Over the last handful of months or so I've noticed occasional BSOD's seemingly out of nowhere on my Win11 daily driver VM.  I have an older Nvidia passthrough video adapter, everything had been pretty much working pretty cleanly for a while until I noticed a BSOD - Memory Management every so often.  Over the last 4-5 weeks or so it seems to have been increasing and certainly over the last week, enough that it's caught my attention and made me realize it is time to do something.  Doesn't appear to be any rhyme or reason when or why it happens and I realize that it could just be a Windows / VM issue, because I have no indication of anything unusual on the Unraid side of things.

 

Windows event log is pretty generic, App Events just show:

Fault bucket , type 0
Event Name: BlueScreen
Response: Not available
Cab Id: 0

 

Figured I would post in here to see if anyone has any ideas before I start thinking about maybe just starting fresh from my last clean build backup.  Another thought I've had is using a linux distro.  I'm not a linux expert but also don't shy away from it, although maybe about 1.5-2 years ago I messed with a few different distro's as a daily driver VM and feel like I ran into some issue where the OS would hibernate and I couldn't wake it up and had to hard reset the VM using Unraid or something along those lines.  Been long enough ago that I don't remember why I abandoned that idea other than the few that I tried, incl Ubuntu, all had that issue and I just moved on.  Maybe it was an older Unraid version or VM misconfig on my part, but I'd be interested to hear of anyone that has a stable distro recommendation that seems to work great as an Unraid VM.

New issue today, VM just powered off out of nowhere.  Powered it back on in Unraid, noticed this in the syslog:

 

Dec  8 08:55:07 unServer kernel: qemu-system-x86[12748]: segfault at 104219410 ip 000055c0589b7764 sp 000014bc31e83e00 error 4 in qemu-system-x86_64[55c05866b000+5c0000] likely on CPU 10 (core 4, socket 0)

Fingers crossed I found the culprit.  I always assumed that if there truly was a RAM issue that somehow Unraid would make it known, not just within a VM.  All of a sudden yesterday though, firefox kept giving me crashed tab errors very frequently.  In searching for what could be the culprit, I saw RAM and using Memtest86+ mentioned, so booted up to the Unraid boot screen and selected the Memtest option.  Sure enough, failures right off the bat.  Went through the process of testing 1 chip at a time.  2 of them were fine, 1 errored out immediately just like the initial test and with just the last one, the server wouldn't even boot up properly.  Has been fairly stable with just the 2 good chips in and waiting on replacements 🤞