LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Ubuntu (https://www.linuxquestions.org/questions/ubuntu-63/)
-   -   Ubuntu 14.04.1 LTS No GUI On Boot worked in 12.04 (https://www.linuxquestions.org/questions/ubuntu-63/ubuntu-14-04-1-lts-no-gui-on-boot-worked-in-12-04-a-4175528163/)

gkasica 12-13-2014 08:13 AM

Ubuntu 14.04.1 LTS No GUI On Boot worked in 12.04
 
2 Attachment(s)
Upgraded to the 14.04.1 release and now have no GUI display just a blank screen. CTRL-SHIFT-F2 will take me to a prompt where I can kill X and get a login prompt back on tty1.

I'm at a loss how to fix this though. Another system here works fine with the same monitor but different video card.

Here is Video Crd Info from the non-working system
root@saturn:/var/log# lshw -c video
PCI (sysfs)
*-display UNCLAIMED
description: VGA compatible controller
product: KM400/KN400/P4M800 [S3 UniChrome]
vendor: VIA Technologies, Inc.
physical id: 0
bus info: pci@0000:01:00.0
version: 01
width: 32 bits
clock: 66MHz
capabilities: pm agp agp-3.0 vga_controller bus_master cap_list
configuration: latency=32 mingnt=2
resources: memory:f4000000-f7ffffff memory:fa000000-faffffff memory:fb000000-fb00ffff

I'm also going to attach the Xorg.0.log which is the non-working one and working-Xorg.0.log which is obviously the working log.

What is the best way to reconfigure this to get it working again?

widget 12-13-2014 03:20 PM

The first thing to try when having a problem like this is to edit the menu entry from the grub menu screen by hitting the "e" key.

You want to add to the intruction string so it looks something like;
Code:

linux /vmlinuz root=/dev/sda5 ro quiet splash nomodset
The "nomodset" being what you need to add.

Then hit Ctrl + x to boot to the edited entry. This is just temperary for that boot only so it doesn't screw with anything. Many times this will get you to the desktop even if the graphics may not be the best.

Booting to recovery is an easy way to get a tty login but how ever you want to get to it that is the thing to do if that doesn't help.

Check with a command like;
Code:

dmesg | grep via
This may give you some information that is useful.

Using your other box would be an easy way to discover the needed driver for your via controller. Once you have that info you can install it on your ailing box at the tty prompt.

flshope 12-13-2014 08:03 PM

The same thing happened to me after updating a vintage 2003 machine to 14.04. Apparently, 14.04 installed an updated video driver that was incompatible with my old nvidia card. I did the ctl-alt-F1 thing and then used the apt command to remove the new driver and install the nvidia-bumblebee driver. That fixed my problem ... until a recent software update did it again. I am still struggling with the current mess, which gives me a blank screen even on the shell window, though I can execute commands, but cannot see the system's response.

gkasica 12-13-2014 11:31 PM

the nomodset got me the same blank display after the blue Ubuntu spash screen. Couldn't get out of it without a reboot via SSH.

dnesg | grep via shows though I think you want dmesg | grep -i via when its a VIA

root@saturn:~# dmesg | grep -i via
[ 0.191987] pci 0000:00:01.0: disabling DAC on VIA PCI bridge
[ 0.191987] pci 0000:00:11.0: Bypassing VIA 8237 APIC De-Assert Message
[ 0.947744] agpgart: Detected VIA P4M800 chipset
[ 0.950440] agpgart-via 0000:00:00.0: AGP aperture is 32M @ 0xf8000000
[ 1.376904] sata_via 0000:00:0f.0: version 2.6
[ 1.377550] sata_via 0000:00:0f.0: routed to hard irq line 11
[ 1.381437] scsi0 : sata_via
[ 1.381612] scsi1 : sata_via
[ 1.383204] pata_via 0000:00:0f.1: version 0.3.4
[ 1.396769] scsi4 : pata_via
[ 1.398111] via_rhine: v1.10-LK1.5.1 2010-10-09 Written by Donald Becker
[ 1.398120] via_rhine: Broken BIOS detected, avoid_D3 enabled
[ 1.398997] scsi5 : pata_via
[ 1.402639] via-rhine 0000:00:12.0 eth0: VIA Rhine II at 0xfc030000, 00:13:d3:64:87:e1, IRQ 23
[ 1.403439] via-rhine 0000:00:12.0 eth0: MII PHY found at address 1, status 0x786d advertising 01e1 Link c5e1

Using the other box isn't an option as they have different video cards in them.

How did you locate the driver that was getting loaded and then replace it with apt-get - I know how to add and remove but the finding the driver to remove & replace is where I'd be stuck.

SO I'm still in the same place here at this point.

makokikadur 12-14-2014 02:19 PM

In 14.10 a similar bug was introduced.
Don't know if it would be the same with 14.04.1 but you can try hitting ESC twice:
Quote:

Machines with ATI/AMD video cards may be getting blank or entirely turned off screen at boot. The screen will switch on when the X server starts. If you need it to turn on prior to that, for example to unlock an encrypted harddisk hitting ESC twice should reset the video card to a working state.

gkasica 12-14-2014 02:31 PM

Ok. I'll try that but how do I get X to do it's thing and start up?

If I hit ESC on boot after the splash screen I can see the boot sequence up to the point I'm assuming X starts or tries to and it goes black and stays that way.

Monitor actually goes into power save/sleep/no signal mode (Orange light).

makokikadur 12-14-2014 04:24 PM

In my case hitting ESC twice at the black screen brings the image back.

