[CentOS] Fixing to bite the dust?

[nate]

sam wrote:
> I dunno Nate,
>
>     It started all this extraenous logging stuff after kicking memory
> from a paltry 256m up to 2 Gigs. The system has on occasion crashed with
> little else in the log files that would indicate any kind of other
> hardware problem, so with all the rejected packets or partial entries
> (none which showed up on this particular snippett leads me to believe
> it's dropping sync somehow.  I'll watch the further logs and stuff to
> see if I can find something more definitive.

System crashes often do not log anything.. if possible connect
a serial console to the system and configure it to put the console
on the serial port, then connect that to another system or terminal
server..

Or you can run something like memtest86, though it can take a while
sometimes to trace memory errors(days). There is a burn-in test
suite that a lot of vendors use created by VA Linux a long time
ago called Cerberus (ctcs), you can find it on sourceforge, in
my experience if the hardware is bad ctcs will crash the system
in a matter of hours every time(assuming motherboard/cpu/ram),
and it helps catch bad disk controllers as well as disks too,
putting the system under incredible strain.

nate