0

I am quite new to Ubuntu. For a month, my system has frozen 2 out of every 3 times when it reached the login screen.

Dual boot with GRUB was fine. Keyboard was responsive and I could switch between Ubuntu and Windows.

With "nomodeset", all the kernel messages were invoked and carried out neatly. After that, four scenarios happen:

  • A pitch black screen.
  • A purple screen is shown but no login box follows. No mouse is found.
  • The purple screen showed up but no login box was seen. Mouse is seen but either stuck or moveable.
  • The purple screen and a login box is shown. I typed my password, hit enter, saw the login box disappear and the monitor was stuck with the purple screen.

Of all cases above, no error message appears. I could not access TTY.

At first, I thought there was something wrong with GNOME display manager. I reinstalled gdm3 and the freeze still persisted. I switched to lightdm, successfully got past the login screen and found the error message below.

Found when loading /etc/profile:
:: ERROR: Unable to proceed: possible causes listed below.
This script must be sourced. Did you execute or source this script?
Unrecognized/supported shell (supported: bash, zsh, ksh, m/lksh, dash).
May fail in dash if you rename this script (assumes "setvars.sh").
Can be caused by sourcing from ZSH version 4.x or older.
As a result the session will not be configured correctly. 
You should fix the problem as soon as feasible.

I tried:

$bash /etc/profile 

and was hit with another error:

:: initializing oneAPI environment ...
   profile: BASH_VERSION = 5.0.17(1)-release
:: ERROR: No env scripts found: No "env/vars.sh" scripts to process.
   This can be caused by a bad or incomplete "--config" file.
   Can also be caused by an incomplete or missing oneAPI installation.

My machine a Dell Precision Tower 5810, with Intel Xeon E5-1607 v4 and NVIDIA Quadro P600. Dell stated no support for Ubuntu 18.04 and 20.04 on Tower 5810. A month or two before the freeze first happened, I removed cuda11.5, installed cuda11.6 using APT packages and installed Intel oneAPI using Intel oneAPI Base Toolkits. Intel oneAPI has its own version of setvars.sh. I believe that was the most recent installation that had anything to do with configuring the hardware. The full content of /etc/profile is posted below:

# /etc/profile: system-wide .profile file for the Bourne shell (sh(1))
# and Bourne compatible shells (bash(1), ksh(1), ash(1), ...).

if [ "${PS1-}" ]; then
  if [ "${BASH-}" ] && [ "$BASH" != "/bin/sh" ]; then
    # The file bash.bashrc already sets the default PS1.
    # PS1='\h:\w\$ '
    if [ -f /etc/bash.bashrc ]; then
      . /etc/bash.bashrc
    fi
  else
    if [ "`id -u`" -eq 0 ]; then
      PS1='# '
    else
      PS1='$ '
    fi
  fi
fi

if [ -d /etc/profile.d ]; then
  for i in /etc/profile.d/*.sh; do
    if [ -r $i ]; then
      . $i
    fi
  done
  unset i
fi
3
  • Hello. There is no version of Ubuntu 18.4 or 20.4 The first set of errors you give are usually related to wrong video driver. It seems you may have been removing things without a good reason to do so. From your question < I removed cuda11.5, installed cuda11.6 > is there a reason you did this?
    – David
    Apr 11, 2022 at 14:19
  • You may find this step by step guide helpful. tutorialforlinux.com/…
    – David
    Apr 11, 2022 at 14:21
  • @David I mean Ubuntu 18.04 and 20.4. My mistake. I was suggested to update cuda version to run a scientific computing software called gromacs. Thank you for your reply.
    – hwa-le
    Apr 11, 2022 at 15:22

0

You must log in to answer this question.

Browse other questions tagged .