If X does not start you can use
Code:

startx
or try to bring up the display manager
Code:

sudo service lightdm start
Another possible workaround would be booting an earlier kernel (holding down LEFT SHIFT key at boot to access the bootloader menu and selecting a previous kernel.)

There's a wiki article about this problem that might be useful:
https://wiki.ubuntu.com/X/Troubleshooting/BlankScreen

widget 12-14-2014 05:24 PM

Have you tried booting to an older kernel?

Have you tried booting to recovery mode?

I found this referencing a driver for your chipset;
http://linux.die.net/man/4/openchrome

But there is no such package listed in the Sid (Debian) repo so I suspect it is a very old page.

I believe your chipset should be handled with the firmware that comes with the kernel.

It is very posible that it just isn't up to the job.

One way to test that would be to install another DE on there but it is a long shot.

I don't like the reference in your dmesg output;
"via_rhine: Broken BIOS detected, avoid_D3 enabled"

I have no real idea what that is about. Doesn't look good.

gkasica 12-14-2014 06:29 PM

Widget hasn't worked since I ran the upgrade process and live CD will work for the same distribution version. Recovery mode will get me a text interface.

startx will run and go blank so I know it's starting. Ps shows that lightdm is running.

gkasica 12-14-2014 07:31 PM

Ok doing some testing here.
At reboot you get the Grub boot menu and picking the regular boot option you get the Ubuntu splash screen and a single ESC will give you a text screen to watch boot up to a point where it goes black and the monitor light goes orange for loss of input etc. Letting it finish boot (I have an email sent out at the end of rc.local) still has a blank screen but a ps shows lightdm and X running.
The last message I can catch on text is libvirtd sfarting up then blackness. After the server is fully up hitting ESC twice does nothing still black. CTL-ALT-F1 will get me a text login prompt. Looking then I see that X11 is still running as is lightdm.

Typing service and a tab shows there is an X11-common which doesn't appear to be running
Starting it gets an ok but then restarting lightdm gets me the back screen. The ctl-alt-F1 gets text session back startx gets a scroll of a whole bunch of stuff then a blank screen ctl-c stops that. The results of that are in the X0 log file I sent along.

Any ideas?
Where is the config file for X11 located. Searched without success or I'm looking for the wrong name.

widget 12-14-2014 08:34 PM

OK, so the same version works as a Live Session. This is great.

If you do a manual upgrade you find that some things don't work quite right. There are things you can do to correct this.

First we need to know some things. Were you, before the version upgrade, using;
A>backports?
B>ppa's?

Was 12.04 fully up to date before running the version upgrade?

Did you read the release notes and any other documentation relating to the version upgrade? This is usually pretty well documented on the Ubuntu site.

Did you use the ISO for 14.04 for the upgrade or was this run using Update Mangler in 12.04? The process is a bit different and using the ISO is more reliable - in my opinion.

As some background for you, if you do not already know this, the LTS to LTS upgrade is pretty well tested. One of things that happens fairly near the release of the new LTS is a "step" upgrade in the old LTS and it should cause you to then be running 12.04.4 (I think this is still the 4th "step". These steps correspond with the release of new ISOs so that the ISO is up to date with package upgrades to the day of the ISO release. There are also some packages added to smooth the LTS to LTS version upgrade.

This testing is, or used to be, done by the ISO testing group of folks. Quite a bit of fun.

What you have actually done is attempt to upgrade from Debian 7 (Wheezy) which was the Debian testing version on which 12.04 was based to Debian 8 (Jessie) which is now in freeze for release sometime in 15.

This change includes the change from using startup to using systemd unless I am badly mistaken. I am not at all sure because my computer died in Feb and really didn't have a box running Debian-testing again utill early Sept so I am not sure just when that change over took place.

Also Debian hasn't used startup but sysvinit. So the Canonical devs had quite a chore assuming that 14.04 is using systemd.

There are always a lot of changes between LTS releases. So there are a lot of things that can go wrong. If you are not very careful in reading the release and version upgrade notes things can get hairy in a hurry.

Hopefully what you have is some screwed configurations or broken packages or perhaps both.

Let's find out if this is the case, rather than the obvious video problem, before you do more to your system which could complicate things.

widget 12-14-2014 08:43 PM

Well the switch to systemd is not the problem. That is supposed to take place as in 15.04. This is good as it makes things less complicated.

colorpurple21859 12-14-2014 09:19 PM

The working xorg-log shows the mach64 driver being loaded, whereas the bad log shows only vesa being loaded, maybe creating an /etc/xorg.conf file with the following will fix the problem:

Code:

Section "Device"
      Identifier  "Card0"
      Driver      "mach64"
EndSection


gkasica 12-14-2014 09:20 PM

As far as I know I didn't have any back ports or ppa's. 12.04.1 LTS was totally up to date and without issues. I did an apt-get dist-upgrade immediately before doing do-release-update from console command line. I don't recall ever being offered a 12.04.4 LTS version. I looked at the release notes and don't recall seeing any issues that I had to be concerned with before the upgrade. The actual upgrade went flawlessly with this one exception. What's the best way to go about working through the broken package or messed over config file? Would it be good to apt-get purge all relevant items and reload? If so what items do I dump? Or is there a better answer. I have a 12.04.1 LTS server still working - the production box - but the video cards are different but go to the same Dell LCD display via a port switching device. Does any of this help?

gkasica 12-14-2014 09:22 PM

Color purple the other system has a different video card hence that difference.


All times are GMT -5. The time now is 09:07 AM.