On Thu, Feb 5, 2015 at 7:23 PM, jdow <[log in to unmask]> wrote:
> On 2015-02-05 05:08, Joseph Areeda wrote:
>>
>> On 01/29/2015 08:41 AM, Phil Wyett wrote:
>>>
>>> On Thu, 2015-01-29 at 08:30 -0800, Joseph Areeda wrote:
>>>>
>>>> Hi All,
>>>>
>>>> I've been getting random crashes of X on my development workstation with
>>>> messages like:
>>>>
>>>> on console: cpu #X stuck in [X:...]
>>>> .xsession-errors contain things like: "gnome-session: Fatal IO error 11
>>>> (Resource temporarily unavailable) on X server :0."
>>>> /var/log/messages has "NVRM: GPU at 0000:01:00.0 has fallen off the
>>>> bus."
>>>>
>>>> Most things I've been able to find suggest a driver vs. kernel problem
>>>> so I updated to the latest driver from NVidia (346.35) with no luck.
>>>>
>>>> Have others seen this?  Any hints?  Could it be the NVidia card failing?
>>>>
>>>> Thanks,
>>>> Joe
>>>
>>> Hi,
>>>
>>> I have seen this once before with a persons system. The issue in that
>>> case was a bad physical connection. Removing the card, cleaning the
>>> connections and reseating the card corrected the issue.
>>>
>>> Regards
>>>
>>> Phil
>>>
>> This took a while because I had to finish a project and couldn't afford
>> random
>> reboots, but I took Phil's advice and used the professional connector
>> cleaner
>> AKA pencil eraser on the connectors, reseated the card.  I have now gone 4
>> days
>> without an incident whereas I was forced to reboot 3 or 4 times per day.
>>
>> Thanks Phil!
>>
>> Joe
>
>
> Just a note about pink erasers - don't. They contain a lot of sulfur. That
> corrodes contacts rather rapidly. So the fix may be lamentably temporary and
> a re-fix may be impossible after the second or third time.

I've had good success with it for edge-connected circuit boards, but
that was a while ago. Contacts were bigger back then, and the gold
tended to be thicker, so there was less chance of scrubbing your way
right through it.