Intermittent crashing causing reboot

[AccidentalDenz]

Over the last couple of days, I've had the occasional crash where the PC freezes entirely for a few seconds and then reboots itself - no BSOD or anything like that. It only seems to happen when I'm using my mouse for extended periods, browsing news, forums, etc. I've had the mouse years and it was only a fairly cheap one, so it wouldn't surprise me if it was on its way out, and so it's possibly a hardware error causing the crash. I don't have another mouse to hand, but could certainly get one. Any thoughts on things I could check in the meantime to see if it is this or indeed if it's a problem with something else?

 

[ubuysa]

Windows by default auto-restarts when a 'system failure' occurs, generally system failures write a memory dump so looking to see whether there are dumps would be useful. Minidumps (which is what are likely being taken) will be found in C:\Windows\Minidumps - upload any you find to the cloud and post a link here. A kernel dump would be more use, but you'd notice if one of those was being done because they take a while to write, you'll find the latest in C:\Windows\Memory.dmp - upload that to the cloud with a link to it here. You also must have a paging file to write a kernel dump, so if you've disabled the paging file you'll never take kernel dumps (which is bad).

You might want to prevent Windows auto-restarting on you and ensure it takes a kernel dump when it does fail. To do that enter sysdm.cpl in the Run command box, click the Advanced tab and then the bottom of the three Settings buttons (the one in Start-up and Recovery). Uncheck the 'Automatic restart' box to stop Windows auto restarting, then you'll see the error details - which may well be a BSOD. Also pull the 'Write debugging information' menu down and select 'Kernel memory dump'. Because these are large (well over 1GB) only the latest is usually kept but when you're having problems you want to keep every one - at least until they've been analysed. To do that uncheck the 'Overwrite any existing file' box, the dumps will then be numbered C:\Windows\Memory1.dmp, C:\Windows\Memory2.dmp and so on. Upload them all to the cloud with a link to them here.

What makes you you suspect the mouse? Does it have its own driver (Logitech for example) or is it using the basic Windows mouse driver? Is it a wireless mouse? Replace the batteries if it is.

 

[AccidentalDenz]

Windows by default auto-restarts when a 'system failure' occurs, generally system failures write a memory dump so looking to see whether there are dumps would be useful. Minidumps (which is what are likely being taken) will be found in C:\Windows\Minidumps - upload any you find to the cloud and post a link here. A kernel dump would be more use, but you'd notice if one of those was being done because they take a while to write, you'll find the latest in C:\Windows\Memory.dmp - upload that to the cloud with a link to it here. You also must have a paging file to write a kernel dump, so if you've disabled the paging file you'll never take kernel dumps (which is bad).

You might want to prevent Windows auto-restarting on you and ensure it takes a kernel dump when it does fail. To do that enter sysdm.cpl in the Run command box, click the Advanced tab and then the bottom of the three Settings buttons (the one in Start-up and Recovery). Uncheck the 'Automatic restart' box to stop Windows auto restarting, then you'll see the error details - which may well be a BSOD. Also pull the 'Write debugging information' menu down and select 'Kernel memory dump'. Because these are large (well over 1GB) only the latest is usually kept but when you're having problems you want to keep every one - at least until they've been analysed. To do that uncheck the 'Overwrite any existing file' box, the dumps will then be numbered C:\Windows\Memory1.dmp, C:\Windows\Memory2.dmp and so on. Upload them all to the cloud with a link to them here.

What makes you you suspect the mouse? Does it have its own driver (Logitech for example) or is it using the basic Windows mouse driver? Is it a wireless mouse? Replace the batteries if it is.

I've made the change to prevent auto-restarting so a kernel dump can be done when the crash next happens - will upload it so it can be looked into.

The mouse flashes when it connects/the PC is switched on and it seems to be disconnecting itself every so often - it's wired and so it shouldn't be disconnected while the PC is on. As I say, I suspect it's probably on its way out, no idea if it's causing system failures though!

 

[ubuysa]

I've made the change to prevent auto-restarting so a kernel dump can be done when the crash next happens - will upload it so it can be looked into.

Perfect!

The mouse flashes when it connects/the PC is switched on and it seems to be disconnecting itself every so often - it's wired and so it shouldn't be disconnected while the PC is on. As I say, I suspect it's probably on its way out, no idea if it's causing system failures though!

Get another mouse then ASAP. That's an easy suspect to eliminate.

 

[AccidentalDenz]

Perfect!

Get another mouse then ASAP. That's an easy suspect to eliminate.

I've had another crash since getting the mouse (first one in nearly a week - told you they were intermittent!). The system has crashed but hasn't reset yet - I'm assuming it's doing the kernel dump as you did say that'd take a while. How long should I expect it to take? The initial crash was an hour ago.

