Help to troubleshoot...Freeze before desktop
Help to troubleshoot...Freeze before desktop
Hi guys! I have KDE Neon 6.2.3 running on my Surface Pro 7+, and up until now it's been running rather smoothly, for years. But today I seem to be unable to fully boot anymore on normal mode. Linux will get stuck on the KDE spinning gear, with the gear frozen. And it becomes unresponsive, to open alternative consoles with control+alt+Fx, or even control+alt+del. The only thing that works is a force shutdown by holding the power button.
I tried with a previous kernel, with the same results (current is linux-image-6.10.10-surface-1, I have a generic 6.8.0-49, but it also freezes). I can boot in safe mode. It opens a different login screen, and after a failed attempt of opening lxde, it fails back to my usual KDE session, in safe graphics mode.
What can I test? How can I see what failed at boot?
You're viewing a single thread.
When you've booted in safe mode you can use journalctl to check what's going on.
https://www.howtogeek.com/499623/how-to-use-journalctl-to-read-linux-system-logs/
https://www.man7.org/linux/man-pages/man1/journalctl.1.htmlThanks. I pulled the journalctl. This is...A bit extensive. Just the entries from today are about 4MB in text. I'm having a very hard time finding the moment I force it off and then start again. Any good troubleshooting guide on how to find crashes?
You need to specify the -b switch or you're just looking at logs for the current session. Something like
journalctl -xe -b 1should do the trick.Just an update. Today I made a rescuezilla full backup of the whole drive, and then wiped and installed KDE Neon brand new from Dec 1st release from their page. Guess what...that one ALSO fails to boot.
OK so just to confirm: you get a boot freeze even on a fresh install before you restore a backup?
Are you able to boot a liveusb?
I would try another distro just to rule out a software problem but it sounds like hardware failure. I would test your ram, graphics card, and storage medium. Do you have an SSD or an HDD?
I actually clicked on the replace partition option on the installer, instead of a full disk wipe. The profile did look brand new once I managed to login in safe mode, but I didn't stay too long to check, as I just concluded hell, if it only boots in safe mode I might as well continue troubleshooting on the original install.
Anyway...After going back to the original image (as I had it backed up anyway), I paid a bit more attention at that weird popup on safe mode as I logged in, mentioning it failed to launch startlxde. So I ran
sudo dpkg --reconfigure sddmAnd on the menu, for some reason sddm wasn't the first choice. After choosing it again, this time it went straight to the boot screen. Some other day I'll check again whichever random login menu is installed, and remove it, as it clearly causes issues.
Ah ha! Problem solved then. Nice troubleshooting skills!
Oh, and before trying fixes - make a backup of your precious files when booted in safe mode.
About that - I´d like to run a backup that takes all apps and their configuration files (more than my own docs and files, which I already have properly backed in Seafile). What would you recommend just for the apps, their configs (and maybe the repos) etc?
