Opensuse 11.4 - Boot problem with frozen screen right at the first second - repeatable

stakanov · June 23, 2011, 5:02pm

I have a repeatable and annoying problem that seems to be connected with the graphics card driver of nvidea. I had another card before (7900 card) and I have after having had a malfunction now a nvidea 430. Mainboard Gygabyte with AMD 790 Chipset, 4 GB Ram. The opensuse system is fully patched and updated, newest kernel but the same error occurred with another graphic card and with the precedent kernel and with the precedent nvidia video driver.

Now, the problem is as follows:
System is dual boot with options:

openSUSE 11.4 64 bit
openSUSE failsafe
Windows XP Sp3

Windows boots normally without trouble.
Failsafe boots normally without trouble.
The option one presents a quite erratic and not really expected error pattern.

if you put in any whatsoever parameter at boot (nomodeset or edd=0 or run parallel no than the machine does boot.
if you do use nvidea driver OR the nouveau driver the machine screws up right after the beginning of the boot screen (no possibility to see the screen output because esc would not make before the lock up.
if you do use the same parameters that make the system start normally as parameters in the grub (through yast) then the same system locks up again. No matter what parameter you try.
occasionally you notice that the system has locks (temporary) when shutting down from the normal boot (that is when you did start with some parameter). This will happen generally with the ethernet at shutdown. Sometimes it blocks shortly when “setting firewall rules”. When moving a mouse or doing ESC it will go through to login normally.
The functionality once you are on the desktop is perfect. No errors or locks, no problems at all.

But the problem remains a repeatable sometimes random lock up right at the beginning of the boot process. Any idea how I could see what happens? I also do not recall the parameter to be put in grub to have verbose output.
Thank you in advance for any help to make this system work.

DenverD · June 23, 2011, 7:02pm

On 06/23/2011 05:06 PM, stakanov wrote:
>
> - if you put in any whatsoever parameter at boot (nomodeset …
> than the machine does boot.

then, it is pretty sure a graphics/nvidia problem, go here
http://tinyurl.com/23mgej6 where you will see that nomodeset is “The
first thing to try” and, since you know that works you don’t need to do
anything else on that long page, instead hop theory guide, here:
http://tinyurl.com/37v9y7m and work your way to a nvidia smile…

at least, i think that is what is going on…

–
DD
Caveat
Hardware
Software
23 June: Sunrise 4:36 AM, Sunset 10:03 PM

stakanov · June 24, 2011, 11:09am

Thank you DenverD. However. This seems not to be a video related issue. What happens is that the system now has the nouveau driver blacklisted and the no kms = yes. I am running the latest Nvidia drivers. Behavior:

by putting nomodeset the system freezes immediately
if I do not put nomodeset the system starts the first time O.K.
if I do restart normally or I do start Windows and then restart openSUSE the system blocks immediately.
if I then restart and change one time (without writing anything to boot options) to failsafe menupoint (without selecting it - that is I do NOT start the system, just … ) jumping back to option one and then I start: system starts normally.
That is: sometimes it starts up to “starting device drivers” and then hangs for a long time. I then press just enter and it goes through up to the desktop, working well without problems.
Shutting down in 50 percent of the cases (but not in all) it blocks when running /etc/init.d/halt.local telling me - “missing”. When pressing enter the shutdown proceeds normally.

Although this is to some extend random, the whole “show” is not. The pattern is stable. No real error messages. 11.1 did run fine, Windows runs fine. Even 11.4 once started with this tedious trial and error thing runs fine. So what could that be and how could I log? I have some doubt about an AHCI issue. Logging for this seems to be disabled. How can I enable logging for AHCI events, does anybody knows?
Thank you.

DenverD · June 24, 2011, 12:10pm

On 06/24/2011 11:36 AM, stakanov wrote:
>
> - by putting nomodeset the system freezes immediately
> - if I do not put nomodeset the system starts the first time O.K.

well, my understanding is that once you have installed the correct
nvidia (which you say you have now done) it should not need nomodeset,
and i’m not surprised you have troubles if you use it…

and, i’m not surprised if you don’t use and it boots fine…

> - if I do restart normally or I do start Windows and then restart
> openSUSE the system blocks immediately.

ok now, define “if i do restart normally” what a normal restart to you?
to me, i click on the desk top, pick leave, then on the popup i pick
“Turn off computer”…is your “normal restart” different? during your
“normal restart” press the Esc key and look and see if you have a bunch
of errors (maybe)…

and, you are also saying that always after running windows it “blocks
immediately”?

and, please define “blocks immediately”…what is that? is it the same
as your “freezes” in the first line above? what does both “block” and
“freeze” look like?? flashing lights or not? hard drive light full on
busy busy busy, or not? black screen? no screen? can’t read screen? can
read screen but nothing moving?

