v11.4 dies at bootup

English Install/Boot/Login
#1

opensuse v11.4
linux 2.6.37.6-0.5-desktop x86_64

This was not a problem initially after upgrading from v11.3. It started about a week afterward. It is now quite consistent.

At a point where the boot sequence initializes the sound adapter, the screen goes black because the video is turned off, i.e., the output from the video adapter is ended. I say the “sound adapter” because I hear a faint thump from the speakers when the sound is initialized.

At this point the system is dead.

I can boot in Failsafe mode, usually. Which maybe implies a timing problem.

There is nothing in </var/log/messages> between the normal shutdown and the successful boot in Failsafe mode. The two intervening system crashes are not recorded.

Any ideas how to diagnose and treat the problem?

#2

How about telling us more about your hardware ?

Boot to failsafe and provide the output of:


/sbin/lspci -nnk | grep VGA -A2

and also immediately after a failed boot attempt, boot to failsafe mode and open the file /var/log/Xorg.0.log.old (which is the X log from the failed boot) and copy and paste it to SUSE Paste. And then post here the website / URL address.

Also while still in that failsafe session (after a failed boot) open with a text editor the file /var/log/boot.0msg (which is the dmesg log file from the failed boot) and copy its contents to SUSE Paste and then post here its website / URL address.

#3 
/sbin/lspci -nnk | grep VGA -A2
01:00.0 VGA compatible controller [0300]: ATI Technologies Inc Radeon HD 4770 [RV740] [1002:94b3]
	Subsystem: XFX Pine Group Inc. Device [1682:2900]
	Kernel driver in use: radeon

The system is still in failsafe mode since the last failed boot. These are the files you requested:

X.org.0.old <SUSE Paste>

boot.omesg <SUSE Paste>

#4

That makes no sense. That says you are using the ‘radeon driver’ from the ‘lspci’ command in FailSafe mode, and the Xorg.0.log.old says you attempted and failed to use the FBDEV driver during a regular boot. I should be seeing the opposite.

ie in FailSafe mode you should be using the FBDEV driver (and hence we should NOT see ‘kernel driver in use radeon’ in the lspci), and in the Xorg.0.log file we should see either radeon or fglrx failing.

It almost looks like you grabbed the xorg.0.log by mistake instead of the xorg.0.log.old (or the error is very unusual)

I can confirm that by the /var/log/Xorg.0.log file which has


[358298.875] Kernel command line: root=/dev/disk/by-id/scsi-SSEAGATE_ST336753LW_3HX0G0GM000073370Y4S-part2 apm=off noresume edd=off powersaved=off nohz=off highres=off processor.max_cstate=1 **x11failsafe** vga=0x375

and I can confirm that by the /var/log/boot.omsg file which has:


<6>    0.000000] Command line: root=/dev/disk/by-id/scsi-SSEAGATE_ST336753LW_3HX0G0GM000073370Y4S-part2 apm=off noresume edd=off powersaved=off nohz=off highres=off processor.max_cstate=1 **x11failsafe** vga=0x375

ie you gave me the log files from a Failsafe boot.

I don’t want that.

I want you to reboot to a regular boot. Then have it fail. Then boot again to Failsafe and obtain the /var/log/Xorg.0.log and the /var/log/boot.0msg. Those will be the log files from the failed boot.

Please try again.

Thankyou.

#5

Okay. The regular boot failed, the failsafe succeeded which what is running now. (I think the previous time there was a failure to boot in failsafe. Very distressing.)

Xorg.0.log.old <http://susepaste.org/77030595>
boot.omsg <http://susepaste.org/84838376>

#6

Unforuntately that is not the log file I need. I note this:


 33960.351] Kernel command line: root=/dev/disk/by-id/scsi-SSEAGATE_ST336753LW_3HX0G0GM000073370Y4S-part2 apm=off noresume edd=off powersaved=off nohz=off highres=off processor.max_cstate=1 **x11failsafe** vga=0x375

That suggests you did two failsafe boots back to back.

Unfortunately that is also not the log file I need. I note this:


<6>    0.000000] Command line: root=/dev/disk/by-id/scsi-SSEAGATE_ST336753LW_3HX0G0GM000073370Y4S-part2 apm=off noresume edd=off powersaved=off nohz=off highres=off processor.max_cstate=1 **x11failsafe** vga=0x375

ie in both cases, those files are showing the content of a ‘failsafe’ boot and not a regular boot.

I need you to attempt a regular boot. That will fail. Then attempt a ‘failsafe’ boot. Then provide those log files (Xorg.0.log.old and boot.0msg).

