Bug 539541

Summary: Installation Freeze Milestone7 i586 DVD
Product: [openSUSE] openSUSE 11.2 Reporter: andrea florio <andrea>
Component: KernelAssignee: Thomas Renninger <trenn>
Status: RESOLVED UPSTREAM QA Contact: E-mail List <qa-bugs>
Severity: Critical    
Priority: P2 - High CC: coolo, jeffm
Version: FactoryFlags: coolo: SHIP_STOPPER-
Target Milestone: ---   
Hardware: x86-64   
OS: Other   
Whiteboard:
Found By: --- Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---
Attachments: Hardware information
y2logs
new logs after a new freeze and reboot
Xorg.0.log
xorg logs
/var/log/messages

Description andrea florio 2009-09-16 09:31:40 UTC
Created attachment 318471 [details]
Hardware information

As Summary the installation always freeze on my PC.
It freeze after about 5 minutes and usually freeze on installation summary.

Before report the bug i got 3 freeze

2 at "installation summary"
1 at "customizing boot options"
Comment 1 Stephan Kulow 2009-09-18 18:15:20 UTC
Are you still able to switch consoles, e.g. alt-ctrl-f2? If so, provide yast logs
Comment 2 Stephan Kulow 2009-09-18 18:15:44 UTC
I assume media check was successful?
Comment 3 andrea florio 2009-09-20 06:38:58 UTC
yes, both md5 and sha1 sum was correct.. i also tried to re-burn the DVD at slower speed and i tried again.

1st time : the machine automatically reboot (about after "new installation/upgrade" screen)
 
2nd time: installation freeze, keyboard and mouse do not works (i want to check ncurses installation, maybe that's X.org freeze)

3rd time: installation successful

using the same media all the 3 times...
Comment 4 andrea florio 2009-09-20 06:55:14 UTC
sorry... one mistake

3rd time: installation freeze in "automatic configuration" after reboot

4th time: installation successful

i think that because the 3rd time froze when it was installed on the pc, y2logs should include freeza data isn't it?
Comment 5 Stephan Kulow 2009-09-20 18:42:02 UTC
#1 is still unanswered/unprovided
Comment 6 andrea florio 2009-09-20 21:16:48 UTC
Created attachment 318995 [details]
y2logs
Comment 7 andrea florio 2009-09-20 21:19:41 UTC
it looks to be a X.org failure, even after a succesfull installation (when i freeze i cannot switch to shell) it freeze several times. When it freeze it also reboot sometimes.

i know you dropped (finally from developer side, why so early?? from end-user side) xorg.conf and i'm afraid the problem is there.

in logs you should fine 2 or 3 times so rpms fail installation because after freeze/reboot rpm db was broken, and i had to manually rebuild it.
Comment 8 Stefan Dirsch 2009-09-21 09:21:37 UTC
For installation we still use fbdev driver - for each and everyone (unless you disabled kernel framebuffer, then you get vesa driver). If the issue also occurs during installation, we would have a general issue in X, since after installation the native driver should be running. More likely is that we see a kernel issue here. In any case please provide /var/log/Xorg.0.log, preferrably without any xorg.conf in place.
Comment 9 andrea florio 2009-09-21 09:43:56 UTC
the log hass been already provided: https://bugzilla.novell.com/attachment.cgi?id=318995
Comment 10 andrea florio 2009-09-21 09:59:49 UTC
Created attachment 319052 [details]
new logs after a new freeze and reboot

added new logs after a new freeze (and automatic reboot)

no xorg.conf used
Comment 11 Stefan Dirsch 2009-09-21 10:04:30 UTC
Created attachment 319054 [details]
Xorg.0.log

nv driver running (GeForce 8600 GT). Nothing obvious in the logfile.
Comment 12 Stefan Dirsch 2009-09-21 10:08:09 UTC
Ok. Please switch to fbdev driver by copying /etc/X11/xorg.conf.install to /etc/X11/xorg.conf (restart Xserver afterwards) to verify if the issue really also occurs with fbdev driver. If it doesn't, nv driver is likely the issue, otherwise we either have a general Xserver issue or it's likely kernel related.
Provide feedback.
Comment 13 Stefan Dirsch 2009-09-21 10:11:04 UTC
(In reply to comment #10)
> Created an attachment (id=319052) [details]
> new logs after a new freeze and reboot
> 
> added new logs after a new freeze (and automatic reboot)
> 
> no xorg.conf used

This time you used the proprietary NVIDIA driver. We cannot investigate that issue. Make sure that the freeze occurs with the nv driver. For this uninstall
the proprietary NVIDIA driver.
Comment 14 andrea florio 2009-09-21 10:27:35 UTC
the issue actually happend on the following stages:

booting in DVD installation time (fbdev driver ??)

booting from HD (yast2 firstboot automatic config) (nv driver)

running time (nv driver)

running time (nvidia proprietary driver)

i just "cp /etc/X11/xorg.conf.install /etc/X11/xorg.conf"  waiting a new freeze
Comment 15 andrea florio 2009-09-21 12:32:38 UTC
Created attachment 319080 [details]
xorg logs

