Frequent crashes, nothing in logs

[distancy]

I'm experiencing fairly frequent crashes running Ubuntu on my laptop. This has actually been an issue on-and-off for a few years. There will be a few weeks or months where it happens often (up to multiple times per day) and then doesn't happen for long stretches.

What typically happens is that:
* the display continues to respond to input (mouse or keyboard)
* the icons in the tray (both top right and applications on the left) either disappear or turn into grey boxes
* audio sometimes cuts out if playing
* sometimes one of the open applications will pop up "this application is not responding, force quit or wait"
* if I try running (eg) htop in the termal I get "Input/output error"
* this state may last for a few minutes -- I've stayed on video calls for several minutes from this point before a full crash rebooting.
* I need to reboot.
* I don't see errors in the logs (journalctl or /var/log/syslog) from the time of the crash.

It's happening often this week and is extremely frustrating. Any help in getting to the bottom of this would be much appreciated!


Some basic system info (from neofetch):

OS: Ubuntu 20.04.1 LTS x86_64
Host: XPS 15 9560
Kernel: 5.4.3-050403-generic
Uptime: 17 mins
Packages: 3305 (dpkg), 1 (flatpak), 14 (snap)
Shell: bash 5.0.17
Resolution: 1920x1080
DE: GNOME 3.36.4
WM: Mutter
WM Theme: Adwaita
Theme: Yaru [GTK2/3]
Icons: Yaru [GTK2/3]
Terminal: gnome-terminal
CPU: Intel i7-7700HQ (8) @ 3.800GHz
GPU: NVIDIA GeForce GTX 1050 Mobile
GPU: Intel HD Graphics 630
Memory: 3612MiB / 23782MiB

[VMC]

I see you have multiple graphics. Did you install any drivers?

[TheFu]

I'd think it was some sort of hardware or data corruption issue. SSDs that are failing have causes this issue that I've seen.

Input/output error

can be a bad cable, failing storage, bad PCI seating, power supply spike, cooling problem.

Have you setup journalctl.conf to retain logs between reboots? If not, it is hard to troubleshoot HW issues.

[distancy]

I see you have multiple graphics. Did you install any drivers?

Thanks for the reply! Here's the driver info that I know how to find, is there anything else that would be helpful to show?

Screenshot from 2021-01-22 18-37-05.png

[distancy]

I'd think it was some sort of hardware or data corruption issue. SSDs that are failing have causes this issue that I've seen.

can be a bad cable, failing storage, bad PCI seating, power supply spike, cooling problem.

Have you setup journalctl.conf to retain logs between reboots? If not, it is hard to troubleshoot HW issues.

Thanks! I don't think I have enabled it explicitly but `journalctl --boot=-1` does show logs from the previous boot (and no errors just before the crash) so I think it is enabled. Any other ideas for how I could investigate this?

[distancy]

I'd think it was some sort of hardware or data corruption issue. SSDs that are failing have causes this issue that I've seen.

can be a bad cable, failing storage, bad PCI seating, power supply spike, cooling problem.

Have you setup journalctl.conf to retain logs between reboots? If not, it is hard to troubleshoot HW issues.

The most recent time this happened my command gave "command not found" rather than input/output error, if that's helpful.

[TheFu]

The most recent time this happened my command gave "command not found" rather than input/output error, if that's helpful.

That just confirms a storage subsystem issue ... or you were mistyping the command. New people who don't use tab-completion often mistype commands. Heck, even with tab-completion, people sometimes mistype commands.

[ActionParsnip]

Test RAM health. Make sure that's OK
Make sure you have the latest BIOS

[distancy]

That just confirms a storage subsystem issue ... or you were mistyping the command. New people who don't use tab-completion often mistype commands. Heck, even with tab-completion, people sometimes mistype commands.

Yep tried `ls` as well, which I think less likely to have mistyped and also got `command not found`.

Do you have any suggestions on how to proceed? Any diagnostics that would help narrow it down?

[distancy]

Test RAM health. Make sure that's OK
Make sure you have the latest BIOS

Thanks for the suggestions!

I think I have the latest BIOS. This page shows that the latest version is 1.21.0, and running

sudo dmidecode -s bios-version

shows me 1.21.0 as well.

Will test RAM and report back.