/my pc issue. There were one or two people who said they had the same or similar issues over on reddit, but I couldn't really find any similarities between the reports.

 

Because this crash thing, whatever it is, would crash my whole pc, getting normal screen recordings was seemingly impossible, as they would all stop before or right around when it would crash (I assume OBS was just never able to save the recording). I managed to get a recording with my phone, that can be found here (skip to about 14:56. Prior to this its just normal gameplay)

 

 

Fast forward to tonight, I was taking a shower and trying to logically think about the bug, and maybe MacGyver my way into a solution. I remembered the previous BSOD issues I had had, and wondered if this was similar? Thinking about it, it somewhat made sense. Random things in the game just stopped loading, maybe the SSD was somehow just... not sending data or something? idk, but I did have an old 2.5in 500g Samsung sata SSD lying around. I installed it and removed my m.2 SSD, reinstalled windows, reinstalled cod, discord, vivaldi, and a whole host of other things I use daily and it has seemingly stopped! Previously I was lucky to play for more than ~30m without it crashing (though it was inconsistent, sometimes it would happen instantly and other times it wouldn't happen in a whole ~60m long match).  Today, though, I've managed to work my way through 6x ~45/60m games without it crashing! I'm unsure if this is just good luck, and it still may happen or not, but I'm hopeful that its solved.

 

But, this leads to my actual question: What the actual freak could be causing this? My (somewhat) self educated mind thinks that my PC just doesn't like m.2 SSDs, but that doesn't make a whole lot of sense, really.

Right now I just have 2x unused 2tb m.2 SSDs, and only a 500g SSD in my PC, which Isn't ideal.

 

So what can I do to further troubleshoot this? Does anyone have any insight into this issue that maybe I don't?

 

 

On an unrelated journey through the internet, I recently stumbled upon this dual m.2 to sata SSD thing. I think its meant for NASs or something, but I'm unsure. So, is it worth buying this and seeing if my PCs apparent allergy to m.2 SSDs is remedied by using this adapter/thing? I mean, I wouldn't mind using it in either JBOD or raid 0 mode, and having a 4tb SSD available. I don't care about redundancy, really.

 

A current list of my components can be found here.

[Mark Kaine]

i suspect memory issues. 

 

full specs please? 

[problemsolver]

3 hours ago, MrMcyeet said:

Hi friends, I'm back with another random and obscure issue that I've not been able to find anyone else having.

 

 

A while back, just before starfield launched, I upgraded the PC that I built ~2019. Threw in 32g of corsair memory and a shiny new 2tb m.2 drive.

After about ~70 hours into starfield, I would get seemingly random BSOD issues (see this thread I started here). I managed to figure out that it was related to the SSD. No idea what was wrong, but I think the conclusion that someone (either here or on reddit, I forget) helped me come to is that there was some issue with the windows xbox (or gamebar, I forget which exactly) services was trying to use it wrong, or expected it to be some xbox specific thing.

 

Strange, I know, but whatever. I ended up replacing that SSD with a Samsung 970 evo plus, which produced the same BSOD error until I installed the Samsung NVME drivers from their website.

Once I did that, the issue was seemingly gone, which is great!

 

 

Well, apparently not. Just recently, following an update to cod: mw3, I started getting a really weird crash (which sucks, since that is what I've spent a large portion of my free time playing recently)

Things in the game would just... stop working? then eventually freeze my whole PC, requiring me to hold the power button to reset it. Super weird, and it was contained to mw3, so I figured that the issue was an mw3 issue and not a me/my pc issue. There were one or two people who said they had the same or similar issues over on reddit, but I couldn't really find any similarities between the reports.

