Wierdest thing ever

[arisythila]

Okay, So I built a server for this company I'm affliated with. I built two of them. One as a primary one as a backup. So I install the first one, a few days later it starts to randomly crash. So I bring the backup down there, and put it in. Bring the primary home, and run various tests, hard-drive/memory/CPU tests on it, Been running flawlessly, I cannot get it to crash. Meanwhile, The backup server starts to randomly crash...

The Primary works FINE at my house, but crashes at the company.
Same with the backup, I was running stress tests on the machine before I brought it down there. passed for 72 hours..

I am stumped. I thought maybe power spikes or something, So I put a UPS down there. Still crashes. Any ideas why it would crash there, but not at my house? (I am 2 towns over)


Thanks

~Mike

[WeStSiDePLaYa]

Okay, So I built a server for this company I'm affliated with. I built two of them. One as a primary one as a backup. So I install the first one, a few days later it starts to randomly crash. So I bring the backup down there, and put it in. Bring the primary home, and run various tests, hard-drive/memory/CPU tests on it, Been running flawlessly, I cannot get it to crash. Meanwhile, The backup server starts to randomly crash...

The Primary works FINE at my house, but crashes at the company.
Same with the backup, I was running stress tests on the machine before I brought it down there. passed for 72 hours..

I am stumped. I thought maybe power spikes or something, So I put a UPS down there. Still crashes. Any ideas why it would crash there, but not at my house? (I am 2 towns over)


Thanks

~Mike


is it hotter where they are storing them?

ive seen alot of smaller companies just jam them into the smallest, out of the way, location. and that becomes a oven.

[Sorrow13]

my bet is on heat or maybe something got bumped in transit. maybe reseat ram?

[arisythila]

Yeah, Temps are way below hot.

32c CPU, 35c PWMIC, and 38c NF4 Chipset

All pretty normal..

~Mike

[arisythila]

my bet is on heat or maybe something got bumped in transit. maybe reseat ram?

Yeah, but I brought the primary server home, and works flawlessly.

~Mike

[Sorrow13]

hmm. that is wierd.
how about magnetic/electrical interference? have you seen where the server was sitting? Next to any large cables/machines?

only other thing i can think of is maybe power spikes (but not on power cables) ie network or (hope not) telephone connection. Seems like that would crash other computers in the area, but i cant say.

double edit- maybe a computer on network has a virus, and is passing it to the servers?
http://www.support.microsoft.com/?scid=kb;en-us;903251

[phelan1777]

are then in a closed in area? If so, how close are the walls...........maybe something running close by that you cant see that is giving off interference.

Lines running in the ceiling, or the floor? Just an odd thought.

dimensional interference, like a spirit or somethign ;-) (I kid I kid)

[arisythila]

Magnetic/Electrical interference, It sits next to a CRT monitor. I think thats the only thing that would cause interference. Im not sure what is in the walls. The server that was in there before had no problems, before I rebuilt it.

Virus... Still works flawless from home, no crashes in 72 hours.

I think the UPS I put in there was too cheap (60 USD) I think Maybe Ill pick up a 500 dollar UPS see if that helps.

~Mike

[couppi]

Taken from Power and You (a guide to protecting your computer) (http://xtremesystems.org/forums/showthread.php?t=74747)


Finally though, I would like to make one final point about UPS’s. Unless you get an online UPS (one that is constantly on and constantly recharging itself), these do NOTHING for line noise at all… and online UPS’s are very rare today, given the additional costs of ownership, manufacture, and – frankly – safety risks.

Unless it's an online UPS, it probably didn't help. Also, maybe a computer connected to it or the network software or something is causing it to crash for some reason.

[phelan1777]

umm, maybe someone is trying to hack it?

[Serra]

Hrmmm.. I should probably update that as there has been an increasing trend to making line conditioning a built-in feature for the last while...

My thought is that the crashes are software-related, because you have clearly ruled out hardware. What do these servers do?

