System restarting during boot

I noticed one of my servers was restarting itself erratically this morning. A quick look at the logs showed that it had been doing this all weekend - restarting after it had been up for an hour, 30 min, 10, etc. My first thought was the the PS was going bad, no big problem. After switching out the PS, the power cable, using different outlets, etc, I've concluded that the PS is probably not involved.
Then the system started rebooting DURING the boot sequence. I'd get into the GUI part of the boot, and the machine would power down and restart. I tried to do a system repair, using the CD and emergency repair disk. Then I started getting an "Unknown hard error" with reference to ntdll.dll. I replaced this file from another W2K machine with the same configuration to no avail. I've removed/replaced every component except the mb and chip, with no visible effect. Now the machine is rebooting during the text phase of boot, and I'm even more confused. The only change to the OS I've made recently is that I may have installed SP3 (the only real problems I've seen related to that so far, though, are drivers for add-on cards, and I don't have any). Any ideas on what's going on?

Check the UPS. Try plugging it straight into the wall to see if it will boot.

I've done that. I plugged it into multiple outlets, changed cables, changed power supplies. It seemed like the power system was the prime suspect at first. But I can't seem to isolate it to the power. Also, I can boot a floppy, and it doesn't seem to power down (although I haven't left it sitting for a long time that way).

checked the cpu temp?

Can you get into the event viewer? If so see whwt it has to say. It may say something about rebooting from a "bugcheck" if so look at updating drivers, maybe a device on your sys is running off drivers that don't like sp3.

I had a w2k server that was rebooting from a bugcheck at random times till I updated printer drivers :( I read of someone else's rebooting due to vid drivers with the bugcheck message in the event log. etc.

Also if you can get to the desktop, right click on My Computer and go to Properties, and hunt around for the setting that says "When a stop error occurs" to reboot.

edit: also if it seems to like a floppy, but doesn't like the hdd, then run scandisk with the thourough option :confused:

Good Luck~

Thanks everyone. I don't have a good way to check the chip temperature. That's a real possibility, considering the state of my server room (frequently above 80 F). But I don't see any difference when the case is off, or when the machine has been idle overnight. While I can't rule it out, I don't think it's overheating (more than usual, anyway).

I can't even get to the log in at this point. It was rebooting after login, then during the gui part of boot, now in the early stages of boot. I didn't see any error messages, though, when I could get to the event log, except for the unexpected shutdown.

I also tried putting a new HD in the system, and doing a clean install. I had the same problem. This suggests to me that the problem isn't with the OS at all. At this point, I'm guessing the problem is the motherboard. I'm going to replace it (good chance for an upgrade - I'll take a clue from your .sig, Radical), do a clean install, and reinstall the database that was living in the server. If that doesn't work, I'm going to go out for a beer. Or three.

It's not easy to check CPU temp when you can't boot and there's no BIOS provision for it.

But, here's something you can do...

1. Power down server and leave off for awhile.

or

2. Open case and blow air inside.

If the system stays up appreciably longer, you know your problem is related to heat.

(Could also be a short with the case open, but less likely, methinks.)

Quote:

---

Originally posted by The Computer Valet
It's not easy to check CPU temp when you can't boot and there's no BIOS provision for it.

But, here's something you can do...

1. Power down server and leave off for awhile.

or

2. Open case and blow air inside.

If the system stays up appreciably longer, you know your problem is related to heat.

(Could also be a short with the case open, but less likely, methinks.)

---

Or do it the RD way, put your hand on the hs, if it hurts, its HOT :D

Best to use something exceptionally delicate and sensitive to heat -- like privates. :D

Quote:

---

Originally posted by The Computer Valet
Best to use something exceptionally delicate and sensitive to heat -- like privates. :D

---

ummmm... OK whatever turns you on CV :eek2: :D :eek2:

I would stick with the old standby. Crack an egg on there. mmmmmmmmmm processor fried egg. :D

Okay, now you guys are telling me you keep your computers are clean enough to eat from. I want YOUR server room! I won't even bring perishables into mine for fear of contamination.

He didn't say you should EAT the egg! :D

Get AC in that server room! 80 degrees is too hot!!

Quote:

---

Originally posted by amyb
Get AC in that server room! 80 degrees is too hot!!

---

Werd. Nothing above 70 degrees! Keep that B*TCH around 68' and you'll be fine. 80 is just ridiculous.

Quote:

---

Originally posted by HIESLanMan
Okay, now you guys are telling me you keep your computers are clean enough to eat from. I want YOUR server room! I won't even bring perishables into mine for fear of contamination.

---

Well after tackling most of it with some dust-off, the rest is just seasoning. mmmmmmmm processor fried egg with seasoning.:D

Quote:

---

Get AC in that server room! 80 degrees is too hot!!

---

Tell me about it! It used to be worse - I finally tapped into the air vents that go past the room. At least I get a little air in the summer now. But if you think that the heat is my biggest concern, you should check out the flooding!

http://www.hies.org/thom/flood-pouring1.jpg
http://www.hies.org/thom/flood-pouring2.jpg
http://www.hies.org/thom/flood-storage2.jpg

So, how's your disaster recovery plan? :D