Strange Problems with Server Machine

[SiMac]

I built a machine which will hopefully be a server with the following parts:

Artec 56x CD-ROM Drive (Master on ATA/100 Bus 2)
Maxtor 40.9GB UDMA/133 HD (Master on ATA/100 Bus 1)
AMD Athlon 1.4GHz CPU
Biostar M7VKS Motherboard with Award BIOS VKS1003B (Yeah, yeah, I had to get a cheap motherboard because I needed to keep it <$500)
SMC 1244TX Ethernet Card
512MB PC133 Crucial Memory
1 case exhaust fan
1 CPU fan
1 PCI slot fan
Generic case
Red Hat 7.2

Now, I'm having very strange problems with the server. Linux installs from CD-ROM (booted from CD-ROM) fine (tested multiple
times). No errors are reported. However, when booting off of hard drive
after installation, “Unable to handle kernel NULL pointer dereference at
virtual address xxxxxxxx” or “Unable to handle kernel paging request at
virtual address xxxxxxxx” (xxxxxxxx is a number) error reported.
Occasionally other errors are experienced, including inability to find
/etc/fstab. Different errors occur during different bootups; rebooting the
system doesn't report the same error most of the time. System appears to be
fine when in rescue mode after booting from Red Hat CD-ROM, and fsck on any
partitions does not report problems. RAM tests fine in another machine with
a RAM integrity testing application. CPU Temperatures usually hover from
60-70°C according to under chip sensor on motherboard. All power supply
voltages reported by CMOS Setup Utility appear to be fine, except the 12V wire is
13.2V (not sure if that is actually all that far off ordinary). The hard drive functions perfectly in another machine and 1.5GB of files checksum the same as the 1.5GB of files the drive they were copied from. Disabling swap does not stop the errors.

Any ideas? Thanks in advance.

[Toaster]

Howdy,
While I do a great deal in Linux, my experience with VIA and AMD under Linux is limited. Keep this in mind.
When you installed Linux, did you use the default installation?
Reason I ask is large drives under Linux (this applies to IDE as I don't see this error on SCSI devices) can be problematic unless a "boot" partition is used.
This boot partition needs to be 16MB or smaller. This partition hands off the boot process to the other volumes.
How does the system work if started from a floppy "boot" diskette? (not the repair disk but the boot disk that Linux offered to build at the end of the install porcess)
What "class" of installation did you choose? (workstation, server, custom, etc)

[SiMac]

I didn't make a boot floppy because I don't have a floppy drive attached to the system. I just tried reinstalling with a 16MB boot partition; it actually booted once fine (not sure why) but came up with errors upon the next reboot. I installed using the default before and experienced the same errors.

Upon looking around this forum, I've noticed that the CPU does seem to be a tad on the warm side. Is it possible that the CPU is causing these problems? If so, can anyone recommend a powerful, inexpensive cooling system?

Thanks.

[Toaster]

Hello again,
While its possible that you have a stability problem of some sort, I wonder if RH fully understands your hardware?
As a first guess, go to the RH site and reserch your hardware.
I don't know if you realize this, but here on the 'Mech we have a forum for non-Micky$oft O/Ss. I'm going to copy this to that forum as well because folks there might be better able to offer assistance.
I started focusing on IRIX for SGI and my AMD knowings are quite limited.

[SiMac]

Given the varying behaviors, it seems most likely that it is a stability problem. The errors occur at different points in the startup process and are different errors, and once I didn't experience any at all. However, perhaps I should try to acquire a Windows boot disk to verify this.

[Statica]

Try disabling your CPU cache from your BIOS and try to see if it works.
I really would suspect the RAM from the nature of your error.

Here's something to keep in mind, just cos Windows will run on it, doesnt mean that either the RAM or the cache are good, thats the unfortunate truth.

[SiMac]

My first suspicion was the RAM also, however, I verified it with a utility in my Mac and it seemed fine. I've tried disabling the cache; it makes the CPU noticeably slower but does not fix the problem. Any other ideas?

Simon

[Statica]

Are you familiar with running something like ksymoops etc? It would help immensely if I had the dump to look at.

Ok to proceed .. how about first getting a look at your /proc/interrupts.

It seems like a bad stock driver with one of your devices, If you are pretty sure about the RAM, then try doing the following...

Disable your Trident integrated video and integrated audio, put in an old video card in and try doing a setup.

RedHat is really slow at getting new kernels into its systems. So it worked great for 2.2.xx but the 2.2.xx kernels were simply stated .. solid performers .. unlike the 2.4 which is so buggy. (that has led me away from Linux because I believe that Linus is the worst thing for the OpenSource movement with his choking stranglehold on Linux.) Also another thing to try is perhaps put in 2.4.17, or the ac patches. I'd really recommend the ac patches, he is much faster at integrating patches and workarounds. in fact the 2.4.7 ac patches have numerous VIA provided fixes.

[SiMac]

I managed to install a kernel rpm from rpmfind of 2.4.16 (with a ton of patience; I had to get to a command line of the system while booted from the HD because I have no floppy drive and no boot CD. I wouldn't want to build a kernel on this thing, so I used the RPM), however, I have the same problems.

I'm not familiar with ksymoops, as I've never had problems like this before. Please enlighten me.

It's probably not the video; this is before I can even start X.

Thanks for the help
Simon

[SiMac]

Okay, got the machine to get all the way to X-Windows and even ran Nautilus and Mozilla. I also ran memtest, the RAM is fine. I've also experienced a few random reboots...is a power supply problem a possiblity?

Thanks