> - if I then restart and change one time (without writing anything to
> boot options) to failsafe menupoint (without selecting it - that is I
> do NOT start the system, just … ) jumping back to option one and
> then I start: system starts normally.

this is VERY strange…i think what you are saying is if you (not using
nomodeset) you boot up to a working system, shut down the next time you
boot up if you just sit on your hands it hangs…BUT, if instead of
sitting on your hands you push the down arrow and then the up arrow (to
get back to the normal boot) it does not hang…is that what you are
saying?

> That is: sometimes it starts up to “starting device drivers” and then
> hangs for a long time. I then press just enter and it goes through up
> to the desktop, working well without problems.
> - Shutting down in 50 percent of the cases (but not in all) it blocks
> when running /etc/init.d/halt.local telling me - “missing”. When
> pressing enter the shutdown proceeds normally.

strange, please show us the output of entering the following into a terminal


ls -hal /etc/init.d/hal*
echo ==================================================
sudo cat /boot/grub/menu.lst

and copy/paste the output back to this thread using the instructions
here: http://goo.gl/i3wnr

> I have some
> doubt about an AHCI issue. Logging for this seems to be disabled. How
> can I enable logging for AHCI events, does anybody knows?

i don’t know anything about AHCI…sorry

–
DD
Caveat
Hardware
Software
24 June: Sunrise 4:36 AM, Sunset 10:03 PM

stakanov · June 24, 2011, 12:49pm

this is VERY strange…i think what you are saying is if you (not using
nomodeset) you boot up to a working system, shut down the next time you
boot up if you just sit on your hands it hangs…BUT, if instead of
sitting on your hands you push the down arrow and then the up arrow (to
get back to the normal boot) it does not hang…is that what you are
saying?

Yeap, this is the precise behavior. Lock up immediately means: no line of bootlog (I do boot now in splash=verbose - otherwise you would see just the frozen screen, even with ESC).
And sometimes when it hangs it is also O.K. to move the mouse, in alternative press enter and it would run up to the desktop without error. And the fun thing: it is quite random but it hangs somehow more reliably:

with udev starting
during shutdown when trying to shutdown eth1
during shutdown when writing: sending all processes the term signal. It would then hang. But it does shutdown if you press the enter key one time (or you move the mouse).

stakanov · June 24, 2011, 3:16pm

Here we go:

linux-fces:~ # ls -hal /etc/init.d/hal*
-rwxr-xr-x 1 root root 3.5K Feb 19 06:23 /etc/init.d/haldaemon
-rwxr-xr-x 1 root root 6.1K Nov  2  2009 /etc/init.d/halt
-rwxr--r-- 1 root root  343 Jun 22 18:31 /etc/init.d/halt.local
linux-fces:~ # echo ==================================================
==================================================
linux-fces:~ # sudo cat /boot/grub/menu.lst
# Modified by YaST2. Last modification on Fr Jun 24 15:06:20 CEST 2011
# THIS FILE WILL BE PARTIALLY OVERWRITTEN by perl-Bootloader
# Configure custom boot parameters for updated kernels in /etc/sysconfig/bootloader

default 0
timeout 60
##YaST - generic_mbr
gfxmenu (hd0,5)/boot/message
##YaST - activate

###Don't change this comment - YaST2 identifier: Original name: linux###
title Desktop -- openSUSE 11.4 - 2.6.37.6-0.5
    root (hd0,5)
    kernel /boot/vmlinuz-2.6.37.6-0.5-desktop root=/dev/disk/by-id/ata-SAMSUNG_HD502HI_S1VZJ1KS407054-part6 resume=/dev/disk/by-id/ata-SAMSUNG_HD502HI_S1VZJ1KS407054-part7 splash=verbose quiet showopts vga=0x361
    initrd /boot/initrd-2.6.37.6-0.5-desktop

###Don't change this comment - YaST2 identifier: Original name: failsafe###
title Failsafe -- openSUSE 11.4 - 2.6.37.6-0.5
    root (hd0,5)
    kernel /boot/vmlinuz-2.6.37.6-0.5-desktop root=/dev/disk/by-id/ata-SAMSUNG_HD502HI_S1VZJ1KS407054-part6 showopts apm=off noresume edd=off powersaved=off nohz=off highres=off processor.max_cstate=1 nomodeset x11failsafe vga=0x31a
    initrd /boot/initrd-2.6.37.6-0.5-desktop

