Freeze (Ubuntu 25.04 loading screen) on the first loading after installation

[Yosha872]

After having a strange issue on installation (the installer asked me if I wanted to restart, I clicked yes, it closed, but nothing happened. I had to shutdown the live OS myself. Maybe I should have do a 'ubuntu-bug' before rebooting, but by the way I don't know if I had to choose 'installation' or 'installer').

On the first boot of ubuntu 25.04 beta the loading screen (computer logo + ubuntu logo) was like freeze. After waiting 1 min then doing CTRL+ALT+DEL, now the OS does boot fine (tried 3 times).

ProblemType: Bug
DistroRelease: Ubuntu 25.04
Package: ubiquity (not installed)
ProcVersionSignature: Ubuntu 6.14.0-11.11-generic 6.14.0-rc7
Uname: Linux 6.14.0-11-generic x86_64
ApportVersion: 2.32.0-0ubuntu2
Architecture: amd64
CasperMD5CheckResult: pass
CurrentDesktop: ubuntu:GNOME
Date: Fri Mar 28 15:30:36 2025
InstallCmdLine: BOOT_IMAGE=/boot/vmlinuz-6.14.0-11-generic root=UUID=f9796f07-9884-418e-9095-b044d814faaa ro quiet splash crashkernel=2G-4G:320M,4G-32G:512M,32G-64G:1024M,64G-128G:2048M,128G-:4096M
InstallationDate: Installed on 2025-03-28 (0 days ago)
InstallationMedia: Ubuntu 25.04 "Plucky Puffin" - Beta amd64 (20250326.6)
ProcEnviron:
 LANG=fr_FR.UTF-8
 PATH=(custom, no user)
 SHELL=/bin/bash
 TERM=xterm-256color
 XDG_RUNTIME_DIR=<set>
SourcePackage: ubiquity
Symptom: installer
UpgradeStatus: No upgrade log present (probably fresh install)

[Daniel van Vugt (vanvugt)]

First boot is usually slowest. I think that may be related to snapd, but let us know if it happens again.

[Yosha872 (yosha)]

Less than 5 reboots later, the OS was not able to boot to gnome any more. Still stuck on the loading screen (with frozen 'loading' sign).

After reinstalling the OS (and cleaning the partition), waited for the first boot is complete (this laptop a has a LED showing disk activity), 2 or 3 boots later I have the same issues. Same thing: I can not even boot to Gnome any more.

When booting with 'nodemodeset' argument, I can see lot of loops related to the Nvidia GPU (such as '(2 of 3) Job gpu-manager-service/start running (4min 4s / no limit)', but I no idea how to output a log file.

[Daniel van Vugt (vanvugt)]

If you can boot with 'nomodeset' then please run:

  journalctl -b-1 > prevboot1.txt
  journalctl -b-2 > prevboot2.txt
  journalctl -b-3 > prevboot3.txt

to collect the logs of the previous three boots which presumably contains the bug. Then attach the resulting text files here.

Separately, if the problem is caused by the Nvidia driver then please try installing without selecting the 3rd party software option.

[Yosha872 (yosha)]

I have no idea why but today Ubuntu did load. Here are the files asked.

[Daniel van Vugt (vanvugt)]

It seems to be the Nvidia driver (570.86.16) holding up the boot sequence. Maybe it stopped happening when you received the 570.133.07 update?...

---

mars 30 00:07:09 Crawler-E30e (udev-worker)[510]: nvidia: Spawned process '/sbin/modprobe nvidia-drm' [834] is taking longer than 59s to complete.
mars 30 00:07:09 Crawler-E30e systemd-udevd[438]: 0000:01:00.0: Worker [479] processing SEQNUM=3955 is taking a long time
mars 30 00:07:09 Crawler-E30e systemd-udevd[438]: nvidia: Worker [510] processing SEQNUM=4518 is taking a long time
mars 30 00:07:41 Crawler-E30e systemd[1]: nvidia-persistenced.service: start operation timed out. Terminating.
mars 30 00:07:41 Crawler-E30e nvidia-persistenced[1596]: Received signal 15
mars 30 00:07:41 Crawler-E30e nvidia-persistenced[1596]: Shutdown (1597)

---

avril 02 13:02:20 Crawler-E30e (udev-worker)[472]: nvidia: Spawned process '/sbin/modprobe nvidia-drm' [845] is taking longer than 59s to complete.
avril 02 13:02:20 Crawler-E30e systemd-udevd[460]: 0000:01:00.0: Worker [504] processing SEQNUM=3904 is taking a long time
avril 02 13:02:20 Crawler-E30e systemd-udevd[460]: nvidia: Worker [472] processing SEQNUM=4455 is taking a long time
avril 02 13:03:43 Crawler-E30e kernel: usb 3-9: USB disconnect, device number 3
avril 02 13:04:20 Crawler-E30e (udev-worker)[472]: nvidia: Spawned process '/sbin/modprobe nvidia-drm' [845] timed out after 2min 59s, killing.
avril 02 13:04:20 Crawler-E30e (udev-worker)[472]: nvidia: Process '/sbin/modprobe nvidia-drm' terminated by signal KILL.

[Yosha872 (yosha)]

It is the same issue with Nvidia driver 570.133.07 update. But I have something perhaps useful: on Grub instead of selecting 'Ubuntu' (fail rate: 100%), if I go in 'Advanced' then select the first line (not 'recovery mode') the Ubuntu does run almost every time. I attach 3 logs.

Perhaps the issue is with my Nvidia card, but as I have dual boot with Windows 11 where it does run every time, I insist on investigating.

[Daniel van Vugt (vanvugt)]

Three issues there:

1. The i915 kernel driver is crashing(?) on startup (I've seen this in another bug report too):

   i915 0000:00:02.0: [drm] Port B asks to use VBT vswing/preemph tables

2. The Bluetooth kernel driver is taking 20-30 seconds to start up (download firmware?).

3. Again "nvidia: Spawned process '/sbin/modprobe nvidia-drm' [870] is taking longer than 59s to complete." but that could just be caused by #1.

So overall it's probably a kernel bug. Certainly the workaround you mention in comment #7 just sounds like booting an older kernel version that's free of the bug(s).

[Rovano (rovano)]

"BOOT_IMAGE=/boot/vmlinuz-6.14.0-11-generic"

Try update OS. Out is long time 6.14.0-13.

Personally, I haven't had anything like this happen to me even on kernel version 6.14.0-11 or on 6.11 line.
Or with drivers. Im was using 570.86.x and im using 570.133.07 for now.
But I have an old Nvidia GPU card without the new firmware stuff.

The only other difference between us in configuration I see is in the parameters.
I don't use splash and quiet in kernel line.
And for dual boot Win10.