NT Server reboots

Most likely it is set to re-boot on crashes in the startup/shutdown tab of the system properties. You could stop that so you can see the blue screen for any clues. It could be a power problem. I have had them. Make sure the server is on a good UPS. Also update the video and NIC drivers. Maybe if all else fails try to update to SP6a.

Dave

One possibility is that you are actually experiencing a BSOD ("Blue Dcreen of Death" aka STOP error) but your server is set to reboot automatically when this happens. First, do two things: 1) update all your drivers, especially NIC and/or video drivers; and 2) set your startup/shutdown parameters to write an event to the event log when a STOP error occurs (right click My Computer, select Properties, go to Startup/Shutdown tab). I'm not sure when the "write an event to the log" was enabled - you might have to run SP6a to get this option. This will cause the system to record the basic information about your STOP error in the Event Log - it will be shown as a "SaveDump" event - and then you can research this using TechNet or the MS knowledgebase on-line. If you are NOT getting BSOD's (i.e., no events appear in the log but the system still reboots), then it is likely being caused by some third party software (not a driver - because drivers almost always cause STOPs) or possibly power problems. Likely 3rd party software candidates are antivirus programs and utilities such as Norton. In that case, uninstall ALL third party software and then see if the system stabilizes. If the reboots go away, then you can try reinstalling 3rd party apps one at a time, with a pause for a few weeks inbetween to see if the reboots recur. Once they recur, you then know that the last app you reinstalled is the one that is causing the problem. Hope this helps!

I don't have the definitive answer, but I do have a couple of comments. First of all, I think you're fine with SP5 on this age of machine and it has been far more stable, in my experience, than SP6a. I generally am not in favor of doing a service pack as a shotgun approach to problem-solving.

Your best clue is going to come from setting the machine up to write a dump file when it re-boots. Once this is set up, most of the time, you'll find the dump file (an image of the RAM at the time ofthe crash) in a file called memory.dmp (or user.dmp) and usually it will be found in the c:\winnt directory.

You can run a utility called dumpexam.exe from the NT Server Resource Kit and, provided the symbol tables are installed, it will tell youthe address of the routine that choked and what module or driver was loaded at that address. Usually, it turns out to be a stale or corrupt driver.

However, in this case, you could also be seeing the beginnings of a hardware memory failure and that could be very, very tough to diagnose.

But the best starting point would be generating and analyzing the dump file.

Best wishes!

paul

If the machine is attached to an UPS, then depending on the model UPS and associated software this can really cause some headaches. Most types of UPS monitor software include scripts that will shutdown the host machine whenever selected events happen. This can be due to anything from a loss of communications with an UPS or from power surges and the like. Also, depending on the interface between the host machine and the UPS there could be a problem where as the computer performs a certain function which the UPS will recognize as the server actually telling it to shutdown (ie. thru a serial interface that becomes disrupted due to another function "polling" the serial ports).

Guys, I did follow all your advice, but the server keeps on rebooting at least once in a day. Got any other suggestions?

sounds very like a power issue, check your UPS if you have one, maybe run the server via the mains for a few days to see if that cures it. Could be the batteries in the UPS are starting to fail and the ups keeps dropping the voltage.

This question was closed by the author