I've been experiencing frequent crashes on my Linux system—at least three times a week. Sometimes these crashes occur while I'm actively using my computer, but often they happen when I'm away and the screen is off. When I return, the system won't wake up; I'm unable to toggle the caps lock key on my keyboard, SSH becomes unresponsive, and the only thing I can do is a hard reboot by holding down the power button. I've attempted to use the magic sysrq key, but that hasn't worked either, despite confirming it's enabled and functional. Strangely enough, there are no error logs available. Here are my system specs:
**OS**: Debian GNU/Linux 13 (trixie) x86_64
**Host**: MS-7B86 (2.0)
**Kernel**: Linux 6.12.41+deb13-amd64
**Uptime**: 12 mins
**Packages**: 2792 (dpkg), 11 (flatpak), 7 (snap)
**DE**: KDE Plasma 6.3.6
**GPU**: NVIDIA GeForce RTX 2080
**Memory**: 5.71 GiB / 31.30 GiB
**Nvidia driver**: 550.163.01
5 Answers
Also, running `sudo dmesg >> ~/dmesg.txt` can help. This command saves the kernel messages, and you can search through it for errors that could indicate hardware issues. Sometimes, those errors can hint at what's going wrong.
In my experience, browser issues can cause crashes on Linux. I had a similar problem with Firefox freezing on my systems, but switching to a Chromium-based browser solved the issue for me. You might want to try that and see if it helps your situation.
Make sure to check your logs! Using `journalctl` can help you identify what's causing the issue. If you don’t find anything, share your logs with us, and we can help troubleshoot further. Also, just a reminder: if there’s a USB disconnect at the time it crashes, that could be a clue!
First off, make sure your BIOS is up to date; sometimes these issues have been fixed in newer versions. Also, consider adding the parameter `processor.max_cstate=1` to your kernel boot configuration, as some Ryzen models have trouble waking from sleep properly. If none of that works, it may be worth investigating whether this is a GPU issue, especially since you’re using Wayland with an older Nvidia driver.
A handy tip is to leave an SSH session open from another machine and run `journalctl -f` to monitor the logs in real-time. This way, when your system freezes, you can catch the last logged event. I'm suspicious that there might be compatibility issues between your Nvidia driver and Wayland, especially since the driver you're using is quite old.
I did look through my logs and found some entries related to NVIDIA right before the crash. I shared the details on Pastebin; do you think it's a driver issue?