Computer restarts when compiling only when X server runing

razcakappa · n00b Joined: 18 Sep 2021 Posts: 2

Hi,

I just came across a strange problem. When I emerge something, it starts to compile fine, however middle of the way computer fans tends to run in fast and noisy and suddenly computer restarts. When I'm compiling without starting X, it runs (heavy load like compiling) for hours even overnights without failing.

I assume this might be a hardware issue like thermal. However before sending this off to repair, I would be relieved if I can find out the problem's relationship with the X here?
Anyway to debug this problem from the software level?

p.s: As a side note, hardware is Macbook PRO 15" 2015 (Retina) and When I'm using OSx with Adobe apps (I assume those are heavy on resource consumption) I'm not seeing such issue.

Zucca · Posted: Sat Sep 18, 2021 6:32 am Post subject:

I suspect RAM.
You have more ram used when you have X open at the same time.
Do you have portage temp on tmpfs? If so, then try to temporarily move your portage temp outside of the tmpfs (or just unmount it). Then if your compilation goes ok with X open, it's most probably RAM issue.

Or course you could straight away just run some memtest. :P
_________________
..: Zucca :..
Gentoo IRC channels reside on Libera.Chat.
--

NeddySeagoon · Posted: Sat Sep 18, 2021 11:41 am Post subject:

razcakappa,

Welcome to Gentoo.

It could be a thermal issue too. It runs for hours without X. Running X and using it is not free. It needs both CPU and RAM.
Booting into memtest is a good start. You can run it in userspace (no top of the kernel) too but don't do that as its not nearly so useful.

A good CPU cooling stress test is Prime95. It works your CPU and cooling system really hard.

Don't go spending any money on fixes until you post results here.
memtest can show other problems ... not just RAM.

Redoing the CPU thermal paste and cleaning the cooling system is on the easy end of of PC fixes.
Being Apple, the hard bit is getting in.
As its six years old, finding a cooling system full of grot would not surprise me.

Do you have lm-sensors installed?
You can use that to keep an eye on temperatures while you do other things.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.

mike155 · Posted: Sat Sep 18, 2021 11:54 am Post subject:

Does cpupower work on your machine?

You could use 'cpupower' to lower the maximum frequency. I regularly do that on notebooks. Many notebooks were designed to be thin and lightweight - not to compile source code for multiple hours.

Run

figueroa · Posted: Sun Sep 19, 2021 3:23 am Post subject:

Lots of solutions without any diagnostics. It could just be your configuration. Post results of "emerge --info"
_________________
Andy Figueroa
hp pavilion hpe h8-1260t/2AB5; spinning rust x3
i7-2600 @ 3.40GHz; 16 gb; Radeon HD 7570
amd64/23.0/split-usr/desktop (stable), OpenRC, -systemd -pulseaudio -uefi

Tony0945 · Posted: Sun Sep 19, 2021 1:51 pm Post subject:

MaDDeePee · Guru Joined: 18 Aug 2004 Posts: 387 Location: Germany

hello.

memtest86+ without error?
after this, run prime95 - any errors?

This sounds like a hardware problem to me.

cameta · Veteran Joined: 04 Aug 2004 Posts: 1320

A faulty power supply can cause this.
_________________
Si algo falla LEE el jodido manual, Si sigue fallando LEE BIEN el jodido manual.

razcakappa · n00b Joined: 18 Sep 2021 Posts: 2

Wow I haven't expected this kind of exposure to this question. Great team Gentoo!

I just started to diagnose the system as you guys explained. I have a good news and a bad news.

First I've done a memtest86 and the results came clean. It took more than 5 hours to complete, however never experienced any crash or reboot (report https://imgur.com/a/YkutJuS).

However when I trying to run prime95, after I chose the stress test I need and when I give permission to start, system crash and reboots. And I can replicate the results all the time. Even without the the X. When I try

figueroa · Posted: Tue Sep 21, 2021 5:24 pm Post subject:

Laptops being harder to troubleshoot than a desktop machine, will this one run on AC power with the battery removed? I had similar symptoms with one Sony laptop. I'm thinking power, CPU, capacitors, most likely culprits.

To my eyes, your portage configuration looks rational. You could try reducing -j4 to -j1 in make.conf. If it makes any difference, it would be diagnostic.
_________________
Andy Figueroa
hp pavilion hpe h8-1260t/2AB5; spinning rust x3
i7-2600 @ 3.40GHz; 16 gb; Radeon HD 7570
amd64/23.0/split-usr/desktop (stable), OpenRC, -systemd -pulseaudio -uefi

NeddySeagoon · Posted: Tue Sep 21, 2021 5:37 pm Post subject:

razcakappa,

When prime98 causes the crash and reboot, how long does it take?

If you run it with the system just powered on from cold, does it take longer to crash and reboot that if you rerun it once its warmed up?
If yes, you have a cooling system problem. The longer first run from cold is accounted for by the thermal inertia of the CPU and heatsink.

If its not thermal then the power supply to the CPU is implicated.

Your MAKEOPTS="-j4" is a good number with 16GB RAM. Reducing it as figueroa suggests may avoid the problem by not working the CPU as hard as only one make job will run at a time.
Not all build system honour it though.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.