Serra

[arisythila]

Taken from Power and You (a guide to protecting your computer) (http://xtremesystems.org/forums/showthread.php?t=74747)



Unless it's an online UPS, it probably didn't help. Also, maybe a computer connected to it or the network software or something is causing it to crash for some reason.


Windows 2003 doesnt say anything, no program errors or anything. Logs are completly clean.

~Mike

[arisythila]

Hrmmm.. I should probably update that as there has been an increasing trend to making line conditioning a built-in feature for the last while...

My thought is that the crashes are software-related, because you have clearly ruled out hardware. What do these servers do?

Serra


File server. Thats it. If it was software, It would say why it crashed. Logs are completely clean. no errors. I have also been putting loading the server that has been crashing at the company, at my house, and it hasnt been crashing.

Works flawlessly at home, crashes at the company.

Both Primary, and Backup server. Same thing. Works perfectly at my house, bad at the company.

~Mike

[faruquehabib]

if you think someone is trying to hack it, then run it at the company where it normally runs, but simply do not hook up the cat 5 cables , just let it run and not do it actual job. might eliminate some possible problems

[Serra]

Alrighty, new list of questions:

1. Are they using any custom software packages?
2. What OS is on the server?
3. What OS are the clients using?

My thinking is that this may be OS-related in some manner (OS's often do a better job of generating error reports based on apps than OS issues in my experience.... assuming as well that you have it properly configured to collect the information that you need...).

[arisythila]

if you think someone is trying to hack it, then run it at the company where it normally runs, but simply do not hook up the cat 5 cables , just let it run and not do it actual job. might eliminate some possible problems

Its only connect to an internal network... no access to the internet.

~Mike

[phelan1777]

Its only connect to an internal network... no access to the internet.

~Mike


Could be a disgruntled working in the company;)

[arisythila]

Alrighty, new list of questions:

1. Are they using any custom software packages?
2. What OS is on the server?
3. What OS are the clients using?

My thinking is that this may be OS-related in some manner (OS's often do a better job of generating error reports based on apps than OS issues in my experience.... assuming as well that you have it properly configured to collect the information that you need...).

Just Windows 2003 Server, Clients are using Windows 2000.

Primary and Backup Servers at my house works fine, no problems at all. If they cause errors at the company, should cause errors at my house.

~Mike

[arisythila]

Could be a disgruntled working in the company;)

Family owned, Family worked.

~Mike

[situman]

I built my friend a computer once. Works flawlessly in my house, plug it into his house, it would freeze pretty quickly. Reason? His sockets and wiring sucks. It couldnt provide enough juice. Bought a UPS, it would last about 30mins before freezing.

[arisythila]

I tested voltage, voltage was 120v so getting good voltage, but not sure on the frequency. I just bought a new 500 dollar UPS, that should activly filter the line. Im going to give it a try.

~Mike

[derektm]

I tested voltage, voltage was 120v so getting good voltage, but not sure on the frequency. I just bought a new 500 dollar UPS, that should activly filter the line. Im going to give it a try.

~Mike

Good luck! :toast:

[epion2985]

try putting it at a different location at the company and see if it still crashes.

[Kalway]

How exactly is it crashing? blue screen? what happens?

If it blue screens with the "IRQL_LESS_THAN_OR_EQUAL_TO" one then it's bad memory.

We're working on upgrading one of our terminal servers at work from a dual p3 700 to an A64. Stupid thing is giving me issues just trying to install win2k3 and now I installed win98 and when it boots it says not enough memory to load everything. Gah I hate windows sometimes.

[sobol]

Hi
Maybe first you could post full spec of those servers?
For me its looks like problem with hard disk operations?
In your home you are testing CPU/Memory stability, whilst in your company those elements are probably not loaded. I assume that it is a file server and in your company they use a lot of its resources (many hdd operations?). Try doing extensive use of your hard drivers, maybe change ide controller drivers?
Regards.

[mursaat]

Its only connect to an internal network... no access to the internet.

