You are not logged in.
Pages: 1
I am running vanilla Arch on a Dell XPS 13 9343, with i3 window manager, no desktop environment. 6.1.3-zen1-1-zen kernel. The computer suddenly went down for a reboot. Unfortunately, I was not present, so I do not know what happened just before the reboot. I had only gone away for a minute or two. When I came back, I was presented with the display manager (SDDM). Here are the relevant journalctl logs:
Jan 06 22:02:32 archlinux kernel: audit: type=1334 audit(1673064152.440:168): prog-id=0 op=UNLOAD
Jan 06 22:02:32 archlinux kernel: audit: type=1334 audit(1673064152.440:169): prog-id=0 op=UNLOAD
Jan 06 22:07:14 archlinux kernel: perf: interrupt took too long (3248 > 3140), lowering kernel.perf_event_max_sample_rate to 61000
Jan 06 22:10:28 archlinux pulseaudio[551]: Battery Level: 80%
Jan 06 22:10:58 archlinux pulseaudio[551]: Battery Level: 90%
Jan 06 22:14:58 archlinux pulseaudio[551]: Battery Level: 80%
Jan 06 22:15:28 archlinux pulseaudio[551]: Battery Level: 90%
Jan 06 22:23:28 archlinux pulseaudio[551]: Battery Level: 80%
Jan 06 22:23:58 archlinux pulseaudio[551]: Battery Level: 90%
-- Boot cb411fcdceb94eac8253e08ed0a0c920 --
Jan 06 22:25:22 archlinux kernel: microcode: microcode updated early to revision 0x2f, date = 2019-11-12
Jan 06 22:25:22 archlinux kernel: Linux version 6.1.2-zen1-1-zen (linux-zen@archlinux) (gcc (GCC) 12.2.0, GNU ld (GNU Binutils) 2.39.0) >
Jan 06 22:25:22 archlinux kernel: Command line: BOOT_IMAGE=/vmlinuz-linux-zen root=UUID=63878d7f-7aba-45fc-9b9a-bc2de8128ed0 rw rootfsty>
Jan 06 22:25:22 archlinux kernel: x86/fpu: Supporting XSAVE feature 0x001: 'x87 floating point registers'
Jan 06 22:25:22 archlinux kernel: x86/fpu: Supporting XSAVE feature 0x002: 'SSE registers'
Jan 06 22:25:22 archlinux kernel: x86/fpu: Supporting XSAVE feature 0x004: 'AVX registers'
Jan 06 22:25:22 archlinux kernel: x86/fpu: xstate_offset[2]: 576, xstate_sizes[2]: 256Can someone tell me what happened?
Last edited by Gabachin (2023-01-07 05:03:49)
Offline
Jan 06 22:10:28 archlinux pulseaudio[551]: Battery Level: 80%
Jan 06 22:10:58 archlinux pulseaudio[551]: Battery Level: 90%
Jan 06 22:14:58 archlinux pulseaudio[551]: Battery Level: 80%
Jan 06 22:15:28 archlinux pulseaudio[551]: Battery Level: 90%
Jan 06 22:23:28 archlinux pulseaudio[551]: Battery Level: 80%
Jan 06 22:23:58 archlinux pulseaudio[551]: Battery Level: 90%Was the device on battery or AC?
Spontanous reboots are typically because of
1. underpowered
2. overheated
3. bad HW (RAM, CPU)
and the battery level development looks suspicious.
Online
The computer was on AC. I thought the battery development looked weird too, because the polybar battery module displayed 100%. The module is just a script that runs
sensorsand prints out the relevant information via
awkI also have a custom polybar module that displays the temperature and fan speed. As far as I know, there are no overheating issues. The machine did run rather hot with youtube videos playing, around 80C, but I undervolted the computer slightly, which dropped the temperature to around 70-75C.
