Ytdyklly

About half a month ago, I installed Scientific Linux 7.6 on a computer that has shut down and restarted daily as expected until today. Details of the operating system are below.

1. Full version number: (3.10.0-957.1.3.el7.x86_64) 7.6 (Nitrogen)


2. Desktop: Gnome

Today when I turn back on the computer, the gnome desktop won't load. The boot up proceeds as expected until the grey-blue screen showing a spinning wheel (the one with radial lines). Then about when the wheel is supposed to pick up speed and take you to the login page with usernames to select from, it breaks into the terminal mode text with status messages. They all show [ OK ] in green except for these lines repeated once:

[ INFO ] dev-virtiox2dports-com.redhat.spice.0.device is not active.

[DEPEND] Dependency failed for Activation socket for spice guest agent daemon

.....

Also the last line (before the display manager load hangs) reads thus:

  OK  ] Started Crash recovery kernel arming.. to Quit.........ice.emon.549..progress polling....

As you can see the messages there are garbled and the [ OK ] label is missing the leading bracket too.

It looks horrid so far, but I am able to navigate to a login screen and appropriate a terminal with (Alt,F2) through (Alt, F6), and then log in as one of the users or as root, likewise.

So I am hoping that more than the window manager is not malfunctioning. When I check on the status of gdm.service, I find that it shows as 'active (running)' but there are also these messages.

<Datetime> <hostname> systemd[1]: Starting GNOME Display Manager...

<Datetime> <hostname> systemd[1]: Started GNOME Display Manager.

<Datetime> <hostname> gdm[5517]: GdmDisplay: display lasted <n> seconds

.....

<Datetime> <hostname> gdm[5517]: GdmLocalDisplayFactory: maximum number of X display failures reached: check X server log for errors

The third line repeats five more times with different values for n seconds.

If it is relevant, I am using the nouveau driver on an Nvidia GTX GeForce 660 (OEM) card, and there is also an Intel on-board graphics cotroller driven by the i915 module, as informed by the output of the 'lspci -k' command.

My questions are:
(1) Is my hunch that only the window manager has crashed and the rest of the system is still functioning correct?
(2) Where can I find the X server log cited above?
(3) I read that it is possible to reinstall the display manager on distributions other than Scientific Linux and want to know if that is a possibility here?
(4) Is there a way to recover display functionality without taking the step of reinstallation if that is an option?

Thank you many times.

I've been able to restore access to the Gnome desktop by black listing the i915 driver for the Intel (on-board) controller.

I read (1) that this might result in interference, and though the context there is different, I tried it. These are the commands I ran in a terminal window.

vi /etc/modprobe.d/blacklist.conf

mv /boot/initramfs-$(uname -r).img /boot/initramfs-Intel-i915-$(uname -r).img

dracut -v /boot/initramfs-$(uname -r).img $(uname -r)

reboot

When the machine restarts it proceeds beyond the blue-grey screen and the spinning wheel right through to the login screen for the Gnome desktop this time.

I don't know if this is the last of the problem, since I am still using the nouveau driver, which is reported here and there as being unreliable, but for now the problem has been evaded temporarily at least. Any comments on the case would still be helpful.