I think you're gonna find that a hard task. There isn't one universal directory where apps saves their configs and bin files. You'll find some in ~/.config
You can use the -b flag to get a specific boot sequence:
https://www.howtogeek.com/499623/how-to-use-journalctl-to-read-linux-system-logs/#reviewing-boot-messagesSometimes it can help to filter by UID 1000:
https://www.howtogeek.com/499623/how-to-use-journalctl-to-read-linux-system-logs/#selecting-log-messages-by-data-fieldsOther than that you can try to filter by KDE services, but as I run GNOME I'm not very familiar with their names. In the KDE forum a developer recommends using this for another problem:
journalctl --user-unit=plasma-plasmashell.service --boot 0 --no-pager
https://discuss.kde.org/t/black-screen-after-login-how-to-troubleshoot-plasma-desktop/1615/4Dec 03 16:27:12 SurfaceNeon wpa_supplicant[1436]: wlp0s20f3: CTRL-EVENT-REGDOM-CHANGE init=DRIVER type=COUNTRY alpha2=HK Dec 03 16:27:12 SurfaceNeon NetworkManager[1428]: <info> [1733214432.6906] manager: startup complete Dec 03 16:27:12 SurfaceNeon systemd[1]: Finished NetworkManager-wait-online.service - Network Manager Wait Online. Dec 03 16:27:12 SurfaceNeon systemd[1]: Reached target network-online.target - Network is Online. Dec 03 16:27:12 SurfaceNeon systemd[1]: Started update-notifier-download.timer - Download data for packages that failed at package install time. Dec 03 16:27:12 SurfaceNeon systemd[1]: Started update-notifier-motd.timer - Check to see whether there is a new version of Ubuntu available. Dec 03 16:27:12 SurfaceNeon systemd[1]: Reached target timers.target - Timer Units. Dec 03 16:27:12 SurfaceNeon systemd[1]: Started cups-browsed.service - Make remote CUPS printers available locally. Dec 03 16:27:12 SurfaceNeon systemd[1]: Started deluged.service - Deluge Bittorrent Client Daemon. Dec 03 16:27:12 SurfaceNeon systemd[1]: Reached target multi-user.target - Multi-User System. Dec 03 16:27:12 SurfaceNeon systemd[1]: Reached target graphical.target - Graphical Interface. Dec 03 16:27:12 SurfaceNeon systemd[1]: Starting systemd-update-utmp-runlevel.service - Record Runlevel Change in UTMP... Dec 03 16:27:12 SurfaceNeon systemd[1]: systemd-update-utmp-runlevel.service: Deactivated successfully. Dec 03 16:27:12 SurfaceNeon systemd[1]: Finished systemd-update-utmp-runlevel.service - Record Runlevel Change in UTMP. Dec 03 16:27:12 SurfaceNeon systemd[1]: Startup finished in 4.484s (firmware) + 6.837s (loader) + 3.208s (kernel) + 11.614s (userspace) = 26.145s. Dec 03 16:27:12 SurfaceNeon systemd[1]: dmesg.service: Deactivated successfully. Dec 03 16:27:14 SurfaceNeon ModemManager[1454]: <msg> [base-manager] couldn't check support for device '/sys/devices/pci0000:00/0000:00:14.3': not supported by any plugin Dec 03 16:27:20 SurfaceNeon systemd[1]: NetworkManager-dispatcher.service: Deactivated successfully. Dec 03 16:27:20 SurfaceNeon systemd-logind[1351]: Power key pressed short. Dec 03 16:27:20 SurfaceNeon systemd-logind[1351]: Powering off... Dec 03 16:27:20 SurfaceNeon systemd-logind[1351]: System is powering down.All I can see in that is that your machine has an LTE SIM slot and that the LTE Modem isn't supported. That shouldn't stop you from booting.
Yeah...It looks to me like it just says it went alright, right until I hit the poweroff button, because it's actually completely crashed.
Just an update. Today I made a rescuezilla full backup of the whole drive, and then wiped and installed KDE Neon brand new from Dec 1st release from their page. Guess what...that one ALSO fails to boot.
If a clean install also fails then I would start considering hardware fault.
You could always check if you can boot Fedora Workstation with GNOME just in case but I think you should start looking for a replacement PC.I actually clicked on the replace partition option on the installer, instead of a full disk wipe. The profile did look brand new once I managed to login in safe mode, but I didn't stay too long to check, as I just concluded hell, if it only boots in safe mode I might as well continue troubleshooting on the original install.
Anyway...After going back to the original image (as I had it backed up anyway), I paid a bit more attention at that weird popup on safe mode as I logged in, mentioning it failed to launch startlxde. So I ran
sudo dpkg --reconfigure sddmAnd on the menu, for some reason sddm wasn't the first choice. After choosing it again, this time it went straight to the boot screen. Some other day I'll check again whichever random login menu is installed, and remove it, as it clearly causes issues.
Sweet, then you know what's going on and solved it!
You might wanna try out if your distro is compatible with cockpit:
https://cockpit-project.org/
https://github.com/cockpit-project/cockpitIt gives you a webUI that you can use to check out logs and services (among other things) and makes it a lot easier to troubleshoot computer troubles where the machine starts but your GUI doesn't.