[SOLVED] Random Power Off / Reboot - where to look next?

[jb_11000]

Hi everyone!

I'm quite new to Linux in general and especially Debian so please be gentle!

I currently have Debian 10 installed (4.19.0-14-amd64) on a dual-boot with Windows 10 on my laptop.
I am having an issue with Debian randomly "powering off" roughly 5-10 minutes after the first bootup of the day, after which the system boots up again. I wouldn't describe it as a proper 'reboot' as the system literally powers off in an instant, as if the power had suddenly been cut off.
After this has happened once, it then runs fine - until the next day, when it happens again. This is the only pattern I am able to ascertain, as it doesn't seem to matter if I boot into Windows, or shut the computer down for a while - it invariably runs fine until the first boot into Debian of the day, and it happens again.

So far I have checked the logs using journalctl, as well as every entry in /var/log, and there is no obvious cause: no error messages (at least not within 5 minutes of the shutdown) or even shutdown related messages. Every time it has happened I have checked these logs and there doesn't appear to be anything happening related to the shut down, for example in the most recent case there was a 'connection blocked' message from UFW around 30 seconds before it happened, with previous records occuring several minutes before the shutdown.

Similarly, when rebooting there is nothing in the logs which appears out of the ordinary when compared to any other session.

After googling about I read somewhere that there could be interference in the power button circuit causing the system to power off, which I didn't think was the case as a) it's a pretty new laptop (1yr) and has run Windows 10 with no issues and b) surely if this were the case there would still be a record of it somewhere in the logs?

Nevertheless I read that I should try editing /etc/systemd/logind.conf to include the line HandlePowerKey=ignore (which was previously commented out), however this has not had any effect.

So does anybody know what else I could check to try to find the cause?

Thanks in advance and sorry for the long-winded post!

[Head_on_a_Stick]

Is the machine over-heating? That should leave some messages in the logs but it's all I can think of.

Have you installed the CPU µcode package? Intel Haswell processors in particular are quite broken without it.

[jb_11000]

Is the machine over-heating? That should leave some messages in the logs but it's all I can think of.

Have you installed the CPU µcode package? Intel Haswell processors in particular are quite broken without it.

Thanks for replying. I don't believe it is due to overheating as this usually occurs within 5-10 minutes of booting up, without having ran any applications. Also I have previously ran Hashcat on the laptop (on Kali) up to a temp of 80 degrees with no issue.

That said, where could I look to see what the recorded temp would have been around that time?

I haven't heard of the µcode package, I have a core i9 which I think is Coffee Lake? Would it be worth while to install it still?

Cheers

[Head_on_a_Stick]

where could I look to see what the recorded temp would have been around that time?

A record of temperatures is not kept (AFAIK) but you can use xsensors or conky to keep track.

I have a core i9 which I think is Coffee Lake? Would it be worth while to install it still?

Yes, I think so. The package is called intel-microcode.

[jb_11000]

Ok, I will have a look into those suggestions. Thanks for your replies.

[Deb-fan]

Would agree with the Hoasinator. For new cpu's like an i9 would really consider installing a newer kernel from Buster backports, also any required firmware along with that, including microcode for anything that has newer backported versions available. Sounds like a real oddball type of tech mystery. Place to start would be what I'd just typed.

It's a head scratcher, only happens on 1st boot of Debian ? Doesn't happen with Win10 ? Kinda at a loss as to what to suggest, a kernel upgrade definitely seems prudent though, isn't going to hurt anything and it's been long standing good practice to keep another kernel as backup just in case.

If you do that above would be good if you'll report back so others know. Also can't hurt giving some information about the spec's of your system, many easy/effective tools to show what type of hardware you're dealing with ie: "lspci -vv" or using inxi that thing can probably tell you the name of the person in China who built that pc, lol.

Other totally baseless (take for what it's worth) speculations, perhaps see if there's a firmware(uefi)upgrade available, not even really saying it's likely because what you're describing is ghost in the machine stuff but buggy bio's certainly do happen and can end up being patched and upgrades put out to address them too. Ah I'm still mostly at a loss here and this is about everything I can think to suggest.

Oops, PS ...

No shortage of system monitoring tools or methods for gnu/nix, ask google. Gonna doubt it's a temp related thing but that's just my opinion or worse, a hardware issue, if it were something of that nature surely there would be more problems and symptoms involved. Think it's just a matter of installing newer software packages and making sure you've got the right ones installed for that system to work well.

[jb_11000]

Thanks guys. As of today, all of 48hrs after installing the intel-microcode package, the shutdown has yet to occur. So I would perhaps tentatively call it solved - I'll update again if it doesn't happen over the next few days. Embarrassingly enough, when I looked into it there are a few Debian guides which recommend installing this package as part of the initial setup so I guess I didn't do my homework properly!

Thanks for your advice Deb-fan, I had started to look at re-configuring the kernel anyway (just to trim out some of the unnecessary bits rather than make any major changes) so I will take that on board.

Interestingly another issue seems to have been righted - I have been running ParrotOS through virt-manager/QEMU/KVM and it was 50/50 whether it would freeze after the Grub menu on boot-up - now it has worked straight away every time since installing the package. I wouldn't know whether this is pure coincidence though!

I had checked for a BIOS/UEFI upgrade but I'm not confident as there isn't a clear guide on Clevo's website (for my MoBo anyway) and I've never done it before so would be quite likely to do something bad!

I copied the output of lspci -vv here in case anybody is interested:
https://pastebin.com/UVWkiXF9

b.t.w. laptop was built by J. Stoner of PC Specialist in Wakefield

[Deb-fan]

^Nice.

Sounds like you're doing your homework though. Only ever flashed bios a couple times and one time Hp borked the laptop involved, was still under warranty, so sent it back and they had to fix it.

If things are working well wouldn't mess with it and if someone decides to do so think they should read the instructions beforehand. Don't get me wrong its not rocket science, do or die only think of it as a sometimes why bother to fix if nothing's broken type of situation.

Anyway welcome to Debian.

[stevepusser]

8th-gen Intel generally works better with firmware-misc-nonfree installed from the nonfree section, too. If you get a new kernel from backports, make sure to update firmware packages from backports, too.

[jb_11000]

I've run daily for a week now without issue, including upgrading the kernel via backport, so thanks again for all the help and advice.