###Don't change this comment - YaST2 identifier: Original name: windows 1###
title windows 
    rootnoverify (hd0,0)
    chainloader +1

###Don't change this comment - YaST2 identifier: Original name: floppy###
title Diskette
    rootnoverify (fd0)
    chainloader +1

DenverD · June 24, 2011, 4:13pm

On 06/24/2011 03:36 PM, stakanov wrote:
>
> -rwxr–r-- 1 root root 343 Jun 22 18:31 /etc/init.d/halt.local

you can see that that file exists, even though it complained that it was
missing… i do NOT have a clue what is going on…sorry…hope you find
the problem…

now, i have not gone back through all of your post to learn if you got
to 11.4 by “upgrading”, or doing a full disk format and install, or
saving /home and formatting the rest…so, i have no idea what the
likelyhood is that the ‘upgrade’ you did left a bunch of junk in the
root directory or in your home…

neither do i know if you have installed programs from factory, unstable,
tumbleweed, playground or any of the “community repos” which may or may
not have perfectly stable code…

nor do i know if you have installed openSUSE over some other distro…or
used alien to install deb packages, or manually installed rpms made for
other distros…

neither do i know if or how many times you may have had forced shutdowns
or crashes or abrupt power offs, any of which may or may not have
corrupted some portions of your hard drive…

so i can’t imagine how your system got to be the way it is–and the way
it is is NOT like mine (which is full format install no code from
potential troublesome repos/packages–and smooth boots/shutdowns)…

not knowing how it got scrambled, i can’t begin to fix it…sorrry…

–
DD
Caveat
Hardware
Software
24 June: Sunrise 4:36 AM, Sunset 10:03 PM

stakanov · June 24, 2011, 4:54pm

The system is a dual boot. The openSUSE installation was done new, not through update on a partition (formatted that previously was of a 11.1. The home was maintained as is. The boot sector is the only one that might be the problem. One workaround to fix this would be: take the XP CD and then do a restore of the MBR following XP standards. Then do a restore of grub binding in the 11.4 partition. Now once upon a time that was easy with the DVD. But that routine has intelligently taken off, so I have no clue how to perform this.
The repos: all repos as usual; OSS, NOSS, Packman, and KDE ver.4.6 stable, Extra. No unstable versions of what so ever was installed. But, the problem presented itself right after the first new install, without any but the default repositories activated. The video driver is Nvidia so these repos are also active.
For me it is the boot sector that is screwed (MBR). But no clue how to fix it.

stakanov · June 25, 2011, 8:18pm

O.K. We will go by order.
The PC had incompatibility with the boot process without showing the typical behavior of the NVIDIA / nouveau driver pattern. HOWEVER it did work with the insertion of nomodeset or what ever else parameter as long as interaction with the keyboard did take place. As I did a change of the video card and a fresh install at the beginning I was able to rule out the problem of software. The fact that either nouveau and Nvidia proprietary driver did present problems and the system did start with fail safe, this made me think about the APIC / powersafe functions. I then did proceed to a BIOS update of the mainboard. This by itself did not help.
System:

Opensuse 11.4 with latest
KDE 4.6.0.0 stable and Kernel Linux 2.6.37.6-0.5-desktop x86_64

**Mainboard: **Gigabyte GA-MA790XT-UD4P (rev.1) with ATI 790 Chipset without graphics and Award BIOS. Bios Version is F8G.

The processor is an AMD Phenom™ II X4 905e Processor. The mainboard is not commercialized any more.
The Graphics card is a Nvidia GeForce GT 430 with 1 GB VRAM and
4 GB Kingston DDR 3 Ram.

So it was reasonable powerful. However boot (if it took place) was slow (up to several minutes) and the performance overall was sluggish.
A look in the BIOS showed that the CPU setting "CE1" was set to “active” instead I switched it to “auto”. Apparently the CE1 function is not completely supported with the kernel, and gives problems with boot up in Linux with the current kernel. You may wish to deactivate the option if you encounter sluggish system performance and problems during boot (hanging, random starts and hanging, hanging during start up requiring input via the "enter"key in order to get to the desktop.
I have now a quite powerful and working dual boot system. Using the proprietary Nvidia driver.
Thank you Denver for trying to give me a help on this. Hope this info makes it easier for somebody “googleling” for this issue.
Cheers.

P.S. Now I have only to make my HDMI sound to work (will bother Oldcpu for that in a new thread).