[wellylug] hardware errors

[Richard Hector]

On 11/06/14 20:23, kingsley at loaner.com wrote:
> Hi Richard,
>
> I searched Google's indices for the error messages
> you wisely linked to.
>
> Evidently
>
>      they're called "Machine Check Exceptions",
>
>      they can be caused by a bad CPU, and
>
>      a package named "mcelog" can collect and
>      decode them.

Cheers, thanks.

I did do a bit of a search, and found some of that info. Unfortunately 
it seems to be possible that the error could be in the cpu, or the 
chipset, or the ram itself ...

Also, Debian Wheezy appears to have a bug that prevents mcelog working 
properly (I installed it a few days ago, but the log is empty, which is 
what the google results predicted). Since the errors only show up 
occasionally (the machine ran for 1.5 days with no errors), I probably 
can't just boot an alternative system for a bit and run mcelog to see 
what happens - there's a good chance nothing would, and the machine is 
still needed.

Luckily, it doesn't seem to be impacting what the machine is actually 
doing - with one possible exception: downloading packages with apt-get 
or aptitude will often fail with a size mismatch or something; I have to 
re-download, possibly multiple times, to get it right. Whether that 
issue is related to the hardware errors, I don't know.

Richard