GA-MA770-S3/SB600/SATA & Random Reboot

R

rjw

Hi all,

I'm currently experiencing an odd and annoying problem with my desktop randomly rebooting. So far it only appears to be happening with FreeBSD however right now I am load testing it under Fedora 12 Live CD to try replicate and further identify the issue.

Originally I was running FreeBSD 9.1 amd64 (custom kernel) with one 40 GB SATA drive for the OS and two 80 GB SATA drives striped using gstripe(). I've been using this setup for roughly six months and the problem has existed the entire time although very intermittently. Due to time constraints it's only now that I can start looking into the problem.

I'll outline below all the things I have done thus far.

  1. Removed the striped disks. Problem still occurred although less frequently.
  2. Set BIOS to Fail-Safe Defaults plus set SATA controller to run in the following modes:
    • Legacy IDE
    • Native IDE
    • RAID
    • SATA->AHCI
  3. Installed FreeBSD 8.4-RELEASE amd64 - Seven minutes after default install and while I was installing subversion ( pkg_add -r subversion) the system rebooted.
  4. Installed FreeBSD 8.4-RELEASE amd64 and compiled a debug kernel. The system rebooted with no crash dump.
  5. Installed Microsoft Windows 7 with some downloaded driver updates. The problem didn't occur however I wouldn't call this test substantial.
  6. Removed the SATA drives, installed two IDE drives and ran the following script:
    Code: 
    #!/bin/sh
while [ 1 -ne 2 ]; do
  for disk in ad0 ad1; do
    dd if=/dev/urandom of=/dev/$disk bs=1M
  done
done
    This ran for many hours and the problem did not occur. I then found a potential fix regarding setting hw.pci.enable_msi=0 and hw.pci.enable_msix=0 with the SB600 controller.
  7. So I reinstalled the SATA disks as per the above, booted using mfsBSD 8.3 i386, applied the sysctl options and ran the above script. A few hours later the system rebooted.
  8. I then booted mfsBSD 8.3 amd64 however I set the options prior to booting into mfsBSD and ran the above script. This time the script ran for roughly eighteen hours without rebooting... until this morning. I truly thought I had found the fix. :(

When the system reboots there is no panic and no crash/core dump which suggests it's a hardware fault. Based on the above and assuming Fedora doesn't reboot then it would appear to me that FreeBSD is somehow triggering a fault in the SB600 controller (which is known to be quirky).

I am at my wits' end with this and my only two options are:

  1. Use the two IDE disks which doesn't give me much capacity.
  2. Purchase a SATA RAID controller.

If anyone has any further suggestions or questions please let me know.
 
I retract my above claim. Fedora also reboots however only randomly during the boot-up phase, so it is likely a hardware issue.
 
An update to this issue. I found the thermal paste on my graphics card was rock hard and largely only surrounding the chip and not actually on it and this was causing the reboots when booting up the PC. After swapping that with a new card that issue was resolved. I figured that may have been the issue all along so I reinstalled FreeBSD. FreeBSD continued to reboot. It is random - it can be when I'm using it or when idle at 3 AM. At no point has FreeBSD's uptime been greater than one week.

I have been running Fedora 19 now for a couple of months now without an issue and its current uptime is thirty days. Although I think this is related to a hardware fault, FreeBSD is the only operating system triggering it. I want to use FreeBSD - as I think it's truly is better than Linux - but I don't want to buy new hardware. For the first time in ten years, my desktop is not FreeBSD. :(
 
You must log in or register to reply here.
Back
Top