Question "New" build random reboots.

[lakedude]

So my buddy gave me a Gigabyte X 570 UD MB, an AMD 3950x, an EVGA GTX 1070, an Intel 660p series M.2 SSD SSDPEKNW512GB and 4 sticks of G.Skill F4-3200C16-16GVK RAM 🐏. Running W10.

I traded out an untrusted water cooler for a new BeCool Rock 2 air cooler.

I solved a fuzzy display issue by fiddling with the extra long HDMI cables conjoined by a double female.

The thing crashed the first day then ran fine playing Alyx the second day. The 3rd day I played Alyx and then did a torture test running both the CPU and GPU stress tests on Furmark (at the same time) pulling 350 watts. It ran fine with temps around 68 C. I downloaded some VR games and then the system crashed (after it had time to cool down) while I was rearranging the icons on the desktop???

How can it handle gaming and stress tests but not moving an icon?

Happy to supply more info, just ask.

 

[Tech Junky]

First thing I would do is tear it down to components and put it back together making sure everything is seated well and proper pasting / pads are in place. Having been used by someone else there's a slim chance something could be loose or not tightened down properly.

I've done it myself with an air cooler and then later on looking at the temps wondering why they were so high. Went to tighten the screws and I had missed one side other than enough to hold it in place. Clamped it down and temps dropped significantly.

Sometimes RAM will cause things to crash if they're not 100% inserted .

Double check all of the power leads to make sure you have good contacts.

 

[lakedude]

SMART reports OK

Log viewer reports 1 type critical error Event 41 Kernel-Power. 5 since I've had the system and then a 2 year gap and then more from 2020.

There are matching 6008 unexpected shutdown errors.

There are nearly 1000 ETW USB 0xC0000022 errors, nearly all since I've had the machine. I suspect the out of range usb keyboard.

 

[lakedude]

I'm not going to start with the nuclear option first, might need to go there but I'd like to try and narrow things down a bit first before completely rebuilding. The system runs well enough.

Oh the original owner suspects the SSD. He says the system ran great until he added that drive.

The SSD is the hottest thing on the board at idle 55 C. The other stuff gets hotter under load of course.

 

[Tech Junky]

Log viewer reports 1 type critical error Event 41 Kernel-Power. 5 since I've had the system and then a 2 year gap and then more from 2020.

Ok, this one can be a total PITA to track down because it's not actually a power issue.

I went through this with my laptop and it turned out to be an intel driver issue causing the problem. I spent about a month tracking it down because I had actually replaced the power brick with a higher output option which just happened to coincide with the Error 41 / reboots.

 

[In2Photos]

How old is the BIOS?

 

[lakedude]

How old is the BIOS?

Not sure.

 

[lakedude]

You know what, my W8 laptop had some weird issues that turned out to be the software load. It never worked correctly until I was about to give it away and discovered some CLI command for windows to do a deep self check. I saved a link to the thread in the MSI forums but they rearranged their site and I can't remember the command.

Update, played Alyx until the battery on the Quest died, no problems. That is what the rig is for so I'm pretty happy. Would love to get down to zero issues but it is usable the way it is...

 

[lakedude]

Found it! sfc /scannow

 

[DAPUNISHER]

Found it! sfc /scannow

LOL that is meme status on PCMR. System file checker commands have rarely fixed issues on clients systems IME. DISM commands success rate, only marginally better. If you are getting results with it though, good for you.

No way I keep the OS install a friend was using though. If you are in the ecosystems e.g. Google, Microsoft, Firefox; doing a fresh windows install is fast and easy. Log in to everything after installing latest AMD chipset and Nvidia drivers, and you are almost done already. Game launchers are next, and just point them to your game drive/s. Whole thing is less than an hour, and you eliminate all the Windows rot and other shenanigans that can go along with the pre-existing install.

 

[lakedude]

Fingers crossed 🤞.

 

[DAPUNISHER]

How's it holding up?

 

[lakedude]

How's it holding up?

Good!

No issues since sfc scannow but I've only been playing Alyx which was not when it bombed out previously. I mean it has never rebooted while playing Alyx. I'm on chapter 9 playing on hard which has been easier than when I played on easy because I know what I'm doing now.

Best guess is that it either isn't really fixed and I just haven't seen the problem recently or bad clone fixed with sfc scannow.

My W8 laptop must have had a bad clone from the factory because it never did work correctly until the sfc scan. Curiously that laptop currently has W10 and an SSD and works perfectly now. Darn nice machine but I hated it because of W8, the spinner and the bad clone.

 

[lakedude]

And then there was this which obviously points to a storage issue.



Reset, no help.

Just when you think the drive is dead, you turn it off and on again and it lives again.

Starting to agree with the original owner and suspect the M.2 drive or communication to the drive.

How do I make sure I have the proper CD key / Windows activation info to start fresh with a new drive? The MB likely lacks the "tatoo" (as Tech Junky says) that signals to the mothership.

 

[Tech Junky]

When considering a reinstall I just download a new ISO with up to date patches already embedded on the image. These are preactivated as well. We already paid the MSFT tax so, there's no issue there. If it's the slot you can get bootable hba's to bypass the slot.

 

[DAPUNISHER]

When you boot from the USB, choose the option that says you don't have a windows key. It is already tied to the mainboard, so it will activate itself. If you want a local account, remember to keep the LAN unplugged until you reach desktop.

 

[lakedude]

Ordered...

If the socket is bad I'll order an adapter.

 

[lakedude]

The new NVMe drive came today. Windows copied over in little more than a blink.

I forgot to leave the network disconnected until it started pestering me for online account info so I killed the install. Curiously after some repair it backed up a few steps and allowed me a local account. If it hadn't I was going to wipe the drive and start over.

Running good so far but we won't know for a while if it is really fixed.

 

[lakedude]

There is weirdness in Alyx. After the short train ride the pistol and gravity gloves dissappear!