Windows 10: Regular System Reboots involving ntoskrnl.exe - on system idle

Hi there, am having a problem with a new computer system. Freshly installed with windows 10 Pro. Have attached the DM log collector files Attachment 90539

For the first week this system ran without problems. Left online for long periods copying network files and downloading etc. Then on July 14th began to suffer from unexplained system reboots after between 30-60 mins of being left without keyboard or mouse input. These don't happen while the system is in use and I've been able to confirm 6-7 hour periods of intense usage without any crash.

Each of the idle crashes comes up with a "The computer has rebooted from a bugcheck." listing in the Windows event review. Each of these reboots basically restarts the machine and comes back to the login screen.

I've downloaded the BlueScreenView minidump reviewer program to find out what drivers are involved and the results of these crashes are as follows :

Between 14/07/2016 and 16/07/2016

Bug check String Caused by Driver / Address

Unknown NTFS.sys +a63e0
Unknown tcpip.sys +1fb2c8
Memory_Management ntoskrnl.exe +142760
IRQL_NOT_LESS_OR_EQUAL ntoskrnl.exe +142940
Memory_Management ntoskrnl.exe +142940
Page_fault_in_nonpaged_area ndis.sys +97860
IRQL_NOT_LESS_OR_EQUAL ntoskrnl.exe +142940
Bad_Pool_Header FLTMGR.SYS +37059
Unknown ntoskrnl.exe +142940
Page_fault_in_nonpaged_area ntoskrnl.exe +142940

On July 14th I installed some Windows Updates, Latest Nvidia Geforce drivers (368.81) and some Epson printer drivers.

Just to re-iterate this crash reboot symptom only happens while the computer is idle for an extended time without keyboard or mouse input and happens regardless of whether the computer is literally doing nothing - or actively downloading from the internet (steam etc).

The PC is overclocked. But I have tested reverting to system defaults (non overclocked this past weekend for 30 hours or so and the last few crashes have happened with the non overclocked status)

Any help or suggestions gladly appreciated.

:)

 

Continuing BSOD issue

OK, I will try that for a while. There has only been one more BSOD in the last two days. Again, occurring while the system was idle (Checked for any scheduled activities during that time - none). Again, it was the same fault ntoskrnl.exe +142940.

 

Windows 10 new machine crashing

I have a windows 10 Laptop that's new. System keeps rebooting from an error. How can I send a memory dmp to be analyzed

ntoskrnl.exe ntoskrnl.exe+188529 fffff801`46080000
fffff801`46956000 0x008d6000 0x5a979177
2/28/2018 10:36:55 PM

Here's a screen shot from Blue Screen View...

 

Attachment 90544

Have added an image of the BlueScreenView Analysis - it actually shows that in another two instances Bad Pool and Page fault - the ntoskrnl.exe was involved in the crash as one of two cited drivers .. so that means currently 8 of the 10 reboots have involved this driver I believe.

 

One of the dumps form the 16th indicates a problem with BHDrvx64.sys. This belongs to Symantec Eraser Control which is part of Symantec AV.

Code: ffffd000`20e703d8 fffff800`61113601Unable to load image \??\C:\Program Files (x86)\Norton Security\NortonData\22.6.0.142\Definitions\BASHDefs\20160711.001\BHDrvx64.sys, Win32 error 0n2 *** WARNING: Unable to verify timestamp for BHDrvx64.sys *** ERROR: Module load completed but symbols could not be loaded for BHDrvx64.sys BHDrvx64+0x173601[/quote] A couple of the dumps indicate memory_corruption but this could be due to a driver.

Please uninstall all Norton/Symantec programs from your computer. Enable the built-in Windows Defender and Windows Firewall and run with them for awhile, See how the system behaves.

 

I'll try that tonight and let you know how it survives through the night. The crashes have been very regular so far so if it stays up should mean we're getting somewhere.

Cheers.

 

Okay tried that, no luck, the system rebooted after about 20mins same symptoms. I've gone ahead and reinstalled windows 10 and will try to leave it fresh for a time with just the browsers and without Norton and see if the problem continues on a fresh installation.

