server crashes / blanks out.

[TahoeEddie]

running a windows 2000 server.
i have been increasingly experiencing an issue with this machine where it will unexpectedly stop running.
the symptoms are that the server will just quit, quite literally.
all of the system's power will remain on (power lights on the computer, and monitor are on), but nothing on the screen - no blue screen, no cursor, not anything.

the only way to recover the machine is to do a hard reboot.
this has been happening since may.
and it is becoming a frighteningly frequent occurrence.

i have checked the event logs to see if there is anything unusual, and i cannot seem to find anything that would indicate why this is happening.

the machine is up to date on all of the patches and fixes.

i do not know if this is a software or hardware problem, but i cannot seem to find any solutions on the software side.
so i was hoping that someone might be able to shed some light as to what would cause a machine to have its power remain on, but just stop all computing, and pose some solutions as well?

thanks,
eddie.

[jdc2000]

Hi. Welcome to VirtualDr.

Does the light on the monitor go from green to orange after this happens? There are two possibilities I can think of. If the PC is set to go into power save mode, it could be a problem with that. On a server, I usually disable ALL power saving features, in the OS, BIOS, everywhere. Or, it could be a hardware problem. The fact that you are getting a blank screen would seem to indicate that it is going to power save mode, then crashing. Check to make sure that all fans are operating, and that heat buildup is not a problem.

[TahoeEddie]

thanks jdc2000 for the quick post.

no, the monitor light stays green.
i had thought about those powersave modes, and i turned as many as i could find off.

i also went so far as to never allow the screensavers to start, so that it should never lock the computer.

some additional information that i forgot to mention earlier,
this problem happened to me today, while i was actually standing there (finally i was able to see it in action).

i was scrolling through the event logs, and the display just went away.

this server is located in a closet, like many; and there is a telecom box on the wall next to it.
and just before the server went out, there was a clicking noise in the telecom box. i don't know exactly what all is in that box, or even if that had anything to do with it, but it was a strange coincidence.

but note that that box clicks quite often, while the server crashes are less often.....

so in any case, the computer should not have been trying to go into any sort of powersave mode while i was using it.

-eddie.

[jdc2000]

Are you connecting to this server through a KVM switch or some other remote access device setup, or are your monitor, keyboard, and mouse connected directly to the server's video card, and keyboard/mouse ports? If you are not connecting directly to the server with your display and KB/mouse, I suggest you connect directly to the server PS in order to determine if it is the server causing the problem, or your connection that is bad.

[Train]

Do the check out that jdc2000 listed. And if it still locks up, I would say you are getting interference from the telcom unit of some sort. Seen that happen with this computer and the interference was in the answering machine. Moved it about 10 foot away and no more problems with the computer. Check where the KVM, if you use one, cables run. Might need to move them is all.

[TahoeEddie]

hey guys, thanks again for continuing to post suggestions.

here's the setup.
the keyboard, mouse, and monitor are all plugged directly into the computer.
they are not running through a switch.

and train, just a couple of questions about your recommendation about moving the machine.

not knowing much about the kinds of interference that you think might be causing it, i was wondering if you might be able to help me understand the basics behind that. i will look into seeing if i can move it anywhere. the space is tight. i might have to move the box into a completely different room.

might take me a while to post back to this recommendation.
but also note that the machine has been set up in this geographical layout for about 1.5 years, and it is only since may that this has been happening.

again, i appreciate all of your suggestions.
thanks,
eddie.

[Train]

Might get the folks to check out the tecom box. What I think you are experiencing is called RFI {Radio Frequency Interference.} But as it was in there so long, I would think maybe the telcom equipment may be the problem.
Just a idea, but I have had to fight RFI before and that is where I would them look first. Could be just a bad relay causing the problem. By the way let them know what is happening so they can check their stuff out real good.

[Admiral]

Sometimes this sort of thing happens if the server fans start to fail or the vents are clogged. Could be worth a look.

[TahoeEddie]

i was in there last night checking it out.

the fans run fine.

i had to manually run the backup since the server had failed prior to the scheduled backup time.

we are running backupexec.
and about 4 minutes into the backup, the air conditioning went on in the building, and caused a power fluxuation in the building, and coincidentally the server crashed, this time with a blue screen, and stop error noting ntfs.sys.

once again, i don't know if there is any direct correlation, espeically since the server is plugged into a surge protector/UPS.

does anyone know of a away to test the efficacy of the UPS?

thanks,
eddie.

[Train]

I would say the batteries need replacing now. Might check the manufactors website under Support and they may have what you want.

If the AC unit causes that much trouble, I suggest paying somefolks to fix your electrical system. The AC unit is drawing to much current for the system to handle. Yes, I am well aware that a new service may need to be hooked in and that is EXPENSIVE. But cheaper than losing everything. Just had to do just that at work.

[TahoeEddie]

ok people, just in case you were interested..

after running motherboard monitor, and not finding anything,
even went and ran memtest (got a couple of good passes) - the machine crashed during memtest even after a few successes.

replaced the memory with a new stick of 512MB RAM.
it has been running smoothly for 20 hours now.
which is better than we were getting.
eventually, yesterday we were not able to bring it back up at all.

this might have been the problem all along.

odd that memtest passed though.

thanks for all of your input along the way!

-e