I leave my computer running for long stretches because it also acts as a plex server. I turned my monitor on last night to open steam, and the window didn’t render in; I didn’t think much of it, but before I could restart my PC, I got a “memory management” BSOD. I turned off XMP, as well as taking out each RAM stick, but I continued to get BSODs. Either “memory management” or “critical process error”. Some other things I’ve attempted:

I can’t reset the PC; when I try, I’m told “there was a problem resetting your PC”.

I can’t use a system restore point; that also fails.

When I open the terminal and run

sfc /scannow

it finds and fixes corrupt system files every single time (I’ve attempted 3 times now), but I still get a “critical process error” BSOD.

I attempted running

DISM.exe /Online /Cleanup-image /Restorehealth

before “sfc” but this got an error 87 and didn’t work.

When I turn on the PC, my lights for DRAM and VGA stay lit up for maybe 5-10 seconds, but turn off while I’m in BIOS or the windows startup repair screen (and my monitor is plugged into the graphics card; my CPU can’t do display out).

I’m at a bit of a loss here. My next guess would be to attempt to reinstall windows, but I don’t have another windows PC handy to create bootable media, so I’m hoping I have a thumb drive laying around with an ISO on it, or I’ll need to wait to get one from a friend.

Also, in the event that reinstalling windows is the fix, should I disconnect the drives holding my plex media beforehand? Wouldn’t want to risk them getting wiped

An update: I have a drive with installation media for Windows 10 laying around, but when I got to the point where it was installing files on my boot drive, partway through it said it didn’t have the required files and cancelled.

THE FIX: Turns out it was an issue with my RAM. I plugged in a thumb drive containing memtest, and after running the test received a ton of errors. Swapping in a new RAM kit seems to have totally resolved my issues. The PC boots up perfectly fine now

  • comador @lemmy.world
    4·
    7 months ago

    Sounds like memory corruption. Before you do like the other poster said and run memtest, I’d suggest reseating the ram, possibly in different banks if you can, reseat the video card and then try again.

    If you still get failures:

    • memtest
    • update bios
    • dive deeper into the windows system logs for clues
    • then check your ssd for bad blocks using the SSD vendor tools

    Always backup your important files!

    • DonnieDarkmode@lemm.eeOP
      1·
      7 months ago

      Are there vendor tools that run off of a flash drive like memtest? I can’t boot into windows at all

        • DonnieDarkmode@lemm.eeOP
          2·
          7 months ago

          Ultimate boot looks like an extremely useful tool, thanks for mentioning that. After testing each RAM stick in the primary slot for my MOBO, I’m running memtest and getting a lot of errors (now over 500 and still on pass 1)

          • comador @lemmy.world
            4·
            7 months ago

            Most RAM have a lifetime warranty, so please take a picture of the errors and use that to rma (return) to the memory maker for a free replacement.

            In the meantime, consider buying new ram (unless you can wait 6 weeks for an rma return lol).

  • themoonisacheese@sh.itjust.works
    4·
    7 months ago

    You should run memtest for 24 hours with all ram sticks in.

    It definitely looks like your ram is somehow faulty, have you tried resetting CMOS beyond just turning off XMP? It could also be your motherboard’s VRM overheating, but i don’t know how you would test that without a bunch of thermocouples.

    • DonnieDarkmode@lemm.eeOP
      2·
      7 months ago

      Running memtest, I’m at over 400 errors before the first pass even completed. I take it that’s probably a RAM issue?

    • DonnieDarkmode@lemm.eeOP
      2·
      7 months ago

      I haven’t reset the CMOS yet; I’ll give that a try. But I did try booting with a single stick inserted, swapping them out, and both attempts gave me a BSOD (the critical system error one). My CPU cooler has a tiny fan for the VRMs, and my case has excellent airflow, so I think that one is less likely

  • thantik@lemmy.worldEnglish
    3·
    7 months ago

    This sounds very similar to a problem I had when an NVME drive was failing; it wasn’t writing certain blocks; they had become read only and the only thing I could do was clone the corrupted data off of the drive and do a reinstall.

    I ended up moving my data to a more reputable drive, and then RMA’d the NVME drive, only to have the replacement be used in my wife’s PC, and the same thing happen 6 months later.

  • AnUnusualRelic@lemmy.world
    2·
    7 months ago

    Fwiw, this wasn’t in windows but the problem was a little similar, I once debugged a hardware problem for weeks until I finally found out that the cpu had crapped out for some reason. I managed to have it swapped on warranty and all was well.

    I’d even changed the motherboard (and ram, and psu) at one point and was running out of ideas, the cpu was the last possibility.

    Just to say that it can be many things.

    But the ram is definitely the usual culprit.

    • DonnieDarkmode@lemm.eeOP
      1·
      7 months ago

      Here’s hoping it’s RAM (memtest seems to suggest so); thankfully DDR4 is probably as cheap as it will ever be right now

      • comador @lemmy.world
        2·
        7 months ago

        Suggestion: Look up the mobo and download the tested memory modules for it. In there you’ll find part numbers compatible with the mobo that you can shop for online.