Here new logs using xorg.conf.install as "xorg.conf" it froze anyway (i was running yast2 sw_single
Comment 16 Stefan Dirsch 2009-09-21 12:50:12 UTC
Sounds more like a kernel issue then. Could you attach the lines of /var/log/messages at the time, when such a freeze occurs.
Comment 17 andrea florio 2009-09-21 13:19:15 UTC
Created attachment 319085 [details]
/var/log/messages

attached messages log (i added few comments to allow you understand where it freeze
Comment 18 andrea florio 2009-09-21 13:19:36 UTC
i forgot to remove "NEEDINFO"
Comment 19 Stefan Dirsch 2009-09-21 13:24:17 UTC
So nothing obvious in kernel log either. Sorry, no idea what's happening her. I suggest to boot into runlevel 3. If it's occuring also in runlevel 3, it can't be an X related issue.
Comment 20 Stefan Dirsch 2009-09-21 19:26:46 UTC
Well, of course without X you can't do much and maybe the issue can only be reproduced by running a special X applications although the freeze might not be related to X at all.

I'm afraid we simply can't investigate intermittent freezes, if they aren't obviously related to X or kernel. Thus the resolution for now can only be
WONTFIX.

BTW, I'm also suffering by such freezes since some weeks now on my laptop. Probably X or kernel related, but also without any traces in X or kernel logs.
So you are not alone here.
Comment 21 andrea florio 2009-10-02 10:50:37 UTC
Stefan, i still have those freeze with M8, and it looks not to be related to X, since also "init 3" freeze too
Comment 22 andrea florio 2009-10-04 12:07:46 UTC
it's definetly a kernel issue, also ubuntu 9.10 with kernel 2.6.31 freeze after few minutes
Comment 23 andrea florio 2009-10-04 12:50:28 UTC
i reported on ubuntu launchpad too:

https://bugs.launchpad.net/ubuntu/+bug/442194
Comment 24 andrea florio 2009-10-04 22:02:50 UTC
i installed 11.1

added lock to kernel-default*
zypper dup to 11.2 

booting on 11.2 with kernel-desktop-2.6.31 ---> freeze

booting on 11.2 with kernel-default-2.6.27 (opensuse 11.1 one) ---> perfectly working
Comment 25 Stephan Kulow 2009-10-05 09:17:08 UTC
we need to find out how many machines it may affect. If this is specific to andea's machine, then it's not a ship stopper, but #20 says differently.
Comment 26 Stephan Kulow 2009-10-15 08:16:43 UTC
ok, there were no other reports against M8, so I don't think it's a ship stopper
Comment 27 Jeff Mahoney 2009-11-19 22:01:33 UTC
Do you have another system and/or does your system support a serial console? Setting up a remote console to capture the freeze would be extremely helpful.
Comment 28 andrea florio 2009-11-20 00:03:57 UTC
i think so.. how can check and how can i setup this serial remote console?
Comment 29 Jeff Mahoney 2009-11-20 00:17:36 UTC
If you're using the serial console, you'll need a null modem cable between the two machines. Then you boot with console=ttyS0,115200 on the node you're testing and run "screen -L /dev/ttyS0 115200" on the other. Change ttyS0 to whatever is appropriate for each node.

Putting it at the end will make it the "primary" console so kernel messages should automatically be directed there. When your system crashes, the messages will be preserved in the screen session on the other machine.

The other option is netconsole, but the modules for that aren't on the install initrd.
Comment 30 andrea florio 2009-11-21 07:28:37 UTC
ok guys... good news.
after spending weeks searching the web, i discover the problem is quite strange, but easy to fix.

looks like new kernel changed a bit the way the call acpi(or apic i always forget it). Any way, most of the asus mother board, now hang because of a bios bug that has been fixed by asus it sealf. A bios upgraded is needed to fix this issue.

in particular my mother board:

asus M2N-MX SE PLUS needs bios >= 0503 (latest is 06something)

i think that should be write into RELEASE NOTES.

looks also that disable acpi on boot should be a workaround but i never tought to use it since i never needed it.
Comment 31 andrea florio 2009-11-21 07:33:04 UTC
as you can see i wans't the only one to have such freeze:

https://bugzilla.novell.com/show_bug.cgi?id=411797#c34

and again looks to be an asus BIOS problem. one more reason to add this warning into the RELEASE NOTES
Comment 33 Thomas Renninger 2010-02-16 13:13:31 UTC
> looks like new kernel changed a bit the way
New kernel means this is fixed for you by a 11.2 latest kernel update?
Or did you try the very latest 2.6.33-rcX kernel?

If latest 11.2 update kernel does not fix it, it would be great if you double check kernel-default.rpm and kernel-default-base.rpm from here:
ftp://ftp.suse.com/pub/projects/kernel/kotd/SLE11-SP1

Best use rpm -ivh ... --force --nodeps
boot it once and if you like to you can easily uninstall it again by:
rpm -e kernel-default-${full_version}.rpm kernel-default-base-${full_version}.rpm

Already setting resolved.