Arch Linux

AmagicalFishy

Member

Registered: 2016-12-10

Posts: 69

[SOLVED] New install hangs at random when shutting down

Hi.

My last Arch install got a little buggy by the end (had it for ~3yr.)

I've just installed Arch on a brand new laptop, and when I shutdown or restart (typically using shutdown -r now) it seems to hang at random. I want to nip this in the bud and figure out exactly what the issues are instead of just getting used to them as I did in my previous install.

I'm familiar enough w/ Arch that I think I can get along pretty well given some direction—but I often don't know where to start.

- I remember a while ago using something that displayed a visualization of the boot process; it showed me how long every unit took so I could easily see e.g. what took a disproportionate amount of time. Does anyone know the name of that program? Is there something similar for shutting down?
- journalctl shows the systemd journal, right? And systemd... is the thing that manages all the services/processes? Does that mean that all logs from any service are shown when I run journalctl?
- What is dmesg showing that's different than journalctl? While the latter is systemd's logs, dmesg's manpage says it's relevant to the "kernel ring buffer". What is that?
- Where are the kernel logs stored that are displayed when I shutdown (I've removed the "quiet" parameter from my grub.cfg [this is what dmesg is showing, right?)

Something I do notice is that there's a spamming of:

...
[   17.969248] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969254] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969260] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969265] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969271] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969277] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969282] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969287] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969293] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969297] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969302] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969308] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969313] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969319] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969324] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969330] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969336] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969342] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969347] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969353] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969359] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969364] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969370] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969375] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969381] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969387] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969392] nouveau 0000:01:00.0: mc: intr 00000040
[   17.969398] nouveau 0000:01:00.0: mc: intr 00000040
...

I see it in both the startup and the shutdown. This is a laptop w/ two graphics cards—one Intel and one NVIDIA. How would you go about solving this? fwiw, I've got KDE Plasma installed.

Last edited by AmagicalFishy (2021-12-16 16:20:04)

V1del

Forum Moderator

Registered: 2012-10-16

Posts: 25,296

Re: [SOLVED] New install hangs at random when shutting down

1. systemd-analyze in particular with critical-chain, don't think this shows shutdowns

2. Yes

3. dmesg/the ring buffer is basically the log of the kernel, it's normally kept in memory and lost on reboot (and the ring buffer has a certain size after which older messages get lost), however the journal monitors it and persists it on disk as well.

4. See above

In general you might want to read https://wiki.archlinux.org/title/Systemd/Journal to know how and why this works. Of immediate interest to analyze a boot failure happening "previously" you should look through/post

sudo journalctl -b-1

which will show you the entire log of the boot happening before your current one.

AmagicalFishy

Member

Registered: 2016-12-10

Posts: 69

Re: [SOLVED] New install hangs at random when shutting down

Awesome, thank you!

95% of all priority 1, 2, or 3 messages in journalctl are noveau related and, according to the internet, are fixed by installing Nvidia drivers. For anyone's reference, I used:

journalctl -b-1 -o verbose | rg -A2 -B9 --color=always "PRIORITY=1" | less -R

... to look at all messages w/ a priority of 1 (relevant ones were 0, 1, 2, or 3, which are emergency, alert, critical, and error respectively).

(Actually I could have just used journalctl -b-1 -p 3)

Looks like I'm gonna have to dive into this NVIDIA thing again. That caused me a lot of problems before. >.<

edit: Ok, so I installed nvidia and nvidia-prime and... everything seems to be working w.o. a problem (all of the PRIORITY=0, PRIORITY=1, and PRIORITY=2 messages are gone, too). Excellent! A few years ago Nvidia drivers kind of messed everything up for me.

Last edited by AmagicalFishy (2021-12-16 16:25:24)