Those log files represent the previous boot attempt. If the failsafe boot fails and you need to boot failsafe twice, that does not help as it will only show me the failed failsafe boot, which is not what I want to review.

#7

I need you to attempt a regular boot. That will fail. Then attempt a ‘failsafe’ boot. Then provide those log files (Xorg.0.log.old and boot.0msg).

That is exactly what I did. That the regular boot files say it’s a failsafe boot is untrue.

Here is a directory listing for the files.

-rw-r--r-- 1 root root  31291 Jul 29 23:18 Xorg.0.log
-rw-r--r-- 1 root root  97083 Jul 29 23:17 boot.msg

-rw-r--r-- 1 root root  31305 Jul 29 23:13 Xorg.0.log.old
-rw-r--r-- 1 root root 105987 Jul 29 23:14 boot.omsg
#8

The times you quote make sense for a location in mid-west USA …

But the content of those two suse pastes look like what I would expect in these files:


-rw-r--r-- 1 root root  31291 Jul 29 23:18 Xorg.0.log
-rw-r--r-- 1 root root  97083 Jul 29 23:17 boot.msg

and not in Xorg.0.log.old nor boot.omsg.

Its unlikely the content of the lines I quoted would be wrong. They reflect the boot code. Without the correct files I am reduced to wild speculation, … which is unpleasant.

So along the wildspeculation front, can you provide the output of this command:


rpm -qa '*fglrx*'

there may be no output

Can you also try booting normally with the boot code ‘nomodeset’ . That should (assuming no xorg.conf nor 50-device.conf edits) boot your openSUSE with the radeonhd driver.

Also, do you have an /etc/X11/xorg.conf file ? Do you have any custom edits in /etc/X11/xorg.conf.d/50-device.conf ?

Apologies for the questions, but without the correct file content being provided/saved by openSUSE I am reduced to speculative question asking …

#9

On 07/29/2011 11:36 PM, jimoe666 wrote:
>
> At this point the system is dead.

is this system booting in a VM, or is it in real machine?

if in a VM, which?
and what is the host operating system/version in which the VM is running?

–
DD
Caveat-Hardware-Software
openSUSE®, the “German Engineered Automobiles” of operating systems!

#10

A real, physical machine.

#11

As oldcpu says I would definitely try adding the nomodeset flag when you do a normal startup, I have a Dell Studio 1558 and needed to add this after installing the ATI drivers from their repository otherwise my machine had exactly the same symptoms as yours.

#12

There was no output.

rpm -qa *radeon*
xorg-x11-driver-video-radeonhd-1.3.0_20100512_80ba041-2.1.x86_64

Can you also try booting normally with the boot code ‘nomodeset’.

That worked! Cool. And it seemed to boot a lot faster besides.

Also, do you have an /etc/X11/xorg.conf file ? Do you have any custom edits in /etc/X11/xorg.conf.d/50-device.conf ?

No. No.

#13

Thats good to read.

Likely you are now booting to the ‘radeonhd’ graphic driver, which in fact is not as fast as the ‘radeon’ driver.

It is possible the ‘radeon’ driver will still work if you using both the ‘radeon’ driver and the ‘nomodeset’ boot code at the same time.

If you wish to test that hypothesis, before testing that I recommend you install the application ‘midnight commander’ (mc) which is a text editor that is easy to use in a full screen text mode. Once it is installed one just types ‘mc’ to launch it and the rest is pretty intuitive. mc may or may not already be installed. Keep that text edit capability in mind in case you ever get stuck in a full screen text mode.

Then with any text editor, edit the file /etc/X11/xorg.conf.d/50-device.conf uncommenting the line with ’ Driver “radeon” ’ in it so that the file looks like:


Section "Device"
  Identifier "Default Device"
#
  Driver "radeon"

#
#  ## Required magic for radeon/radeonhd drivers; output name
#  ## (here: "DVI-0") can be figured out via 'xrandr -q'
#  #Option "monitor-DVI-0" "Default Monitor"
#

EndSection

and save the change and reboot using the boot code ‘nomodeset’ . If that fails, then reboot to failsafe and remove the edit. Do NOT keep backup files in the directory /etc/X11/xorg.conf.d/

#14

I did try that. And it worked without nomodeset.

However, the video adapter’s fan had become so loud that I removed the adapter and reverted to the MB adapter. Now there is a 50/50 chance that a boot succeeds without the nomodeset option.

I can see no difference in video performance between the ATI hd4770 (add-on) and the onboard hd3200. As a bonus, the CPU is running cooler and the noise level is quite pleasantly quiet.