Hi. After upgrading my tumbleweed to 20230526, my PC fails to boot using LightDM. It boots to tty1 prompt where I logged in and run startxfce4 manually.
It’s been a while since I played with DMs etc to remember what to look at. So I did an ‘unconditional update’ on the XFCE4 pattern (which included the LightDM package) but that didn 't help.
I checked the Alternatives which still has lightdm as the default display manager.
When you log in on tty, is lighdm process present (ps -ef | grep lightdm)?
Well, it shows that lightdm was at least started during boot
May 28 13:46:25 asus-roc.lan dbus-daemon[917]: [system] Activating via systemd: service name='org.freedesktop.Accounts' unit='accounts-daemon.service' requested by ':1.14' (uid=0 pid=1517 comm="/usr/sbin/lightdm")
But there are no further messages from it. There are also no usual crash or display manager stop messages so the next step is to determine whether lightdm just sits there doing nothing or it was terminated.
chris@asus-roc:~>$mpv https://www.youtube.com/watch?v=ZU2Roq-emxw
(+) Video --vid=1 (*) (vp9 1920x1080 25.000fps)
(+) Audio --aid=1 --alang=eng (*) (opus 2ch 48000Hz)
File tags:
Uploader: Odyssey - Ancient History Documentaries
Channel_URL: https://www.youtube.com/channel/UCV-yCAXMS-nWrHtORuDfSuQ
libEGL warning: DRI2: failed to authenticate
[vo/gpu/opengl] Suspected software renderer or indirect context.
[vo/gpu/drm] VT_GETMODE failed: Inappropriate ioctl for device
[vo/gpu/drm] Failed to set up VT switcher. Terminal switching will be unavailable.
[vo/gpu/drm] Listing DRM devices with drmGetDevices failed! (No such file or directory)
[vo/gpu/drm] Failed to find a usable DRM primary node!
Segmentation fault (core dumped)
chris@asus-roc:~>$
There seems to be something more fundamentally wrong with tumbleweed that just maybe LightDM.
I’m now looking at doing a snapshot restore.
chris@asus-roc:~>$sudo snapper list
[sudo] password for root:
# | Type | Pre # | Date | User | Cleanup | Description | Userdata
----+--------+-------+--------------------------+------+---------+-----------------------+--------------
0 | single | | | root | | current |
1* | single | | Fri 08 Feb 2019 12:53:42 | root | | first root filesystem |
51 | pre | | Wed 17 May 2023 16:44:47 | root | number | zypp(zypper) | important=yes
52 | post | 51 | Wed 17 May 2023 16:45:21 | root | number | | important=yes
55 | pre | | Sun 28 May 2023 11:31:21 | root | number | zypp(zypper) | important=yes
56 | post | 55 | Sun 28 May 2023 11:42:44 | root | number | | important=yes
57 | pre | | Sun 28 May 2023 11:45:12 | root | number | zypp(zypper) | important=yes
58 | post | 57 | Sun 28 May 2023 11:45:53 | root | number | | important=yes
81 | pre | | Sun 28 May 2023 13:40:10 | root | number | yast sw_single |
82 | pre | | Sun 28 May 2023 13:40:37 | root | number | zypp(ruby.ruby3.2) | important=no
83 | post | 82 | Sun 28 May 2023 13:45:20 | root | number | | important=no
84 | post | 81 | Sun 28 May 2023 13:45:39 | root | number | |
89 | pre | | Sun 28 May 2023 14:05:47 | root | number | yast alternatives |
90 | post | 89 | Sun 28 May 2023 14:06:05 | root | number | |
91 | pre | | Sun 28 May 2023 14:10:44 | root | number | yast snapper |
chris@asus-roc:~>$
I started the zypper dup at around 11:30 today so I guess I restore #55? chris@asus-roc:~>$sudo snapper rollback 55 ???
Educated guess - you do not have any KMS capable device/driver and are using vesafb console driver so systemd-logind waits for suitable DRM device and lightdm waits for systemd-logind to announce such device.
Does adding nomodeset to kernel command line help?
Sorry but that was all Greek to me.
But I will add the nomodeset command at boot and see if that works. I’m currently doing an extra backup of my data files via rsync so this may take a while before I can respond.
But I have a bad feeling in the pit of my stomach…
Hi, I’m on a xfce in tumbleweed. I just did a zypper dup 15 minutes ago and LightDM seems all good here. Did you check if your default default system target is graphical interface in yast2-services manager.
It happened again - same symptoms. And snapper rollback saved my backside.
I’m am stumped as to what is going on as I’ve never had real issues with zypper dup for a number of years. I had a look at zypper.log but nothing stands out - not that I know what to look for other than error worded text.
EDIT: I did notice when my system was in this ‘corrupted’ mode that there seemed to be Nvidia Card issues with one the Firestorm application with the message -
2023-05-28T08:09:01Z WARNING # newview/lltoastalertpanel.cpp(201) LLToastAlertPanel::LLToastAlertPanel : Alert: Display settings have been set to recommended levels because your graphics card changed
from 'NVIDIA Corporation NVIDIA GeForce GTX 1050 Ti/PCIe/SSE2'
to 'Mesa llvmpipe (LLVM 16.0.2, 256 bits)'
Could this be an Nvidia driver issue? These are my drivers -
The dup for 20230526 includes an update for the kernel to version 6.3.2-1.1 but the Nvidia driver is only nvidia-gfxG05-kmp-default | package | 470.182.03_k6.3.1_1-54.4.
Could it be that the Nvidia driver hasn’t caught up with the kernel version yet?
NVIDIA driver should be compiled during installation and the version you have should include kernel 6.3 build fix. So either this fix does not work or module was not compiled. Show as root
This is from my rollbacked snapshot but I guess you want from the “faulty” zypper dup. I can do another upgrade tomorrow as I need my PC to be 100% ok for a few things apps for the next 24 hours.
I saw your driver is 470.x.x. I used to have an older nvidia graphic card that uses that version.
I started to have a problem installing that driver every new kernel lately so I upgraded my graphic card
to avoid that kind situation.
If you are still on this installation - can you show
zypper -si kernel-default
ls -l /usr/src/linux-obj/x86_64
ls -l /lib/modules/6.3.1-2-default/updates/nvidia.ko
Driver should be compiled when kernel-default-devel package is installed. It depends on the link created by installation. I have seen cases when this link was incorrect.
I see the kernel has been updated to 6.3.4 today. Will run the zypper dup again tonight.
But doesn’t the nvidia driver kmp file need to follow the kernel version? It’s still k6.3.1. nvidia-gfxG05-kmp-default | package | 470.182.03_k6.3.1_1-54.4