 

Because this crash thing, whatever it is, would crash my whole pc, getting normal screen recordings was seemingly impossible, as they would all stop before or right around when it would crash (I assume OBS was just never able to save the recording). I managed to get a recording with my phone, that can be found here (skip to about 14:56. Prior to this its just normal gameplay)

 

 

Fast forward to tonight, I was taking a shower and trying to logically think about the bug, and maybe MacGyver my way into a solution. I remembered the previous BSOD issues I had had, and wondered if this was similar? Thinking about it, it somewhat made sense. Random things in the game just stopped loading, maybe the SSD was somehow just... not sending data or something? idk, but I did have an old 2.5in 500g Samsung sata SSD lying around. I installed it and removed my m.2 SSD, reinstalled windows, reinstalled cod, discord, vivaldi, and a whole host of other things I use daily and it has seemingly stopped! Previously I was lucky to play for more than ~30m without it crashing (though it was inconsistent, sometimes it would happen instantly and other times it wouldn't happen in a whole ~60m long match).  Today, though, I've managed to work my way through 6x ~45/60m games without it crashing! I'm unsure if this is just good luck, and it still may happen or not, but I'm hopeful that its solved.

 

But, this leads to my actual question: What the actual freak could be causing this? My (somewhat) self educated mind thinks that my PC just doesn't like m.2 SSDs, but that doesn't make a whole lot of sense, really.

Right now I just have 2x unused 2tb m.2 SSDs, and only a 500g SSD in my PC, which Isn't ideal.

 

So what can I do to further troubleshoot this? Does anyone have any insight into this issue that maybe I don't?

 

 

On an unrelated journey through the internet, I recently stumbled upon this dual m.2 to sata SSD thing. I think its meant for NASs or something, but I'm unsure. So, is it worth buying this and seeing if my PCs apparent allergy to m.2 SSDs is remedied by using this adapter/thing? I mean, I wouldn't mind using it in either JBOD or raid 0 mode, and having a 4tb SSD available. I don't care about redundancy, really.

 

A current list of my components can be found here.

For all troubleshooting below, keep your working SATA drive disconnected

1. Samsung Magician (on you NVMe Windows install) reports everything as normal correct? (Just ruling out the drive is also bad)
2. Did you have anything connected to the SATA 0 port AND your NVMe was in the upper M.2 slot when you had the M.2 issues? If you did/do, disconnect it from the SATA 0 port. Put your NVMe M.2 drive back in the upper slot and test it
3. Assuming #1 is not applicable or doesn't work, swap the NVMe M.2 drive to the other M.2 slot
4. Verify that you have everything connected to the correct slots/ports so that you don't accidentally disable something
5. See if that fixes it

Also, either way, follow the house keeping portion of this guide and skip steps 2 and 3 to make sure it's not a known issue that was fixed in an update.

If none of that works, time to leave that NVMe drive connected, and collect some dump files for us.
Follow this guide: 

And it will probably tell us quite a bit. It sounds like it could be all kinds of things.

Good news is your motherboard has 2 M.2 slots, so you can try the other one.

[MrMcyeet]

22 minutes ago, problemsolver said:

Samsung Magician (on you NVMe Windows install) reports everything as normal correct?

Yes. This was one of the first steps I took when this started happening. Running each of their drive checking options (there were like 3 or 4) told me that everything was fine.

 

24 minutes ago, problemsolver said:

Did you have anything connected to the SATA 0 port AND your NVMe was in the upper M.2 slot when you had the M.2 issues?

Yes, I had a 2tb hdd in the sata0 port and an 8tb hdd in the sata1 slot. I installed these ages ago and mostly use them for backups and cold storage.

 

 

Just for the sake of a sanity check, I will retest your #2 there, but this did persist regardless of the m.2 slot I used.

Originally I had the TEAMGROUP SSD in the top m.2 slot. When I went to replace it (which also reported good with crystalDiskInfo), I started by installing the samsung drive into the bottom m.2 then removing the TEAMGROUP SSD.

 

Through the course of my trying to troubleshoot this issue, I did swap it to the top slot, though to no effect. When I had it in the top slot, the sata0 port *was* populated.

 

 

In the morning I will remove all other drives, and try using just the samsung m.2 in the top slot and report back. 

 

 

 

31 minutes ago, problemsolver said:

If none of that works, time to leave that NVMe drive connected, and collect some dump files for us.

Currently, when my PC freezes/crashes, there are no dump files created (additionally, there is seemingly no relevant logs in event viewer, at least none that I am able to find)

 

Regardless, like I said, will walk through your recommendations in the morning and post an update with the findings.

