OS: Rocky Linux 8.10
Hello, my PC has been freezing/hang a lot lately. And I don’t know what is the cause.
I usually leave my PC on 24/7, but recently I come back to PC completely froze.
Screen is frozen, mouse not moving, keyboard not working (caps lock not turn on/off when pressed). Motherboard light still on, PC fan still on.
Everytime this happens I need to hard restart by pressing the power button on my PC case to turn it off and then turn it back on. Weirdly enough this seems to happen only when I leave my PC, sometimes nothing is running, maybe I leave a twitch livestream on. But hasn’t froze when I am actively using it. (It now freezes too when I am actively using it)
Last night I tried to use the older kernel that I thought would work (when PC boot, theres options to pick kernel I think? and I chose the third one because this PC has been updated twice recently). And the PC successfully lasted overnight, with livestream on. I started to play games through Steam, and after I’m done and leave the PC, it froze.
And now scarily enough, when I hard restart, it just froze on login screen. I haven’t even touched the PC. So I hard restart again and now here I am.
I am not sure what is the cause..
(Yes I am using Rocky Linux for every day use, just because of Davinci Resolve.)
EDIT: it finally freezes when I am actively using it, was just browsing discord and nothing else on the background. It also froze the next time when I tried to edit this post… its getting a lot more frequently now
it seems to be more aggresive now? It just froze after I created this topic, closed my Flatpak Firefox, and went to grab some water.. and came back to frozen PC (keyboard doesnt work, mouse doesnt work)
Maybe my system files is corrupted? Is there something similar to “sfc /scannow” on Windows?
These symptoms are familiar to me for always on systems. The usual culprit is the power supply, or power connections to the MB and drives. The first thing to do is power off and then unplug/replug each power connection several times to the MB sockets and drives. Oxidation does occur overtime. Do the same for your drive sata connections while you are at it.
My PC case (from 2016) and Motherboard do have rust in them for a very long time, because we live in a humid area and unfortunately it was too late until I realize I needed to have a dehumidifier.
Even with dehumidifier in my room everything is still getting rust today (not just PC).
It might be the PSU, it is very old (2016) and there was a sudden power outage in the past 3 days while I was using my PC. I’m not sure if it started happening since then or before it, but definitely started in the past 4~ days. Everything was fine since I switched to Rocky Linux 8 in November 2024, sure there’s some crashes, power outage during that, but it has never crash/hang this frequently before.
Things I will try:
I will absolutely try to do the things you mentioned.
Clean the PC and reseat everything.
Get new PSU and UPS .
Have spare B450 motherboard that I could use to test.
I will run MemTest86 (Free Edition).
I also tested with other Live USB and they also hang/freeze/restart
Ubuntu 24.04 LTS (after 1h 40m. The time is varied, mostly immediately, if i’m lucky it’ll last longer)
EDIT: celebrated too early, but it lasted for 3h 26m~ before system froze up again (Ubuntu 24.04 LTS) had to hold power button again and turn it back on. It froze again few seconds after I logged in.
I guess it’s Memtest time..
Yup, I just did that. And I think the temp was fine before (I did look at it somewhat)
So I’ve done:
unplug-replug the PSU connections multiple times (24pin was a tough one)
reseat CPU, apply new thermal paste
reseat RAM, GPU multiple times (fan header too while I’m at it)
unplugged all internal sata storage for now
clean the PC a little bit, will go more thorough tomorrow because it’s already 12 AM
And so far so good. I am now able to transfer files without freezing/restarting (it keep failing before). able to install Ubuntu 24.04 LTS to external SSD and I am using that for now (still struggling to reinstall Nvidia drivers on Rocky Linux).
No Memtest yet but I’ll use the PC for now.
I think I still want to get a new PSU, and save up for a UPS.
It’s worth checking the last few entries in the logs, just before the freeze. Make sure you have journalctl persistence, and then use ‘-b’ to check the previous boot. If it’s a hardware freeze it might not show anything interesting, the computer just stopped at that point, and the o/s didn’t log anything, but sometimes failing hardware will cause the o/s to log something.
You can also check voltages either direct in the bios, or using something like lm_sensors. It won’t help with transient spike, but if the 5V rail is reading 4.7V…
When it freezes, does the screen retain the image of what ever was on the screen at the time, or is it just black?
EDIT: It may be too early to say this, but seems like the PSU was the problem. Swapping in the new one and I’ve been using it for 3h 46m. Thank you everyone for all the help
Thank you so much for taking your time to help, I really do appreciate it so much!
It’s worth checking the last few entries in the logs, just before the freeze. Make sure you have journalctl persistence, and then use ‘-b’ to check the previous boot. If it’s a hardware freeze it might not show anything interesting, the computer just stopped at that point, and the o/s didn’t log anything, but sometimes failing hardware will cause the o/s to log something.
Yes I did try to open up the Logs program on Ubuntu, but I don’t know what to look for..
You can also check voltages either direct in the bios, or using something like lm_sensors. It won’t help with transient spike, but if the 5V rail is reading 4.7V…
Yes I will try to look that up when the new PSU arrive (before swapping), right now I am just running Memtest86+ (with XMP enabled) while waiting for it to arrive later today
When it freezes, does the screen retain the image of what ever was on the screen at the time, or is it just black?