Last edited by Gabachin (2023-01-07 14:07:11)
Offline
did run rather hot with youtube videos playing, around 80C, but I undervolted the computer slightly
1. underpowered
2. overheated
Revert the undervolting and then let's have a brief look at the system getting too hot (fan config and youtube playing, I'd suggest to avoid the browser and look at external players like youtube-viewer. I watch youtube at ~0% CPU load)
Online
Now this is strange. This morning, the same thing happened on another computer. A spontaneous reboot on a Dell Vostro, same OS, same WM, same Linux kernel. Maybe the charger is failing? Or a bug in the last update? A problem with Chrome? It's not really a reboot, though. It is a logout. Both times this has happened (on the two different computers), it went directly to the display manager. A snippet of the journalctl log that I think is around the time of the problem:
an 07 11:15:24 arch kernel: [ 138597] 1000 138597 296208315 21247 970752 1832 300 chrome
Jan 07 11:15:24 arch kernel: [ 138683] 1000 138683 296474569 274080 3530752 1802 300 chrome
Jan 07 11:15:24 arch kernel: [ 139165] 1000 139165 687018 24071 1032192 0 0 mpv
Jan 07 11:15:24 arch kernel: [ 139723] 1000 139723 296182787 3634 544768 1873 300 chrome
Jan 07 11:15:24 arch kernel: oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),cpuset=/,mems_allowed=0,global_oom,task_memcg=/user.sli> Jan 07 11:15:24 arch kernel: Out of memory: Killed process 138683 (chrome) total-vm:1185898276kB, anon-rss:1059160kB, file-rss:17592kB, >Jan 07 11:15:24 arch systemd[1]: session-3.scope: Killing process 705 (lightdm) with signal SIGKILL.
Jan 07 11:15:24 arch systemd[1]: session-3.scope: Killing process 744 (i3) with signal SIGKILL.
Jan 07 11:15:24 arch systemd[1]: session-3.scope: Killing process 804 (nm-applet) with signal SIGKILL.
Jan 07 11:15:24 arch systemd[1]: session-3.scope: Killing process 808 (polkit-gnome-au) with signal SIGKILL.
Jan 07 11:15:24 arch systemd[1]: session-3.scope: Killing process 814 (dunst) with signal SIGKILL.
Jan 07 11:15:24 arch systemd[1]: session-3.scope: Killing process 816 (xscreensaver) with signal SIGKILL.It seems like chrome was killed and then the other processes followed suit. So maybe it is a chrome issue?
Or maybe Xorg is crashing?
cat /var/log/Xorg.0.log | grep -E "(EE)|(WW)|\error|\failed|\Time"
grep: warning: stray \ before e
grep: warning: stray \ before f
grep: warning: stray \ before T
[ 11831.561] Current Operating System: Linux arch 6.1.3-zen1-1-zen #1 ZEN SMP PREEMPT_DYNAMIC Wed, 04 Jan 2023 16:28:17 +0000 x86_64
(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
[ 11831.561] (==) Log file: "/var/log/Xorg.0.log", Time: Sat Jan 7 11:15:24 2023
[ 11831.567] (WW) The directory "/usr/share/fonts/Type1" does not exist.
[ 11831.571] (WW) Open ACPI failed (/var/run/acpid.socket) (No such file or directory)
[ 11831.586] (WW) Warning, couldn't open module nv
[ 11831.586] (EE) Failed to load module "nv" (module does not exist, 0)
[ 11831.588] (WW) Warning, couldn't open module fbdev
[ 11831.588] (EE) Failed to load module "fbdev" (module does not exist, 0)
[ 11831.590] (WW) Falling back to old probe method for modesetting
[ 11831.816] (WW) NVIDIA(G0): Unable to get display device for DPI computation.
[ 11831.821] (II) NVIDIA(G0): ACPI: failed to connect to the ACPI event daemon; the daemon
[ 11831.842] (II) Initializing extension MIT-SCREEN-SAVER
[ 11831.842] (EE) AIGLX error: dlopen of /usr/lib/dri/i965_dri.so failed (/usr/lib/dri/i965_dri.so: cannot open shared object file: No such file or directory)
[ 11831.842] (EE) AIGLX error: unable to load driver i965Last edited by Gabachin (2023-01-07 18:17:44)
Offline
It's not really a reboot, though. It is a logout.
Means the session or the display server crashed - that's not the same as the spontanous reboot at all.
