Login loop after upgrade to 16.04

Question

I encountered a login loop after updating xenial beta to the latest updates.

I tried a lot of things:

• xauthority or /tmp privileges change

• Reinstalling nvidia drivers, removing them.

• reinstalling unity, unity-greeter, ubuntu-desktop, upstart, lightdm and compiz.

• resetting compiz and unity settings.(I get dbus launch errors:

  Failed to commit changes to dconf: error spawning command line 'dbus-launch -auto launch=..................)
  

• using older kernel and failsafe (failsafe report no screen found)

• tried to launch other DE (steam big picture)

Here's xsession errors log:

unity:
open connection: connect: No such file or directory
Cannot connect to brltty at :0
upstart: gnome-session (unity) main process (5441) terminated with status 1
upstart unity-settings-daemon main process (5434) killed by TERM signal

steam big picture:
open connection: connect: No such file or directory
Cannot connect to brltty at :0
ERROR: ld.so: object '/usr/lib/i386-linux-gnu/libmodeswitch_inhibitor.so' from LD_PRELOAD cannot be preloaded (cannot open shared object file(: ignored.
ERROR: ld.so: object '/usr/lib/i386-linux-gnu/libmodeswitch_inhibitor.so' from LD_PRELOAD cannot be preloaded (cannot open shared object file(: ignored
ERROR: ld.so: object '/usr/lib/i386-linux-gnu/libmodeswitch_inhibitor.so' from LD_PRELOAD cannot be preloaded (cannot open shared object file(: ignored
Could not create GLX context
ERROR: ld.so: object '/usr/lib/i386-linux-gnu/libmodeswitch_inhibitor.so' from LD_PRELOAD cannot be preloaded (cannot open shared object file(: ignored

In Unity (compiz) log there are a lot of errors about opengl plugin

Compiz (opengl) - Fatal: glXQueryExtenisonsString is NULL for screen 0

And a lot of plugins unloaded because of opengl.

Lightdm have sudden SIGTERM signal with no prior errors in the log.

xserver have 2 errors:

xf86EnablePorts: failed to set IOPL for I/O (Operation not permitted)
(EE) Failed to initialize GLX extension (Compatible NVIDIA X driver not found)

But it continue through input operations until it suddenly get terminated.

Answer 1

In my case original nvidia drivers was the cause and switching back to open source nouveau helped. First, I removed all stuff from nvidia:

sudo apt-get remove --purge 'nvidia-*'
sudo apt-get install ubuntu-desktop
sudo rm /etc/X11/xorg.conf
echo 'nouveau' | sudo tee -a /etc/modules

(source: How can I uninstall a nvidia driver completely ?)

After that I could login, but unity didn't load, there was no launcher etc. It turn out I had to re-enable Unity plugin:

sudo apt-get install compizconfig-settings-manager
DISPLAY=:0 ccsm
Find the Unity plugin and enable it.

(source: Unity doesn't load, no Launcher, no Dash appears)

Answer 2

I had the same error on 16.04 as well. In my case it was also an NVidia driver issue. I have 8 monitors and I like how the NVidia diver performs with them. So to get it working, I:

1. Pressed CTRL + ALT + F1 at the login prompt and logged into through terminal.

2. Backed up may old /etc/X11/xorg.conf file, as it was a pain to configure 8 monitors.

3. Downloaded the driver file from NVidia: NVIDIA-Linux-x86_64-367.44.run

4. Ran the following command: sudo service lightdm stop (driver install will fail if this step is skipped).

5. Reinstalled the driver: sudo ./NVIDIA-Linux-x86_64-367.44.run I was notified that the auto generated script did not complete and given the choice to continue the install which I accepted. I also chose to have the driver installed as a kernel module and have the program generate a new xorg.conf file (backed mine up before). At the end of the process it complained about a missing link to a lib, but gave instructions on linking to it. I put off dealing with that for later.

6. Rebooted and was able to login through the GUI.

Answer 3

The problem is that, when you update Ubuntu (and even when you install certain programs that call /sbin/ldconfig), there are changes made to /etc/ls.so.cache.

Try it and you'll see that the ld.so.cache file after such an installation/update will have a different size than before the operation.

This, in Ubuntu 16.04, results in a login loop when logging in from lightDM. Therefore, you should keep a copy of the original ld.so.cache since, with the new version generated after the installation/update, lightDM will no longer authenticate.

I have had this issue many times and have had to be very careful with this. I always keep a copy of the original file (/etc/ld.so.cache.orig) that lets me log in without the login loop and use that to log in. Then, once I have logged in successfully, I copy the new one (/etc/ld.so.cache.new) over. To do so, I have modified my ~/.bash_profile file and have added this line:

sudo /bin/cp /etc/ld.so.cache.new /etc/ld.so.cache

Then, in order to copy the original file back when exiting the session so that I'll be ab;e to log ina gain next time, I have added this line to /etc/gdm/PostSession/Default:

/bin/cp /etc/ld.so.cache.orig /etc/ld.so.cache

For that to work, I have modified /etc/sudoers (by running sudo vosudo) to allow non privileged users to run the copy command:

ALL     ALL= NOPASSWD: /bin/cp /etc/ld.so.cache.new /etc/ld.so.cache

This might look like an ugly hack, but it gets around this login-loop issue and, also, to keep the new /etc/ld.so.cache file.

Answer 4

I was using xfce and I manually checked each folder and file by renaming them. Deleting ~/.config/xfce4/xfconf/xfce-perchannel-xml/displays.xml worked for me.

I had nvidia drivers as well, but they were working fine when I logged in as root and a backup user.

Answer 5

In my case this issue was caused by accidentially enabling Secure Boot in UEFI, which effects in third party drivers unusable (e.g. Nvidia). That caused X session error. After disabling Secure Boot, everything went back to normal.