 

[ubuysa]

I've had another crash since getting the mouse (first one in nearly a week - told you they were intermittent!). The system has crashed but hasn't reset yet - I'm assuming it's doing the kernel dump as you did say that'd take a while. How long should I expect it to take? The initial crash was an hour ago.

Oh it's quicker than that. The dump is written initially to the paging file so it's pretty fast, a couple of minutes.

 

[AccidentalDenz]

After 90 minutes of waiting, I pressed the power button (I know that's not really advised) and it turned off immediately. Pressed it again and PC booted up - no kernel dump. Was typing up an update message here and checking a few things when it crashed again. In the Event Viewer is two 161 warnings "Dump file creation failed due to error during dump creation." and two error 41 - I suspect the error 41 is probably because I pressed the power button rather than leaving it to sit until it decided it was going to shutdown itself.

Any thoughts on the next step as kernel dumps don't seem to be happening?

 

[ubuysa]

After 90 minutes of waiting, I pressed the power button (I know that's not really advised) and it turned off immediately. Pressed it again and PC booted up - no kernel dump. Was typing up an update message here and checking a few things when it crashed again. In the Event Viewer is two 161 warnings "Dump file creation failed due to error during dump creation." and two error 41 - I suspect the error 41 is probably because I pressed the power button rather than leaving it to sit until it decided it was going to shutdown itself.

Any thoughts on the next step as kernel dumps don't seem to be happening?

Make sure you have a system managed paging file set. If the page file isn't big enoght to hold the dump you won't get one.

I think Ctrl-Slt-Del will reboot from a BSOD but the power button is fine. With auto restart disabled it will just sit on the blue screen.

 

[AccidentalDenz]

Make sure you have a system managed paging file set. If the page file isn't big enoght to hold the dump you won't get one.

I think Ctrl-Slt-Del will reboot from a BSOD but the power button is fine. With auto restart disabled it will just sit on the blue screen.

My paging file was already set to be system managed. In looking around to see if anybody else has had similar issues on other forums, I noticed a few files in the LiveKernelReports folder. Are these any use to you?

 

[ubuysa]

My paging file was already set to be system managed. In looking around to see if anybody else has had similar issues on other forums, I noticed a few files in the LiveKernelReports folder. Are these any use to you?

Microsoft OneDrive

1drv.ms

Those are minidumps taken by the kernel and TBH I never realised the kernel wrote minidumps in there! Minidumps don't contain all the kernel data structures but they are useful nonetheless - especially in this case.

WATCHDOG-20210110-1316.dmp is a VIDEO_TDR_TIMEOUT_DETECTED which means the graphics driver didn't respond in a reasonable time. The driver at fault is clear in this dump, it's nvlddmkm.sys, the Nvidia graphics driver. The stack trace of the active thread shows that it was the DirectX component that was processing but it was the Nvidia driver that failed. Unfortunately the minidump doesn't contain any further details.

WATCHDOG-20210110-1132.dmp is identical - in every single way, including the stack trace.

This could be a graphics card failure, but it's more likely to be a driver error. I would use DDU to uninstall all existing Nvidia graphics driver versions and then download and install the latest driver for your card and OS from the Nvidia website.

Important Later Edit (having looked through your loaded modules list): I notice that you have the Nvidia Virtual Audio Driver installed (nvvad64v.sys). There are other threads on here with people reporting conflicts between the Nvidia Audio driver and their Realtek drivers. I would unindtall this driver if you can. If you can't, use DDU to uninstall the entire Nvidia graphics driver and then reinstall deselecting the audio driver.

 

[AccidentalDenz]

Those are minidumps taken by the kernel and TBH I never realised the kernel wrote minidumps in there! Minidumps don't contain all the kernel data structures but they are useful nonetheless - especially in this case.

WATCHDOG-20210110-1316.dmp is a VIDEO_TDR_TIMEOUT_DETECTED which means the graphics driver didn't respond in a reasonable time. The driver at fault is clear in this dump, it's nvlddmkm.sys, the Nvidia graphics driver. The stack trace of the active thread shows that it was the DirectX component that was processing but it was the Nvidia driver that failed. Unfortunately the minidump doesn't contain any further details.

WATCHDOG-20210110-1132.dmp is identical - in every single way, including the stack trace.

This could be a graphics card failure, but it's more likely to be a driver error. I would use DDU to uninstall all existing Nvidia graphics driver versions and then download and install the latest driver for your card and OS from the Nvidia website.

Important Later Edit (having looked through your loaded modules list): I notice that you have the Nvidia Virtual Audio Driver installed (nvvad64v.sys). There are other threads on here with people reporting conflicts between the Nvidia Audio driver and their Realtek drivers. I would unindtall this driver if you can. If you can't, use DDU to uninstall the entire Nvidia graphics driver and then reinstall deselecting the audio driver.

