I am posting this in case it helps someone else. My Fedora 7 server was randomly freezing, either with a straight kernel panic or a weird freeze that left existing shell windows open but which would also freeze if any network interaction was attempted.
The system:
- Motherboard: MSI K9MM-V (MS-7312) with VIA K8M800 & VT8237R chipsets
- Processor: AMD Athlon 64 X2 4400+ (2.3 GHz) Socket AM2
- Memory: Crucial CT2KIT6464AA80E 1GB kit (512MBx2), 240-pin DIMM, DDR2 PC2-6400 DDR2 PC2-6400 CL=5 Unbuffered NON-ECC DDR2-800 1.8V 64Meg x 64
I tested the memory for 18 hours using Memtest86+ and it all checked out ok, I also tried out using maxcpus=1 and acpi=off separately, but eventually found the following combination of BIOS configuration and kernel options which have appeared to solve the freezes.
BIOS: Disable everything you don’t need (e.g. Onboard FDC Controller and Parallel Port) and make sure that ACPI is off. I also found that I had to keep the VGA Share Memory Size at 32MB if I wanted to use the embedded video adapter.
Cell menu:-
Memory Voltage(V): 1.8
Cool’n’Quiet control: Auto
Auto Disable PCI Clock: Disabled
Spread Spectrum: Disable
Adjust CPU FSB Frequency: 200MHz
AGP & P2P Bridge Control:-
AGP Aperture Size: 32M
AGP3.0 Mode: 8X
AGP Driving Control: Auto
AGP Driving Value: DA
AGP Fast Write: Disabled
AGP Master 1 WS Write: Disabled
AGP Master 1 WS Read: Disabled
AGP 3.0 Calibration cycle: Disabled
DBI Output for AGP Trans.: Disabled
VGA Share Memory Size: 32MB
DRAM Configuration:-
Current FSB Frequency: 200MHz
Timing Mode: Auto
Memory Clock value or Limit: DDR 400
VIA OnChip PCI Device:-
USB Controller: All Enabled
USB2.0 Controller: Enabled
OnChip LAN: Enabled
Onboard Lan Boot ROM: Disabled
USB Emulation: OFF
AC97 Controller: Auto
Power Management Setup:-
ACPI function: Disabled
Suspend Time Out(Minute): Disable
Power Button Function: Power Off
Restore On AC Power Lost: On
Now add the options ‘acpi=off maxcpus=1’ to the kernel line in /etc/grub.conf
title Fedora (2.6.23.1-49.fc8)
root (hd0,0)
kernel /vmlinuz-2.6.23.1-49.fc8 ro root=/dev/md0 acpi=off maxcpus=1
initrd /initrd-2.6.23.1-49.fc8.img
These settings compromise performance and I don’t benefit from the second core and AMD’s Cool’n’Quiet technology, but at least the server stays up now!