GUI Boot Failure on Ubuntu System with NVIDIA GPU

Graphics / Linux Linux
1

System Configuration:

  • Operating System: Ubuntu 20.04.1 LTS
  • Kernel Version: 5.15.0-58-generic
  • GPU: NVIDIA GeForce RTX 2060
  • NVIDIA Driver Version: 530.41.03
  • CUDA Version: 12.1

Problem Description: I am experiencing a persistent issue where my Ubuntu system fails to boot in GUI mode but boots successfully in CLI recovery mode. This issue has emerged despite the driver and kernel combination previously functioning without issues. The system is configured to dual boot with Windows, which operates without any problems, suggesting that the hardware is functioning correctly.

Symptoms:

  • The system hangs when attempting to boot into GUI mode.
  • It is possible to initiate an xterm session through startx from CLI mode, indicating that some graphical operations are possible.
  • The failure occurs regardless of whether the system is booted normally into Ubuntu or resumed from a recovery session to GUI mode.

Troubleshooting Steps Taken:

  1. Reinstalled NVIDIA drivers and reconfigured them multiple times.
  2. Checked and monitored system logs for errors related to NVIDIA drivers and Xorg.

Critical Errors Observed:

  • Repeated failures related to the NVIDIA kernel module initialization.
  • Xorg log files indicate issues with device recognition and configuration that seem to stem from NVIDIA driver interactions.

I would appreciate any guidance to resolve this issue, recommendations for further diagnostic tests or configuration adjustments.

Thank you.
script_june-19-2024_58_GUImode_systemGPU.txt (44.4 KB)

2

Please run nvidia-bug-report.sh as root and attach the resulting nvidia-bug-report.log.gz file to your post.

3

nvidia-bug-report.log.gz (305.4 KB)

4

Unfortunately, there’s no log left from a non-recovery mode boot. The nvidia driver seems to be working fine, though. Did you set “nomodeset” kernel parameter? If so, please remove it. Otherwise, please try to create an nvidia-bug-report.log from a normal boot.

5

Getting the log left from a non-recovery mode boot in my case is tricky because the system hangs. So I did the following in /etc/systemd/system/diagnostic_nvidia.service:

[Unit]
Description=Run diagnostics script at boot
Before=gdm.service

[Service]
Type=oneshot
ExecStart=/usr/bin/nvidia-bug-report.sh
RemainAfterExit=yes

[Install]
WantedBy=multi-user.target

and I execute the process as follows :

#!/bin/bash

Define the log file and its location

LOG_FILE=“/tmp/foo/x/dual-june-29-2024/nvidia-log_GUI.txt”

Function to log the execution of commands

log_command() {
echo “Running: $@” >> $LOG_FILE
echo “Output:” >> $LOG_FILE
sudo $@ >> $LOG_FILE 2>&1
echo “-------------------------------------------” >> $LOG_FILE
}

Start logging

echo “System Diagnostic nvidia Log for $(date)” > $LOG_FILE
#log_command “/usr/bin/nvidia-bug-report.sh”
log_command “systemctl enable diagnostic_nvidia.service”
log_command “systemctl start diagnostic_nvidia.service”

Complete logging

echo “Diagnostic log completed on $(date)” >> $LOG_FILE

nvidia-bug-report.log.gz (311.3 KB)

6

Any progress on the latest log file I uploaded on 6-30-2024?

7

In my case switching to DP solved the issue. HDMI does not seem to work with 555.58 for me.

Fedora 40 Silverblue, 555.58, 3080ti

8

“nomodeset” set or not set makes no difference