[MrMcyeet]

2 hours ago, Mark Kaine said:

full specs please? 

I left a pcpartpicker link with the full specs attached, that can be found here.

 

The system is running an i7-8086k paired with a Gigabyte Z370 AORUS GAMING WIFI (rev. 1.0) motherboard, an RTX 3080, 32gb of Corsair's Vengeance RGB Pro 3200mhz memory (not running with XMP) powered by an rm750x PSU. I have 2x 2tb m.2 drives that arent working in addition to a 500gb samsung 850 evo SSD, and two seagate HDDs, a 2tb and an 8tb.

[problemsolver]

Quote

 

Yes, I had a 2tb hdd in the sata0 port and an 8tb hdd in the sata1 slot. I installed these ages ago and mostly use them for backups and cold storage.

 

 

Just for the sake of a sanity check, I will retest your #2 there, but this did persist regardless of the m.2 slot I used.

Originally I had the TEAMGROUP SSD in the top m.2 slot. When I went to replace it (which also reported good with crystalDiskInfo), I started by installing the samsung drive into the bottom m.2 then removing the TEAMGROUP SSD.

 

Through the course of my trying to troubleshoot this issue, I did swap it to the top slot, though to no effect. When I had it in the top slot, the sata0 port *was* populated.

 

 

In the morning I will remove all other drives, and try using just the samsung m.2 in the top slot and report back. 

 

Your top M.2 slot shares the PCIE lanes with the SATA 0 port (this is a common design btw.) And is why I made the recommendation that I made to disconnect anything in the SATA 0 port.

Which means: nohing good can happen from having any  NVMe drive in the top slot and the SATA 0 port populated. But it also shouldn't crash the computer of course; as theoretically, only a SATA M.2 drive would affect the bandwidth of the SATA 0 port and vice versa (and both of your drivers are NVMe M.2 drives). But it's worth ruling out when you are having problems since they physically share traces on the motherboard.

 

Quote

 

Currently, when my PC freezes/crashes, there are no dump files created (additionally, there is seemingly no relevant logs in event viewer, at least none that I am able to find)

 

Regardless, like I said, will walk through your recommendations in the morning and post an update with the findings.

 

Wait... what? The BSOD collection tool the guide tells you to install couldn't find ANY dump files?

Sorry, I'm just surprised... I'll have to do some research on what to do in that case.

[MrMcyeet]

2 hours ago, problemsolver said:

Wait... what? The BSOD collection tool the guide tells you to install couldn't find ANY dump files?

Sorry, I'm just surprised... I'll have to do some research on what to do in that case.

Forgive me, I haddn't read the BSOD instructional you had linked.

 

Since moving to this older 500gb drive, ive not had it crash any more, so running it now would be useless (which is why I didn't read through it yet).

When I said that no dump files were created, I meant that there are never any MEMORY.DMP or minidumps in the windows folder for me to try and analyze (or have analyzed)

[MrMcyeet]

5 hours ago, problemsolver said:

See if that fixes it

Here is a list of steps and events that I have done and have happened:

 

• Remove all sata drives

• Reinstall samsung nvme ssd (top m.2)

• Reinstall samsung magician

• Run samsung magician

•  - Both short S.M.A.R.T and short scan returned normal
• Run cod:
•  - Crash ≈9 minutes into first game

• Move samsung nvme ssd (bottom m.2)

• Run cod:

•  - Crash ≈2 minutes into first game
•  - Clock read 10:42am (12/23/23) at time screen froze

• Run synapse tool on reboot

•  - Attached here is that report .zip

• Run `perfmon /report` after synapse tool

•  - Performance and resource monitor app was open for 10+ minutes, in spite of this, it just said "Collecting data for 60 seconds..."

•  - All options under the `file` menu were disabled except for the exit option?

• Attached here is a screenshot of that. Not sure whats wrong with it?

 

Im not really sure how to progress past here. In the windows folder there are still no .DMP files and no Minidumps folder.

In the event viewer, apparently nothing happened between 10:40 and 10:47, which is a damn lie, as the clock froze at ~10:42 and the logs at 10:47 are from when I restarted the PC.