 

Okay ... Fresh Installation of windows, have been careful in what I've installed.
Chrome, Steam. Razor Mouse Drivers. WD Access - No Norton Security Suite as yet.

Its just stayed stable while idle for 2 hours - which beats its previous post-problem record. I'm wondering if the Norton uninstall I did on the old version of windows didn't completely remove the problem driver or something.

Additionally - I lost the Power Management Custom session that my Manufacturer had running and had to recreate another custom one derived from the High Performance default - I notice that this one has the Minimum processor state set to 100% whereas my previous 3XS custom one had a min state of 10% - not sure what the implication is here but that change + clean installation no Norton .... and a general refresh of the windows drivers / systems seem to be the only real differences.

Will keep testing today and let you know if its sorted out or the problem reoccurs!

Thanks

 

When you got rid of Norton, you should use the Norton Removal Tool. Sorry, I should have mentioned that.

Keep track of what you add then it will (or should be) obvious whena installed program causes a problem.
I'm hoping all is well and you will continue to run without problems.

 

Okay couple of weeks on - still haven't reinstalled Norton. Same graphic drivers and the problem has started happening again. Here's an updated set of log files.

Literally there's nothing much installed on the computer but windows / steam/uplay / an WD external HD access program, Sonos controller and GPU-z

Symptoms are as initially described in the thread - its fine in use, even for long periods - but if its left idling it crashes and reboots every 30-60 mins without fail.

Any help again gladly appreciated!

 

PS. I know you chaps don't really like the Bluescreen viewer program but the summary screen shows the bug check strings and all 4 of these show the ntoskrnl.exe highlighted in red at the time of the crash.

One thing that I also notice is that even when i do a controlled shut down with by clicking power/shutdown in the windows and the machine powers off ...

It nonetheless creates an error 41 "The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly." message in the event viewer.

This does seem strange to me.

Cheers.

 

The Razor Mouse Drivers has caused some problems like BSODs. Even though I don't see any indication of it in your dumps. it might be a good idea use a different mouse, just a plain USB connected would be great.

What else is connected to the PC? Try disconnecting everything except Mouse (not Razer), keyboard, display and Ethernet cable.

If that doesn't help, try doing a Clean Boot, see if that helps.

See here: How to Perform a Clean Boot in Windows 10 to Troubleshoot Software Conflicts

 

Have ditched the Razer mouse drivers and gone for a basic USB mouse.
Also removed the realtek sound selection program.

Only things I've got connected is

Asus g-sync monitor by display port (main display is running at 144mhz)
ethernet cable (I've disabled the unused connected)
USB wireless headset G930 dongle (in a USB 2 slot)
USB 3 hub (into which plugged the standard usb mouse and keyboard + and g930 usb charger)

If it still crashes with the razer and realtek drivers gone I'll try bypassing the hub and putting the 3 usb devices directly into the motherboard power slots.

Will play around with the clean boot options tomorrow - cheers.

 

Okay ditched the hub and plugged keyboard and mouse on the default windows drivers directly into the motherboard. Ran a clean boot with only Nvida and Steam helper enabled (aside from MS services). Still got a crash / reboot after approx 60 mins of idle. Am attaching the updated log file - the crash was around 2.43 am.

While waiting for a response I'm trying to run the desktop on 60mhz refresh rate for a while and see if that makes a difference ...

Cheers.

 

Okay interesting findings today ...

System has been up for 6 hours without a crash since I changed the desktop screen refresh rate to 60 mhz from the 144 mhz that was the default for the g-sync monitor ( ASUS ROG Swift PG278Q ) - interestingly looking at the power draw on 1080 card and idle temps with gpu-z the load is about 20x as much in 144 mhz and the temps much higher.

I'm wondering if the system crashing from idle is to do with some issue about the nvidia drivers and the way 144 ghz refresh rates are handled in win 10. It certainly makes windows run smoothly but probably not 20x the power draw smoothly truth be told!

Will leave it on tonight to test further but this seems something of a breakthrough ... it might correlate with my earlier findings post re-installation that it ran for a week or so with no problems but I probably switched up the refresh rate with the new nvidia drivers...

Again any thoughts gladly received ...