When I first read this message, the options of a GPU failure or driver error were not too bad. That's a great spot with the virtual audio driver. I've uninstalled that and will see how things go from here. A little hesitant for now, but thanks so much for your help over this!

 

[AccidentalDenz]

As requested last Sunday, I used DDU to remove the graphics driver and re-installed it but without the audio driver. It's crashed again, and using my entirely untrained eye over the minidump, it seems to be similar.

WATCHDOG-20210116-1050.dmp

DMP File

1drv.ms

Any thoughts on next steps?

 

[ubuysa]

As requested last Sunday, I used DDU to remove the graphics driver and re-installed it but without the audio driver. It's crashed again, and using my entirely untrained eye over the minidump, it seems to be similar.

WATCHDOG-20210116-1050.dmp

DMP File

1drv.ms

Any thoughts on next steps?

I've not looked yet because I'm away from the PC, but minidumps are not ideal for diagnosing driver issues. They don't contain all the kernel data structures. If you've got a kernel dump that would so much better.

 

[AccidentalDenz]

I've searched and can't find a full kernel dump where they are supposed to be. All I can find is the kernel minidumps. Had another crash while I wasn't even at the PC a little while ago.

WATCHDOG-20210116-1834.dmp

DMP File

1drv.ms

Tempted to throw the PC out the window or burn the house down or something!

 

[jamiephillips909]

2TB SEAGATE FIRECUDA check that the cable is in properly sounds daft but you would be suprised

 

[ubuysa]

Ok, as it turns out the minidumps do have enough data in them. They are identical, both are VIDEO_TDR_TIMEOUT_DETECTED errors. TDR is Timeout Detection and Recovery and this BSOD happened because a video device (graphics device) did not respond. The driver responsible is clearly identified as nvlddmkm.sys - the Nvidia graphics driver.

The problem is either the Nvidia graphics driver (uninstall with DDU and install the latest version manually from the Nvidia website) or the graphics card itself. If the latest driver (and DDU) don't fix the problem I'd suspect the card.

 

[SpyderTracks]

Ok, as it turns out the minidumps do have enough data in them. They are identical, both are VIDEO_TDR_TIMEOUT_DETECTED errors. TDR is Timeout Detection and Recovery and this BSOD happened because a video device (graphics device) did not respond. The driver responsible is clearly identified as nvlddmkm.sys - the Nvidia graphics driver.

The problem is either the Nvidia graphics driver (uninstall with DDU and install the latest version manually from the Nvidia website) or the graphics card itself. If the latest driver (and DDU) don't fix the problem I'd suspect the card.

On top of this, when installing nvidia driver, do a custom install and DON'T install the HDAudio component. it's been buggy in the last 2 issues causing these exact issues.

That's good advice at any time though, HDAudio is not needed in the first place and often causes issues.

 

[AccidentalDenz]

Ok, as it turns out the minidumps do have enough data in them. They are identical, both are VIDEO_TDR_TIMEOUT_DETECTED errors. TDR is Timeout Detection and Recovery and this BSOD happened because a video device (graphics device) did not respond. The driver responsible is clearly identified as nvlddmkm.sys - the Nvidia graphics driver.

The problem is either the Nvidia graphics driver (uninstall with DDU and install the latest version manually from the Nvidia website) or the graphics card itself. If the latest driver (and DDU) don't fix the problem I'd suspect the card.

Thanks for the clarification.

I used DDU to uninstall the driver and then installed the latest driver from Nvidia last weekend as you advised earlier in the thread.

On top of this, when installing nvidia driver, do a custom install and DON'T install the HDAudio component. it's been buggy in the last 2 issues causing these exact issues.

That's good advice at any time though, HDAudio is not needed in the first place and often causes issues.

I originally had the audio driver installed, but after last weekend's crashes, I used DDU to uninstall the drivers and then re-installed without the audio driver as @ubuysa recommended.

It certainly does seem to be the card - I guess I'll be starting the RMA process. Hopefully I won't be too long without a card!

 

[ubuysa]

Thanks for the clarification.

I used DDU to uninstall the driver and then installed the latest driver from Nvidia last weekend as you advised earlier in the thread.



I originally had the audio driver installed, but after last weekend's crashes, I used DDU to uninstall the drivers and then re-installed without the audio driver as @ubuysa recommended.

It certainly does seem to be the card - I guess I'll be starting the RMA process. Hopefully I won't be too long without a card!

I'd want to do a clean install of Windows and the latest Nvidia driver before going down the RMA, route. You need to completely eliminate software as the potential cause.

If it fails then it probably is tha card.

 

[AccidentalDenz]

I'd want to do a clean install of Windows and the latest Nvidia driver before going down the RMA, route. You need to completely eliminate software as the potential cause.

If it fails then it probably is tha card.

That makes sense. I'll do that tomorrow and see how I get on.