Arch Linux

samobot

Member

Registered: 2020-09-03

Posts: 6

[Solved] Complete system hangs with no mention in journals

Hi,
I'm having trouble with my arch installation.  I've had it for about 6 months now and I'm really enjoying it.  The only problem that I have is I sometimes get complete system crashes under medium to heavy loads.  I've checked the journals for every crash for about 4 or 5 times now, and there is nothing in common between them, and after most crashes there aren't even any error or warning messages from the time of the crash.  I've had my PC for about 2 years and have never seen similar crashes when I used windows.  My system currently dual boots windows and arch, although arch is my daily driver.  The crashes are fairly regular (at least once a week, but often more if I'm doing heavy work like Blender or playing games)

My specs are:
Asrock A320M-HDV
Ryzen 5 2400g
GTX 1060 3gb
16gb dual channel memory

The crashes seem to be from something deep rooted, as after a crash I am unable to switch to virtual console or even run sys requests (which are completely enabled), however there isn't any mention of a kernel panic in the journals either.  I've considered updating my BIOS but have yet to do so since I've read that you shouldn't unless you are sure you need to.  On the other hand for a crash with no journal information, it seems a BIOS problem could be the culprit.  It seems independent of kernel version since I keep the kernel up to date and it's been about 6 months of crashes.

Should I upgrade my BIOS and what other things could I check/do?  I'm fairly proficient at linux so I'm happy to provide any extra information you guys need.

Thanks in advance.

Last edited by samobot (2020-12-07 13:41:39)

twelveeighty

Member

Registered: 2011-09-04

Posts: 1,456

Re: [Solved] Complete system hangs with no mention in journals

When a crash happens, do you have any control left on it, or do you have to hard-reset the box? You state that "after most crashes" there are no warnings or errors, which means that there *are* some warnings/errors sometimes? Post those here, since they may hint at what's going on.

Also post some additional details on your installed system - what kernel and graphics drivers (with versions) are you running? Are you running a DE such as GNOME or KDE? X or Wayland?

A common cause for high-load crashes is overheating of the CPU/GPU - are you monitoring those temps under heavy load?

samobot

Member

Registered: 2020-09-03

Posts: 6

Re: [Solved] Complete system hangs with no mention in journals

After it crashes I have no control.  I have to reset holding the power button.  I suppose it put it wrongly, I have yet to see any errors in the time surrounding the crashes.  I can send journal logs but I'll have to try to trigger a crash or wait until one happens since I don't have any saved.

As for software, I'm running GNOME 3 under X as a DE; kernel is

5.9.10.arch1-1

.  For graphics I'm using the NVIDIA drivers

nvidia 455.45.01-2

.

I haven't checked temps but I've never had any similar problems under windows.  I will do some stress testing and post the temps.

Edit: I just ran some stress tests with the kind of thing it crashes under (several tabs in chrome + modded minecraft + minecraft server).  The CPU temps look fine, it didn't go over 75C, and the GPU didn't even get to 45C.

Last edited by samobot (2020-12-06 04:02:17)

V1del

Forum Moderator

Registered: 2012-10-16

Posts: 25,361

Re: [Solved] Complete system hangs with no mention in journals

We'd need a journal to check whether it's the same issue but there is a known crasher with certain memory allocations in the 455 series of drivers, try patching nvidia-dkms: https://bbs.archlinux.org/viewtopic.php … 3#p1938113

Also instead of the power button try enabling SysRq and rebooting with the REISUB sequence, if it really is just the graphics driver issue I'm thinking of  that should leave traces in the journal, unless you've actively configured it to not retain anything at all across reboots, which would be a suboptimal choice for debugging issues.

However the fact that you mention six months doesn't match with that issue anyway. So I'd also lean towards a BIOS/firmware issue and that's definitely something you should give a shot.

Last edited by V1del (2020-12-06 11:23:35)

samobot

Member

Registered: 2020-09-03

Posts: 6

Re: [Solved] Complete system hangs with no mention in journals

I do have SysRq enabled fully (not just limited to the partial set it was defaulted to), and after a crash I am unable to reboot even with REISUB.  I haven't deleted any of the journals, however I didn't keep the times or dates of the crashes so to get a journal log I'll probably have to try to get it to crash again.

I'm working finding the right BIOS update for my mobo now, and I'm willing to try patching the drivers I just don't know how to apply the patch.  It was a long time ago now but I don't seem to remember any crashes from the first month or so running arch so a bad driver could be the problem.

I've installed quilt to apply the patch I just can't figure out how to do it.

Edit: I've updated my bios to the newest version and I now have a new problem, windows overwrote grub, which was fine since I still have the arch installation media, but after running grub-mkconfig, os-prober and installing grub it only finds Windows and doesn't create an entry for arch in the menu, even though the I used arch-chroot to get into my arch partition and can go and find the kernel manually

Edit 2: I've since fixed the booting issue and the os loads fine

Edit 3: since updating my BIOS the crashes have gotten worse, I've had system hangs, the computer completely restarting without warning, and even hangs while in grub and in the BIOS menu.  This seems to be not a problem with arch linux, but rather with my hardware or bios at this point

Final edit: I ran memtest86 and it came up with 142 errors, I think I'll need to replace my ram, considering that it's no longer an issue with arch, I'll mark the thread as solved

Last edited by samobot (2020-12-07 13:40:59)