Date: Thu, 7 Aug 1997 14:21:54 -0700 From: Frank MacLachlan <fpm@n2.net> To: Donal <donal@brewich.com> Cc: hackers@FreeBSD.ORG Subject: Re: vm_fault problem (fwd) Message-ID: <19970807142154.26219@n2.net> In-Reply-To: <Pine.BSF.3.96.970807123229.252p-100000@localhost>; from Doug White on Thu, Aug 07, 1997 at 12:32:48PM -0700 References: <Pine.BSF.3.96.970807123229.252p-100000@localhost>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Aug 07, 1997 at 12:32:48PM -0700, Doug White wrote:
> [snip]
>
> I am having a problem, and I hope you can help or direct me to someone who
> can.
> I recently upgraded my comm server to 2.2.2 and upgraded some of the
> hardware. The system is now:
> Gigabyte HX motherboard (new)
> Award BIOS
> on board IDE
> 2 com ports
> 2x16MB 60ns EDO RAM (new)
> 1.2GB Maxtor IDE (new)
> NE2000 card
> Trident ISA video card
> Usenet II 4 port Serial card
>
> Since the upgrade the machine has been rebooting at random intervals
> (anything from 5 minutes to 2 days) and I can find no log trace or core files.
> Then yesterday I caught it doing it and managed to write it down! I THINK
> this is exactly what it said...
>
> vm_fault: fault on nofault entry: address f296c000
>
> then it had something about syncing the disks (it failed) and that I could
> press a key to stop the reboot. I did press a key but it rebooted anyway.
>
> PLEASE help!!! Is it a bad SIMM??? MB Cache? Should I disable the cache?
>
> This machine is my gateway. My ISDN is connected to it, as are all 5 of my
> dialup lines. This reboot thing is crippling my business! Thank you in
> advance for your help.
>
I saw similar problems on a system fitted w/ a Tyan Tomcat III motherboard
(Intel HX chipset) and an AMD K5-PR166.
I'd suggest that you systematically reconfigure the system making one change
at a time and then testing the system. Iterative make worlds makes a great
stress test. If it still fails, make another change.
Things to check:
1. L2 cache
2. BIOS chipset settings. Use conservative settings for memory timing,
turn off advanced features, etc (I would, however, enable ECC if
you're using parity memory).
3. Replace your simms w/ known good simms.
4. Make sure that your CPU fan is working. Are you using heat sink
compound between the CPU and the heatsink? Does the heat sink
have ample surface area? Does the edge of the CPU feel hot if
you touch it?
5. Slow the bus speed down to 60 mhz for testing.
6. Make sure that the IDE signal cable is no longer than 18 inches.
7. Pull nonessential cards out of the system.
8. Does the case have adequate air flow? Some cases have very poor
air flow.
9. Double check motherboard settings. Is the CPU voltage set correctly?
10. Use your imagination.
My Tyan system would fail within 2 or 3 make worlds w/ the L2 cache enabled,
but successfully completed 22 passes w/ the L2 cache disabled. I replaced
the motherboard w/ an Asus P55T2P4 motherboard while the Tyan MB is being
repaired/replaced.
--
Frank MacLachlan (fpm@n2.net)
N2 Networking, San Diego, CA
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19970807142154.26219>