The reason is apparently an OOM => add a swap file or partition.
The OOM killer is basically just gonna terminate the most memory hogging processes (chromium ~1GB at the time and browsers are notorious RAM consumers)
This doen't even have to have a misbehaving process, it's just that you tried to use more RAM than is available.
An xorg crash would not be captured by that grep, so please don't randomly grep logs unless you're looking for a very specific token (and even then you remove pot. relevant context)
Online
Got it. Thanks. On the XPS 13, I had created a swapfile but I forgot to edit my fstab file to make it permanent. On the Vostro, though, I did create a swap partition of 2G when I installed the OS. Apparently, this is not sufficient? The Vostro has 8G in RAM and an i5 10th gen processor. I suppose I could add more swap by creating a swapfile. I'd rather not mess with the partitions. Should I do this? As far as randomly grepping logs, well, I got it as a suggestion on reddit. Mea culpa...but I guess I should have done more research before issuing the command. BTW: I installed gtk-youtube-viewer and added a keybinding to i3, and it works like a charm. Temps stay <60C.
Last edited by Gabachin (2023-01-07 21:44:01)
Offline
"free -m"
If there's sufficient swap, before any OOM you should experience some slowdown (for all the swapping out)
Whether 2GB is sufficient depends on the available RAM and the use-profile.
If you've 2GB RAM and 2GB swap and intend to do 4k video editing, that's not gonna fly ![]()
Online
The Vostro has 8G in RAM and an i5 10th gen processor. I thought 2G in swap would be enough. Just to learn about the m flag, a question: when I do
free -mI get
free: Multiple unit options doesn't make senseI am used to just doing
freewhich gave me (on the 9343)
total used free shared buff/cache available
Mem: 7797 1620 518 591 5658 5282
Swap: 4095 2 4093which I can read ok.
What is the m flag supposed to do? And what does the error mean? Guess I could just RTFM.
Last edited by Gabachin (2023-01-07 21:54:10)
Offline
The m option gives a readout in mebibytes. I prefer -h though.
Also:
free: Multiple unit options doesn't make sense
Is free an alias or something?
type freeJin, Jîyan, Azadî
Offline
What is the m flag supposed to do? And what does the error mean? Guess I could just RTFM.
Yes, and who or whatever configured that system setup an alias:
type freeEdit: blast.
There also seem to be 4GB swap.
swaponLast edited by seth (2023-01-07 22:03:06)
Online
Yes, I have 4G in a swapfile, which I just created. Before, I had none, because I had forgotten to edit fstab to make it permanent. And yes! I aliased free -m to free. I had forgotten. Ugh.
Last edited by Gabachin (2023-01-08 00:46:10)
Offline
So that's the dell w/ the spontanous reboot, not the vostro w/ the OOM crash?
free and swap aren't relevant to the dell w/ the spontanous reboot only the vostro w/ the OOM condition.
If you want to investigate the latter as well, you should better start a new thread w/ the relevant data there and ITT focus on the situation of the dell system.
Namely whether you removed the undervolting and whether that had beneficial outcome.
As for browser video acceleration, both big browsers have relevant wiki entries and there's a general wiki on video hw acceleration (whether it's available and works) and for chromium there's also an eternal thread https://bbs.archlinux.org/viewtopic.php?id=244031&p=29
If there're further issues, you should start a 3 thread for that.
Back to the dell system w/ the spontanous reboot:
Status update?
Online
Just a few cents on the OOM situation: For systems with little memory you may want to set up a userspace OOM daemon to handle OOM situations more fine-grained. The kernel handler is rather aggressive; a user space daemon will let you prioritize and isolated process to be killed, and will ideally help you to specifically kill memory-hungry applications instead of taking down the entire session.
The wiki lists a few alternatives at https://wiki.archlinux.org/title/Improv … conditions. I only know systemd-oomd, but my experience has been good so far. If you consistently run applications and processes through systemd-run (to move the process into a separate systemd scope, i.e. cgroup) systemd-oomd is very good at killing only specific memory-hungry processes while leaving your session alive.
Offline
Pages: 1