Sometimes it retains the image (Mouse doesn’t move, keyboard doesn’t work, Ctrl + Alt + F3 doesn’t do anything) and sometimes it restart itself (a lot more common after replug & reseat) but I don’t remember it being just black screen.
Hoping that the new PSU will solve it
even if it doesn’t, I think it’s still a good time to replace
Thank you so much again for your time and your help
Hi everyone, I thought everything was good but today my PC is unusable again.
It started freezing when I was trying to rename files, really low power consumption.
And then it kept freezing and restarting by itself, sometimes this [Hardware Error] message shows up for a short time and then goes into splash screen. But it has been crashing when I logged on less than 1 minute. When I try to look for logs, freeze. When I am inputting my password on login screen, freeze. When the splashscreen loads (spinning circle) can freeze there too..
This [Hardware Error] message shows up for both Rocky Linux 8.10 and Ubuntu 24.04 LTS, which was working fine since the PSU swap (until today). I have to hold power button to turn off every time (which is frequent), so theres probably corrupted files? thankfully(?) my internal HDDs doesnt auto mounts (as far as I know) so hopefully it’s fine there..
It’s combination of just freezing (screen stays on what was I was looking at), and freeze into self-reboot
The rough timeline is this:
Cleaned my PC pretty decently before swapping PSU, including reseating CPU, RAM, M.2, and reapplying thermal paste
Swapped the old PSU to the new PSU (to ENERMAX Revolution D.F. 12 850W)
Live USB doesnt freeze anymore, and able to try out distros, but eventually comes back to Rocky Linux 8 because how easy it is to install and get Davinci Resolve working.
There has been freeze/restart here and there, but I think it has been only 10? and I thought Steam on Rocky was the culprit, so I installed Ubuntu on external SSD when I want to play a game.
PC has stayed on long hours (overnights), and also gaming for long hours without issue. it hasn’t crashed while I was playing game, even with my old PSU. (perhaps I’m just extremely lucky)
I’ve run PassMark MemTest86 FREE after swapping PSU, and it passed the test without errors. It ran for 8h 49m, with CPU temps of 43C/70C/61C (Min/Max/Ave).
Currently running MemTest86 again, 5h 40m in, Pass 3/4 and No errors yet.
Is my PC doomed? RIP CPU? Motherboard? GPU?
I’ve been running my PC with AMD “Eco Mode” for as long as I remember, it’s either 45W EcoMode or 65W EcoMode.
MemTest is run with 65W EcoMode and XMP turned on. I haven’t tried to run it without it yet, will do after MemTest finishes.
I do have another B450 Tomahawk Max and Ryzen 5 3600 that I could use to test. But maybe someone can direct me on what to do to troubleshoot.
Oh also I want to ask, is it safe to post syslog online? I assume it could contain sensitive data..
I asked Google Gemini about your post, and it provides (among much longer answers) some interesting information:
This error code (Bank 5: bea0000000000108) is a well-documented and notoriously frustrating issue specifically associated with AMD Ryzen processors on Linux systems.
It is important to understand that this is a hardware-related stability signal, not necessarily a sign that a component is physically broken. In many cases, it points to the CPU receiving insufficient voltage for its current frequency, or an interaction issue between the CPU, the motherboard’s power delivery, and specific power-saving states (C-states).
Machine Check Exception (MCE): This is a mechanism where the CPU reports an internal error to the operating system.
Bank 5: On many Ryzen chips, Bank 5 errors are frequently linked to the processor’s internal cache hierarchy or power management logic.
The Code bea0000000000108: According to developer analysis, this often indicates a “thread has stopped executing” due to a timeout. Effectively, the CPU core became unresponsive for a split second, causing the kernel to panic or log the error.
When a user enables “Eco Mode” (which effectively enforces a lower PPT/TDC/EDC power limit), the motherboard and CPU are forced to operate within a constrained power envelope. If the CPU’s default voltage curve (VID) was calibrated for the higher performance envelope but is now being forced to adhere to lower power limits, it can lead to “under-voltage” scenarios during rapid frequency transitions, which manifests as that specific Bank 5 Machine Check Exception.
Recommendation: Suggest they try disabling Eco Mode entirely to see if the MCEs disappear. If the system becomes stable at full stock power limits, then the issue is confirmed to be the conflict between their cooling/power delivery and the Eco Mode constraint.
You COULD yourself, but apparently DID NOT, because you have apparently have not tried (or reported) trying the specific recommendation, of disabling Eco mode.
Replaced the CPU and things seems to be working again. It’s probably the CPU problem from the start but I have no regret getting a new PSU.
Found couple of reddit threads with the similar mce [Hardware Error] issues with Ryzen CPUs. And they also had to replace theirs after many troubleshooting/similar freezing issues.
I’m extremely relieved that it’s not RAM (2nd MemTest passed without errors) or GPU (Tested with another system and it runs benchmarks fine.)
Many years ago I had a different but related issue with unexpected hanging and freezing on a self-built desktop computer. After trying different Linux distributions, FreeBSD, Windows, it was clear that the issue was with hardware and not operating systems. Replacing the power supply didn’t help. After a long time of trial and error I found it was the motherboard that had to be replaced (bought the same model used). BTW I wouldn’t so easily discount responses from Gemini and other LLMs, they are often more useful than random Reddit posts that Google search forces on us.
