Strange failure, probably video card

[Kevin O'Gorman]

SOLVED: it was hardware.  New video card, no more stuff in syslog

On Sun, Dec 8, 2013 at 2:27 PM, Kevin O'Gorman <kogorman at gmail.com> wrote:
> On Sun, Dec 8, 2013 at 2:19 PM, Kevin O'Gorman <kogorman at gmail.com> wrote:
>> On Sun, Dec 8, 2013 at 1:59 PM, Nils Kassube <kassube at gmx.net> wrote:
>>> Kevin O'Gorman wrote:
>>>> Kernel is 2.8.0-31-generic SMP x86_64 (amd 4-core)
>>>
>>> I suppose you mean 3.8.0-31 ... that would be the lts-raring kernel of
>>> Ubuntu 12.04 (although the latest lts-raring kernel is 3.8.0-34). Or are
>>> you using Ubuntu 13.04?
>>>
>> Yes.  3.8.0-31.  I actually seem to have -19, -31, -32, -33, and -34 in /boot.
>> I don't know why it's running -31, since I rebooted after the last upgrade.
>> After I send this, I'll see if I can get grub to try -34.
>>
>>>> I don't think it's the driver because I have had this card for several
>>>> years, but the log messages started just a week or so ago.
>>>
>>> Agreed, that looks suspicious, but it still might be triggered by some
>>> software upgrade.
>>>
>>>> However, I'm not sure how to even tell what driver I'm using, and I
>>>> fool around with many things on this machine -- I don't remember what
>>>> i did about drivers.
>>>
>>> If you fool around ith many things, could it be that you were playing
>>> with virtualization when the problem started? The AMD-Vi error message
>>> seems to be related to virtualization / iommu (not that I know much
>>> about it, but that's what I found when searching for the error message).
>>>
>>
>> No.  I haven't fooled with virtualization in some years.  And not at all since I
>> installed 64-bit from scratch.
>>
>>>> It may help to know that among the installed packages are
>>>>   nvidia-304
>>>>   xserver-xorg-video-nouveau
>>>>   nvidia-settings-304
>>>
>>> Then you probably have the proprietary nvidia driver in use. You can
>>> check it with the command
>>>
>>> lsmod|grep nvidia
>>>
>>> if it shows "nvidia" it is the proprietary driver, if it returns
>>> nothing, it is the "nouveau" driver.
>>
>> It shows 'nvidia'
>>
>>>
>>> Actually I have the same video card and also the lts-raring kernel in
>>> this machine. I think I had some problem with the nvidia-304 package and
>>> therefore I installed the nvidia-304-updates package. I don't see your
>>> error messages in the logs, but I don't have two monitors and I have an
>>> Intel CPU.
>>
>> Changing again.  Settings-> Software&Updates -> Additional drivers
>> from 313-update to 310-update.
>>
>> doing update-grub
>
> Very odd.  I had to hand-edit the grub entry to get it to use 3.8.0-34.
> I'm running it now, but it's still logging many many of those errors.
>
> Software&Updates says I'm running 310-update.
>
> Since nothing has budged.  Since 3.8.0-31 was installed 'way before the
> trouble started (Sept 10 and the trouble started 2 days ago according to
> the logs), I'm about to go buy a new card.  If nothing else, it will eliminate
> another variable, and my current frustration level makes it worth the money.

Thanks for your help.  It gave me pointers to stuff to look at, reassured me
I'd checked what could be checked, and I went to Best Buy with a clear
conscience.

And it worked.  New card nVidia GeForce GTX 650ti, system comes up
clean, no junk in syslog.

Whew!

-- 
Kevin O'Gorman

programmer, n. an organism that transmutes caffeine into software.
Please consider the environment before printing this email.