Increasing number of "sudden deaths": How to start forensics?


I am using Manjaro for two years now on a Lenovo Ideapad with AMD Ryzen 4700, 16 GB RAM and always the latest kernel (5.17.1-AMD-znver2). It is a consumer-convertible but I used it intensively as a SOHO device during the pandemic.

In the last 3-4 months, I observe an increasing number of sudden “freezings” of the laptop: No keyboard, no mouse, no network interaction any more - and the screen frozen but not black: I have to power off and to restart the system.

Furthermore, I observe sometimes an erroneous pixel on my screen, that persist until I restart the computer. The first time, this happened, I thought the display-hw is broken. But since it vanished after restart, I suppose it is either a procssing-HW (aged RAM?) or SW problem.

In my perception, the frequency of these incidents is increasing: IIRC the first time happened in November/December 2021 and once in a month, now I observe it every 4-5 days.

Since the warranty ends in May, I want to start to do some forensics on the hardware. Can anybody help me, how to log the moment, when the system crashes as well as where to look and how to analyze the loggings?

Thank you very much & kind regards,

$ inxi
CPU: 8-core AMD Ryzen 7 4700U with Radeon Graphics (-MCP-)
speed/min/max: 1522/1400/2000 MHz Kernel: 5.17.1-AMD-znver2 x86_64 Up: 1h 55m
Mem: 8561.3/15374.0 MiB (55.7%) Storage: 953.87 GiB (69.6% used) Procs: 425
Shell: Zsh inxi: 3.3.14

RAM is always a good start. It could be also be overheating, a faulty filesystem or anything else.

You need to look at the journal for hints.

You’re also using a custom kernel, right? Could you try the “official” one?

Yes, you are right: It was from a time, when there were some issues with the Ryzen-4000 CPUs. Nowadays it shouldn’t make a difference any more - hopefully :wink:

Thanky you very much - sometimes it is really strange, when one starts reading the relevant topics: The last time I did somthing similar was on debian 5 - with tail -f /var/log/… . When starting with Manjaro, for some strange reason, I avoided using journalctl. Especially journalctl -b -1 seems to be very helpful now :+1:

I will change to the common kernel and come back to this thread, when I observe the next freeze.


If you come back, please provide the full inxi output:

inxi --admin --verbosity=7 --filter --no-host --width

Also provide error logs from one of the crashed logs:

journalctl --boot=-1 --priority=3

Where the --boot parameter is the number of boots ago that you had a crash. For example, if the crash was the previous boot, you’d set it to -1, as in this example.