21.10.2011 02:12, Dan Youngquist wrote:
Even if the CMOS battery is dead it won't make the
system unstable; it'll
just use default values for everything. But if the clock is still keeping
time, the CMOS battery is OK. The clock is the first symptom of a dying
battery.
I'd reseat the CPU, memory, and all cards, then swap in a different power
supply; sometimes a power supply will die a slow death and make things weird
for a while. And swap in different memory if you have any, or if no extra
memory, take half of it out and test with the other half, then swap.
I second all above. Exhausted CMOS battery would lead to the "CMOS checksum
error" message each time you completely poweroff your workstation (i.e. turn it
off and remove the 110/220/240V (whatever is used in your country) voltage from
being provided to the power supply unit. Even when CMOS settings are lost worst
problem you would hit is system trying to boot from incorrect boot device,
incorrect settings for the operation mode of your SATA controller (might switch
into "IDE/Compatibility" mode from "Native/AHCI" one) and some linux
services
hanging at startup due to not being prepared to find system time to be way back
into past.
My general experience in diagnosis of "magical problems" suggests that the
first
candidate to blame usually turns out to be faulty memory modules (you had tested
it and proven that's not the case - 6 memtest86+ passes are good enough to be
sure that the memory is OK with 99% probability), then comes the slowly dying
power supply unit and third one to blame are usually bad/oxidised contacts. If
you've got some spare and known to be functional PSU to test with - that would
be your best bet to start with. It wouldn't hurt in any case to eject and insert
back CPU, memory and all other ejectable expansion cards your system is equipped
with. Booting from live CD of one of the LTS distros and trying to work at that
environment for a while would also help to distinct if your problem is caused by
hardware or by software.
Good luck with your fight for stability, I know how bad is it to have faulty system.
--
Best regards,
Alexey Loukianov mailto:mooroon2@mail.ru
System Engineer, Mob.:+7(926)218-1320
*nix Specialist