1

I have a laptop with Ubuntu 20.04.2 LTS, 64 bit.

From some time now, it suffers from inexplicable crashes. At random time, without any heads-up or warnings, the audio stops, the screen gets black, the keyboard doesn't work anymore and I lose ANY kind of control over the laptop. Yet, the leds are still on, showing the laptop is still running... something. Waiting doesn't solve it, nor closing the screen.

Alt-Stamp-K or (Ctrl-)Alt-F* have no effect whatsoever, just like any other combinations of keys, and all I can do is wait for the battery to drain, or long press the power button to kill it.

I have also temperature control on, but it doesn't seem to be correlated. The only common thing to all crashes was that I had Chrome open, but that may be because I always have chrome open...

My plan is, at the next crash, to annotate the exact time at which it happens. What I am asking, is: where can I find logs/recordings/a djinn that tells me what the artichoke is going on in the pc while it is crashing, what are the causes of the crash, and that I can consult when I restart?

2 Answers 2

1

I'd call that a hang, btw, rather than a crash.
I don't know nix well-enough to tell you what logs to look for, so this is a broader scope, based on the symptoms.

Tough to diagnose, but your symptoms are common in three cases - overheat, under-current or failing GPU.

Overheat doesn't have to be CPU temperature, it could be GPU, as could the other two reasons, which is why it's tough to diagnose at consumer-level without spares you can swap in to test.
One test, not truly conclusive but could go part way, would be to have music playing. If at the hang the music starts to cycle one tiny section over & over [may be from a few milliseconds; little more than a buzz, to 100ms or more with an almost identifiable repeat], then it's probably not a CPU issue, it's a GPU hang… though it doesn't tell you whether that's overheat or under-power.

First thing would be to strip & clean the whole thing, including splitting off CPU & GPU coolers & re-applying thermal paste [or pads]
2nd would be to install some temperature monitoring that will log to disk, so you can check back afterwards.

If it leans towards being a power issue, probably time for a new battery, but run those preliminary steps thoroughly first - it's cheaper ;)

1
  • 1
    yes, the music does repeat endlessly a few milliseconds of music when it crashes. I didn't even know that you could split the coolers in a laptop... but anyway, I already have a GPU temperature monitoring, and it does not detect particular high temp before the crashes, but I'll try to make it log somewhere
    – Exodd
    May 9, 2021 at 13:15
1

In ubuntu all logs saves in /var/log/

kernel - save messages from you kernel
syslog - from all process`

You may add to /etc/sysctl.conf

#To enable SySRQ MagicKey - like Alt+SysRq+F, Alt+SysRq+Prnscreen and other
kernel.sysrq=1

#restart Laptop after 5sec after panic
kernel.panic=5

And run command 'sysctl -p' or 'reboot'

0

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .