You are not logged in.
Hello, me again!
2 days ago, I had this issue, which got solved pretty quickly.
Today, I did a yay update again, which crashed AGAIN... But this time, it won't even boot, because of this:
error while loading shared libraries: /usr/lib/libmount.so.1: file too short
So, I tried following the same steps as previously. The problem is that pacman is unusable in arch-chroot, so instead I tried using pacman --root, or even pacstrap.
But no matter what I try, I can't seem to reinstall anything, because of that corrupted libmount file.
What can I do to fix this time?
Last edited by TotoShampoin (2024-03-03 15:31:27)
Offline
This is the result of
sudo LC_ALL=C pacman -Qkk --root=/mnt 2>&1 | grep -v ', 0 altered files'
Trying to do "pacman --root=/mnt -S --dbonly ..." outputs the same "libmount file too short" error
Trying to do "pacstrap -K ..." also gives the error
Last edited by TotoShampoin (2024-03-03 11:50:18)
Offline
/usr/lib/libmount.so.1 is provided by util-linux-libs
Do NOT chroot into the installed system and use "pacman --root /mnt --cachedir /mnt/var/cache/pacman/pkg -S --dbonly util-linux-libs util-linux" from the install iso
Then "pacman --root /mnt --cachedir /mnt/var/cache/pacman/pkg -S util-linux-libs util-linux"
Then the rest of the packages.
pacman --root /mnt --cachedir /mnt/var/cache/pacman/pkg -S --dbonly gcr-4 krita lib32-expat lib32-libnghttp2 libadwaita libass libdbusmenu-qt5 libfontenc libjxl libluv libnghttp2 libunibreak libvirt libvirt-python linux mdadm msgpack-c noto-fonts openshadinglanguage pahole python-aiohttp python-ytmusicapi
pacman --root /mnt --cachedir /mnt/var/cache/pacman/pkg -S gcr-4 krita lib32-expat lib32-libnghttp2 libadwaita libass libdbusmenu-qt5 libfontenc libjxl libluv libnghttp2 libunibreak libvirt libvirt-python linux mdadm msgpack-c noto-fonts openshadinglanguage pahole python-aiohttp python-ytmusicapi
If there're remaining issues, post the entire context.
pacman … 2>&1 | tee /dev/stderr > /tmp/pacman.session
should™ hopefully log everything (idk whetehr the iso ships https://man.archlinux.org/man/core/util … cript.1.en )
You maybe want to look into why and how the system frequently "crashes" on updates.
Offline
Offline
By the way, I have installed Timeshift last time, 2 days ago. Just saying, in case it might help
Last edited by TotoShampoin (2024-03-03 14:39:09)
Offline
The errors are from some alpm hooks (incl. packageshit…) just continue w/ the installation of the packages (ie. same command w/o dbonly)
Offline
Well, the errors were very scary, but I took your word for it, and... They worked so far.
Now, I'm going one last package upgrade in chroot, and hopefully everything work again.
Ah, the update is done now. Let's try and reboot...
Everything is working again!
Thank you Seth, you are a life saver! ;–;
But now, I have to look into what's going on with those update crashes...
In both cases, it crashed at
( 2/16) Reloading system manager configuration...
Last edited by TotoShampoin (2024-03-03 15:35:14)
Offline
https://bugs.archlinux.org/task/77789 ?
sudo /etc/systemd/do-not-udevadm-trigger-on-update
But probably post your pacman.log, or at least a segment covering a succesfull and a bad update to illustrate what's going on at the time.
Thank you Seth, you are a life saver! ;–;
https://en.wikipedia.org/wiki/Set_(deity) has a way less favorable take…
Last edited by seth (2024-03-31 23:35:30)
Offline
/Hi seth,
this kind of crash happened to me last week and today, just now. The end result is that the pacman.log doesn't get updated at all, the pacman cache files are with 0 bytes in /var/cache/pacman and the installed files during the upgrades ends with 0 bytes too. This is quite nasty.
I confirm too that today's crash happened when the "( 2/16) Reloading system manager configuration..." hook message appeared on the screen.
The problem is I don't have a way to reproduce it in a deterministic manner (and I don't have journal logs for this either; I'm using XFS & nvidia drivers if that helps). I'm considering switching to the LTS kernel in the meantime.
I've noticed that this is happening a lot to many people, example: https://www.reddit.com/r/archlinux/comm … perienced/
In fact, I was watching videos on firefox while doing the upgrade so maybe it's related to the nvidia drivers and the 6.7 kernel
Offline
@TotoShampoin, wrt your DM
error: error while reading file /var/lib/pacman/local/python-aiohttp-3.9.3-1/mtree: Unrecognized archive format
altered files
python-aiohttp: no mtree file
python-aiohttp is still broken - the rest looks pretty harmless - but you may have copypasted it out of a terminal? The format is weird at places and there seem stray tokens.
I guess you moved around /boot/intel-ucode.img ?
@reztho
REMOVED: Rule 4. This qualifies as a shitpost.
What was/is the nature of the crash? Insta-reboot or did the system just freeze and you rebooted w/ the power button?
The latter is prone to leave behind corrupted files. Esp. w/ caching SSDs - so that's not overly suspicious then.
If the system spontanously reboots, that's a CPU, temperature or RAM issue.
"Reloading system manager configuration" reloads the systemd daemon:
You two could correlate the output of
find /etc/systemd -type l -exec test -f {} \; -print | awk -F'/' '{ printf ("%-40s | %s\n", $(NF-0), $(NF-1)) }' | sort -f
The next hook would (most likely, check your pacman log - should be same on successful updates) be packagekit. Do you have/use that, too?
Offline
but you may have copypasted it out of a terminal? The format is weird at places and there seem stray tokens.
Ah damnit, that was visible? Sorry for that
error: error while reading file /var/lib/pacman/local/python-aiohttp-3.9.3-1/mtree: Unrecognized archive format altered files python-aiohttp: no mtree file
python-aiohttp is still broken
I saw that, I fixed it
I guess you moved around /boot/intel-ucode.img ?
... Not that I remember? I think I kinda remember doing something along the lines of manually doing "pacman -S intel-ucode", but why would that do something that pacman doesn't expect to see?
the rest looks pretty harmless
Well... Okay then. But what do they indicate?
Offline
You two could correlate the output of
find /etc/systemd -type l -exec test -f {} \; -print | awk -F'/' '{ printf ("%-40s | %s\n", $(NF-0), $(NF-1)) }' | sort -f
Here you go : https://0x0.st/H7j8.status
What was/is the nature of the crash? Insta-reboot or did the system just freeze and you rebooted w/ the power button?
In my case, it's the latter. Both times. So, I guess that means it is a kernel issue, if I'm not the only one experiencing those crashes?
Last edited by TotoShampoin (2024-03-05 07:46:57)
Offline
Not necessarily the kernel, see eg. https://bbs.archlinux.org/viewtopic.php … 5#p2073025
For visibility, you're using
bluetooth.service | bluetooth.target.wants
dbus-org.bluez.service | system
dbus-org.freedesktop.nm-dispatcher.service | system
dbus-org.freedesktop.timesync1.service | system
display-manager.service | system
docker.service | multi-user.target.wants
fstrim.timer | timers.target.wants
gcr-ssh-agent.socket | sockets.target.wants
getty@tty1.service | getty.target.wants
gnome-keyring-daemon.socket | sockets.target.wants
libvirtd-ro.socket | sockets.target.wants
libvirtd.service | multi-user.target.wants
libvirtd.socket | sockets.target.wants
NetworkManager.service | multi-user.target.wants
NetworkManager-wait-online.service | network-online.target.wants
p11-kit-server.socket | sockets.target.wants
pipewire-pulse.socket | sockets.target.wants
pipewire-session-manager.service | user
pipewire.socket | sockets.target.wants
power-profiles-daemon.service | graphical.target.wants
remote-fs.target | multi-user.target.wants
speech-dispatcherd.service | multi-user.target.wants
speech-dispatcher.service | system
supergfxd.service | getty.target.wants
systemd-timesyncd.service | sysinit.target.wants
timeshift.service | multi-user.target.wants
virtlockd.socket | sockets.target.wants
virtlogd.socket | sockets.target.wants
wireplumber.service | pipewire.service.wants
xdg-user-dirs-update.service | default.target.wants
You could also try to explpicitly reload the daemon and see whether that shows any roadbumps.
Offline
Not necessarily the kernel, see eg. https://bbs.archlinux.org/viewtopic.php … 5#p2073025
For visibility, you're usingbluetooth.service | bluetooth.target.wants dbus-org.bluez.service | system dbus-org.freedesktop.nm-dispatcher.service | system dbus-org.freedesktop.timesync1.service | system display-manager.service | system docker.service | multi-user.target.wants fstrim.timer | timers.target.wants gcr-ssh-agent.socket | sockets.target.wants getty@tty1.service | getty.target.wants gnome-keyring-daemon.socket | sockets.target.wants libvirtd-ro.socket | sockets.target.wants libvirtd.service | multi-user.target.wants libvirtd.socket | sockets.target.wants NetworkManager.service | multi-user.target.wants NetworkManager-wait-online.service | network-online.target.wants p11-kit-server.socket | sockets.target.wants pipewire-pulse.socket | sockets.target.wants pipewire-session-manager.service | user pipewire.socket | sockets.target.wants power-profiles-daemon.service | graphical.target.wants remote-fs.target | multi-user.target.wants speech-dispatcherd.service | multi-user.target.wants speech-dispatcher.service | system supergfxd.service | getty.target.wants systemd-timesyncd.service | sysinit.target.wants timeshift.service | multi-user.target.wants virtlockd.socket | sockets.target.wants virtlogd.socket | sockets.target.wants wireplumber.service | pipewire.service.wants xdg-user-dirs-update.service | default.target.wants
You could also try to explpicitly reload the daemon and see whether that shows any roadbumps.
Thank you for answering, Seth. Yeah, I had to poweroff with the power button so that explains the 0 byte files.
Here's what we have in common:
bluetooth.service | bluetooth.target.wants
dbus-org.bluez.service | system
dbus-org.freedesktop.nm-dispatcher.service | system
display-manager.service | system
fstrim.timer | timers.target.wants
gcr-ssh-agent.socket | sockets.target.wants
getty@tty1.service | getty.target.wants
gnome-keyring-daemon.socket | sockets.target.wants
libvirtd-ro.socket | sockets.target.wants
libvirtd.service | multi-user.target.wants
libvirtd.socket | sockets.target.wants
NetworkManager.service | multi-user.target.wants
NetworkManager-wait-online.service | network-online.target.wants
p11-kit-server.socket | sockets.target.wants
pipewire-pulse.socket | sockets.target.wants
pipewire-session-manager.service | user
pipewire.socket | sockets.target.wants
remote-fs.target | multi-user.target.wants
virtlockd.socket | sockets.target.wants
virtlogd.socket | sockets.target.wants
wireplumber.service | pipewire.service.wants
xdg-user-dirs-update.service | default.target.wants
(I did your command plus comm -1 -2 with the output in my computer and their output)
(About the reddit link, that poor folk had the exact same issue whilst updating with pacman)
I'm, for now, on LTS kernel to see if I avoid the issue again and I don't use PackageKit.
Last edited by reztho (2024-03-05 16:50:03)
Offline
From that list, libvirt looks at me.
Otherwise I'd almost have said supergfxd.service looks like a natural contender.
Can either of you cause the freeze by simply reloading the daemon? ("sudo systemctl daemon-reload")
I'd suggest to record your complete service list, it might be the same impact from different causes (eg. if you're using sth. like bumblebee which is damn close to supergfxd)
Offline
Hi Seth, thank you. I don't use bumblebee, just pure nvidia prime (not reverse, but direct) configuration (using XFCE but apart from that, the only nvidia things I have active with systemd are nvidia-resume/nvidia-suspend for suspend / hibernation). As a matter of fact, before your last comment, I uninstalled libvirt too just in case since it looks I had it installed without any use here.
Anyway, here's my full service list:
acpid.service | multi-user.target.wants
auditd.service | multi-user.target.wants
bluetooth.service | bluetooth.target.wants
chronyd.service | multi-user.target.wants
clamav-freshclam-once.timer | timers.target.wants
cups.path | multi-user.target.wants
cups.service | multi-user.target.wants
cups.service | printer.target.wants
cups.socket | sockets.target.wants
dbus-org.bluez.service | system
dbus-org.fedoraproject.FirewallD1.service | system
dbus-org.freedesktop.home1.service | system
dbus-org.freedesktop.NetworkManager.service | system
dbus-org.freedesktop.nm-dispatcher.service | system
dbus-org.freedesktop.resolve1.service | system
dbus-org.freedesktop.thermald.service | system
dbus.service | system
dbus.service | user
display-manager.service | system
earlyoom.service | multi-user.target.wants
firewalld.service | multi-user.target.wants
fstrim.timer | timers.target.wants
gcr-ssh-agent.socket | sockets.target.wants
getty@tty1.service | getty.target.wants
gnome-keyring-daemon.socket | sockets.target.wants
linux-modules-cleanup.service | basic.target.wants
NetworkManager.service | multi-user.target.wants
NetworkManager-wait-online.service | network-online.target.wants
nvidia-resume.service | systemd-hibernate.service.wants
nvidia-resume.service | systemd-suspend.service.wants
nvidia-suspend.service | systemd-suspend.service.wants
p11-kit-server.socket | sockets.target.wants
pipewire-pulse.socket | sockets.target.wants
pipewire-session-manager.service | user
pipewire.socket | sockets.target.wants
reflector.timer | timers.target.wants
remote-fs.target | multi-user.target.wants
smartd.service | multi-user.target.wants
systemd-homed.service | multi-user.target.wants
systemd-resolved.service | multi-user.target.wants
systemd-userdbd.socket | sockets.target.wants
thermald.service | multi-user.target.wants
tlp.service | multi-user.target.wants
wireplumber.service | pipewire.service.wants
xdg-user-dirs-update.service | default.target.wants
And no, I tried to do the daemon-reload a lot yesterday after the latest crash to see if I trigger the failure, but nothing.
Last edited by reztho (2024-03-05 18:36:58)
Offline
I tried to do the daemon-reload a lot yesterday after the latest crash
Might be coincidental. Is this a very recent problem?
Try to add "numa=off" to the kernel parameters, https://lore.kernel.org/all/20240216111 … ul@sk.com/ can pot. throw a lot of things off
Offline
I tried to do the daemon-reload a lot yesterday after the latest crash
Might be coincidental. Is this a very recent problem?
Try to add "numa=off" to the kernel parameters, https://lore.kernel.org/all/20240216111 … ul@sk.com/ can pot. throw a lot of things off
Alright, I'll do a little sacrifice and I'll be back on kernel 6.7.8 with numa=off for a while. Let's see how it goes now without libvirt too.
Edit: Regarding how recent these events are, at least for me, first crash was last Wednesday and second one yesterday.
Last edited by reztho (2024-03-05 23:47:01)
Offline
It went well for a week or so, and since I didn't want more of these random hangs everytime when I update with pacman, I decided to go with the LTS kernel... now guess what, just now, it happened to me too again (kernel LTS 6.6.21)... the only packages being installed were ghostscript and ostree. I had firefox and a video on the background too... not sure if that's the pattern for me.
Last edited by reztho (2024-03-16 11:29:50)
Offline
LTS kernel + nvidia 550xx?
Tried LTS + nvidia 535xx/545xx?
Offline
nvidia dkms 550.54.14-5 & kernel 6.6.21-1-lts, yes. Haven't tried any other combination.
Right now, I tried to setup the systemd-pstore service to see if in next hang, this gives me any info. Let's see.
Last edited by reztho (2024-03-16 11:29:24)
Offline
You most likely want to downgrade to 545 or 535 dkms
Offline
Hi seth, I've been experiencing the same issue in my system -3 times this year already- (in my case my kernel panics, confirmed since the capslock key was blinking https://bbs.archlinux.org/viewtopic.php?id=294216). The 2 coincidences I see with this post is 1) nvidia drivers (using nvidia package with latest 550 release) and 2) yay update calls pacman.
I do want to look into the why my system crashes on updates.
"The problem with quotes on the Internet is that it is hard to verify their authenticity." ~ Abraham Lincoln
Offline
Hi gugah, could you run this command line and post the output here? Thanks. I'm suspecting of the pacman hook for running systemd tmpfiles "cleaning" currently.
for i in /usr/lib/tmpfiles.d/*.conf; do pacman -Qo $i ; done | cut -d ' ' -f5 | sort -u
Last edited by reztho (2024-03-28 22:13:58)
Offline
reztho:
$ for i in /usr/lib/tmpfiles.d/*.conf; do pacman -Qo $i ; done | cut -d ' ' -f5 | sort -u
alsa-utils
audit
dbus
dhcpcd
filesystem
geoclue
kio-fuse
libutempter
man-db
mkinitcpio
openssh
pam
sddm
shadow
sudo
systemd
tpm2-tss
udisks2
util-linux
(couldn't reproduce the bug yet, running paru now instead of yay)
"The problem with quotes on the Internet is that it is hard to verify their authenticity." ~ Abraham Lincoln
Offline