CentOS - Dec 2016 - Kernel Panic at Boot

Hi All,

This is a Centos 6.4 system used as a home server and Myth backend 
machine. It has just suffered a motherboard failure and I have got a 
identical replacement.

This machine normally runs in run level 3. It has a NVidia graphics card 
and the NVidia driver installed but I found it crashed the machine as 
soon as I started X on there. So I never bothered with X and left it in RL3.

So as part of the re-commissioning of this machine I decided to remove 
the NVidia driver and go back to Nouveau. I shouldn't have bothered.

Even starting in RL1 it kernel panics at bootup toward the end of the 
boot process. If I boot it in level 3 all the daemons start and it runs 
a little bit of script I have in rc.local to enable ACL's and then panics.

The panic is something like

panic occurred, switching back to text console.
general protection fault: 0000 [#2] SMP
last sysfs file: /sys/devices/virtual/block/dm-5/dm/name
CPU 1
Modules linked in: "........list of modules......."

Pid: 147, comm: plymouthd Tainted: G         D -------------------  
2.6.32-358.6.2.el6.x86_64 #1 Gigabyte Technology Co., Ltd P43-ES3G/P43-ES3G

There's a list of CPU registers and a call trace after that. I have a 
photo of the screen if that information is useful

Any idea how I can trace this.

Many thanks

Ken


-- 

Ken Smith
K-Net Technology
020 8651 5722

On 12/7/2016 12:09 PM, Ken Smith wrote:
> Gigabyte Technology Co., Ltd P43-ES3G
thtas a rather old motherboard, like circa 2008?    my longtime 
experiences with consumer desktop grade hardware suggest that at 5 
years, 50% of them go flaky.   that one is about 8 years old now.



-- 
john r pierce, recycling bits in santa cruz

John R Pierce wrote:
> On 12/7/2016 12:09 PM, Ken Smith wrote:
>> Gigabyte Technology Co., Ltd P43-ES3G
>
> thtas a rather old motherboard, like circa 2008?    my longtime 
> experiences with consumer desktop grade hardware suggest that at 5 
> years, 50% of them go flaky.   that one is about 8 years old now.
>
>
I know its old. I have quite a few (5) PCI video capture cards in there 
for MythTV and very few modern motherboards have PCI interfaces these 
days. Moving to something more recent would also mean scrapping those 
cards.

That said, I doubt that the age of the motherboard is related to the 
kernel panic after I stupidly de-installed the NVidia driver. It would 
be a sad double fault that the previously unused (but old) replacement 
mother board suffered a failure at the same time as I de-installed the 
NVidia driver. That would be very bad luck.

I have subsequently found that the 2.6.32-358.el6.x86_64 kernel still 
boots fine on there so I very much doubt I'm on the trail of a hardware 
fault.

Would I be safe just to reinstall the 2.6.32-358.6.2.el6.x86_64 kernel?

Thanks

Ken