Could be a hardware problem. Check the errata/bios updates for your
board or system. I had a similar problem and it turned out to be a
hardware BUG with an Intel server board. It is funny the way they
describe the bug -- it is a manufacturer problem that affects only one
of 5 (or 6, I do not remember) PCI slots. If you have a card in that
slot, the system locks up under loads randomly. They solve it with a
BIOS upgrade by degrading PCI clocks on all of the slots!
The BIOS upgrade did solve my problem and the system uptime has been
above 4 months now. Before the upgrade, the system crashed once each
24-48 hours.
--
----------------------------------------
Constantine Gavrilov
Unix System Administrator and Programmer
Xpert Integrated Systems
1 Shenkar St, Herzliya 46725, Israel
Phone: (972-8)-952-2361
Fax: (972-9)-952-2366
----------------------------------------