Regardless, here is an event viewer log thing that can be looked through, maybe it will provide some insight that I am unable to gleam?

[problemsolver]

4 hours ago, MrMcyeet said:

Im not really sure how to progress past here. In the windows folder there are still no .DMP files and no Minidumps folder.

In the event viewer, apparently nothing happened between 10:40 and 10:47, which is a damn lie, as the clock froze at ~10:42 and the logs at 10:47 are from when I restarted the PC.

Regardless, here is an event viewer log thing that can be looked through, maybe it will provide some insight that I am unable to gleam?

Did you follow what I previously stated here:

 

"Also, either way, follow the house keeping portion of this guide and skip steps 2 and 3 to make sure it's not a known issue that was fixed in an update."

 

When you run the PowerShell script (step 4), please edit your original post, insert a Spoiler then insert a Code block inside the spoiler, set the Language to Python, and paste in the output from your Downloads folder LTTForumsComputerInfo.txt

 

Looking through the Event Logs definitely shows various issues. And yes, it appears Windows isn't saving any dumps.

 

Get back to us after following the House Keeping portion of the guide above.

[Bjoolz]

Not being able to create dump files makes the storage suspect (Lumping the M.2 itself into "storage"). There are some other things that can cause it.

 

• Not enough room to create the dump. They all start as full kernel dumps which can be as large as your amount of RAM + 3-4GB. I've seen 67GB dump files on a system with 64GB RAM. They are then pruned to minidumps later if it's set to create minidumps.
• Page file is too small. If it's on System Managed size, it should just expand it, but if you've set a manual size then it needs to account for huge dump files. The page file is where it dumps all the data.
• Drive encryption. The dump process happens outside of the main Windows so it doesn't have access to encrypted drives.

If none of that is an issue, do you have a SATA drive connected (Not M.2) that you could move the page file to? Note that having it on an external drive is not a good idea and might cause crashes that are not related to the actual issue. 

[MrMcyeet]

13 hours ago, Bjoolz said:

Not enough room to create the dump.

I dont think this is the culprit, as both of my m.2 drives have well over 1tb free. My system only having 32gb of memory, even if the dump size was somehow double that of my memory, it should still be fine

 

13 hours ago, Bjoolz said:

Page file is too small.

I have selected the "Automatically manage paging file size for all drives", so I don't suspect that is an issue, either.

 

13 hours ago, Bjoolz said:

Drive encryption

Never played around with this or thought I needed it, unless windows somehow has this enabled by default, I also don't think that this is the culprit.

 

 

16 hours ago, problemsolver said:

Did you follow what I previously stated here

Just walked through the housekeeping portion of that guide, though no change is observed, I still freeze/crash when using an m.2 drive. Do yall need the LTTForumsComputerInfo.txt file?

 

 

13 hours ago, Bjoolz said:

do you have a SATA drive connected (Not M.2) that you could move the page file to?

I can reconnect one of my HDDs to the 2nd or 3rd sata port (as to not interfere with the m.2 lanes or whatever) and move the paging file (and the dump files to via 'system properties' -> 'startup and recovery' as well?)

I am quite busy today, though I should have some free time on the tail end of tomorrow, I should be able to try this out then.

[Bjoolz]

7 hours ago, MrMcyeet said:

(and the dump files to via 'system properties' -> 'startup and recovery' as well?)

This should not be needed as the drive has to be back online for it to boot Windows and create them from the previous dump  to the page file. 

 

And I write a big edit on my post when I realized I was the one who helped you last time (Here and on reddit) which I guess I didn't save. Going to do a quick TL;DR: If I forgot mentioning it last time, it can equally be the M.2 slot just from the error, the SSD is just way more common. The gameflt driver thing was basically debunked. An engineer at a debugging forum pushed it pretty hard, but all of those people where it seemed fixed eventually came back later with the same issue.  

[problemsolver]

8 hours ago, MrMcyeet said:

Just walked through the housekeeping portion of that guide, though no change is observed, I still freeze/crash when using an m.2 drive. Do yall need the LTTForumsComputerInfo.txt file?

It would be helpful, since we don't really have any data at all to look at at this point.