~MikeI'm with faruquehabib, try not to plug the ethernet cable at all, so you can eliminate "environmental" problems.
I built my friend a computer once. Works flawlessly in my house, plug it into his house, it would freeze pretty quickly. Reason? His sockets and wiring sucks. It couldnt provide enough juice. Bought a UPS, it would last about 30mins before freezing.Maybe it's plugged on a very crowded line? You can get the 120v, but not enough amperage or something. Try to plug only the comp and monitor?

Also as someone else said, a bit insight of the server hardware and the errors output could help ;)

[arisythila]

I'm with faruquehabib, try not to plug the ethernet cable at all, so you can eliminate "environmental" problems.Maybe it's plugged on a very crowded line? You can get the 120v, but not enough amperage or something. Try to plug only the comp and monitor?

Also as someone else said, a bit insight of the server hardware and the errors output could help ;)


Computer hardware is

DFI Ultra-D
3800+
OCZ 512x2
2 x 120gb drives (mirriored)
450W PSU
CDrom
Floppy
6200TC

Doesnt report any errors. IT is completely errorless, From how it reboot, it reboots like someone pushed the reboot key on the front of the case. No file errors, no hardware errors, Machine works fine. It has to be enviromental.

~Mike

[Complete]

Here is my only ideas.

I would use a fluke meter and double check the power line. Then go to another outlet and check it again.

Try the computer in a different outlet see what happens.

Also, if there is a high power outlet line? Like 480 or 240 that is near the computer?

Double check your computer power line. It sounds like it could be a frequency/power issue.

just my .02

[LowRun]

I made a puter for a friend with old parts some weeks ago but recently it started to reboot on its own so she brought it back for me to check it. After running ok for the time needed to do some quick test it started to reboot for no reason. To make a long story short i had to up vcore a little to get it to run with no more reboots, everything was set on auto in bios and no OC before i upped the volts.

[derektm]

Computer hardware is

DFI Ultra-D
3800+
OCZ 512x2
2 x 120gb drives (mirriored)
450W PSU
CDrom
Floppy
6200TC

Doesnt report any errors. IT is completely errorless, From how it reboot, it reboots like someone pushed the reboot key on the front of the case. No file errors, no hardware errors, Machine works fine. It has to be enviromental.

~Mike

I don't think a DFI ultra-d on nforce4 is good for a server that needs to have great uptime etc. This is for a company...

[arisythila]

I don't think a DFI ultra-d on nforce4 is good for a server that needs to have great uptime etc. This is for a company...


a DFI Ultra-D setup correctly will have awesome uptime. I have one running at the house, that has been up for almost a year now... 378 days.

Its not the hardware, It cant be. Its something enviromental. It would crash at my house if it was hardware/software. I am running the hard drive like no ones business, while running SP2004 CPU/memory blends on priority 10. Its been up for 86 hours now. no problems. I have been copy and pasting 20 gigabyte files back and forth from another computer to it thought the network.

~Mike

[celemine1Gig]

If it's really line noise causing those reboots, I'd try a different PSU.

Still I can only agree with derektm: A DFI isn't really intended for Server use. I would've bought a Tyan or something similar. But that's just me. As long as it works as it should, everything's OK.

[arisythila]

If it's really line noise causing those reboots, I'd try a different PSU.

Still I can only agree with derektm: A DFI isn't really intended for Server use. I would've bought a Tyan or something similar. But that's just me. As long as it works as it should, everything's OK.

DFI works great. Very stable.. IT was the socket, I plugged a 500 dollar UPS in there, and no more problems...

~Mike

[Skip]

500$ is an expensive solution! but glad to see it worked.

[situman]

DFI works great. Very stable.. IT was the socket, I plugged a 500 dollar UPS in there, and no more problems...

~Mike

Told u so.

[arisythila]

Yeah, I guess it wasnt actively filtering the power. I bet you the Hz were off maybe fluctuating too much for the power supply.

98 hours of uptime now.

Thanks for the help guys.


~Mike

[LazyBum]

Most likely a motor of some sort dropping the line voltage during its start phase. This time of year it may have even been the compressor motor on the A/C if it's fed from the same panel.

[darkhelmet]

um....if it's just a file server.....use linux......