Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations Mike Lewis on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

W2K crashes daily...

Status
Not open for further replies.

elwood78

Technical User
Jul 11, 2001
20
0
0
DE
we have a problem with a W2K-Server, that crashes daily without any error messages in the event logs... one week ago, we dated all BIOS and Firmware up, and it runs about 9 days 16 houres 9 Minutes and 27 secs ;-)

On that server are all(!!!) services. It is in a domain with only one W2K-Server, himself, and a NT4-Exchange5.5 Server. In the first 4 Month the server doesn't crash... then it crashed 1 or 2 times a week, and then daily...
And now, we hope that it runs 10 days...

can it be, that we must to make an 'only'-PrintServer?

many greetings
Elwood

PS: the server is a:
Dell 4400, 2 x 1Ghz CPU, 1 GB Ram (W2K SP2)
 
How many users are accessing it? Is the hard drive full?

What happens when it crashes - does it reboot? Blue-Screen? Freeze?

Let us know more if you can.

Pbxman
Systems Administrator

Please let Tek-Tips members know their posts were helpful.
 
hi,

there are about 30 users on that system... we have 2.5 GB free on C: and 85 GB free on D:... RIS is on D:, but there are no new PC-Installs at the moment...

it reboots... (wenigstens)

thx
Elwood
 
If it reboots, it's most likely set that way to reboot after a stop error occurs. What errors are in the event log?

Pbxman
Systems Administrator

Please let Tek-Tips members know their posts were helpful.
 
Ok,

we tried it, but there are no Messages in the EventLog.

Now we changed the System Board, and now it's about 12 days up...

So we are waiting on the next crash...

Thanks and a happy new year

elwood78
 
If your swap file is on C: it could be that the drive is filling up and then the system is rebooting.
 
Go to the control panel, select "System" and then the "Advanced" tab. You should see a setting in there to 'Automatically reboot' under the 'System Failure' heading. Deselect this to stop rebooting upon a system failure adn select the other options (i.e. "Write an event to the system log", and "Send an administrative alert").

Hope this helps.

Regards.
 
JimEkleberry: first: thank you... the swap-file is on C:, Size 1.5GB, 2.5GB are free... And is there not a Message in the eventlog when the drive is filling up?

wongyc2k: thanx, yep, we maked that, but nothing... black screen, and the server hangs... only power off/on 'helps'...

 
ok, it wasn't the SystemBoard... the server crashed after 16 days...

the server crashes only, when the users are working on it... so, something overflows... but what?

 
I don't know if this will help or not, but I had the same problem with Pro and it happened consistantly with the same program. It turned out that the driver for the video device was corrupt and only crashed like this under a specific use. Later I had a win 2k server do the same but it was services that would cause the crash. It was running an SQL server 2000 and the system services simply canabalized themselves. This turned out to be a virus surprise surprise. It was a development server and prone to these kinds of misshaps so I did a complete reinstall and continued on. I got plenty of messages stating could not find file. Also I noted that the virus scanner had deleted some of the infected files and that some of these had been crucial to system service. I imagine you've already ruled out viruses. Try reinstalling the services, if that doesn't work try reinstalling the OS. From the sound of it your running a 24/7 live server and this might not be the most viable option.

Hope you find my rambleing usefull.

Shawn
 
Hi...well I was wondering, after reading all the replys, if I could find what was the even log generated just before the system lost control/rebooted.
Could you tell the event log details also.

prasanna.
 
pallapalli: nothing... there is absolutly nothing in the enventlogs...

 
windows generates an event ID 6005 immediately after the booting up the machine which indicates that event log service is started. If you could find any event ID before tell us that.....
Okay....of all the ideas are....
1. Is there any Dell management softwares installed and taking controll of any hardware devices - if so then disable then for sometime since win2k's HCL is wide enough to accomodate management of most of OEM servers.

2. Are the users trying to access only the files on the server.....then probably you can enable the "file access" success and failure logging on the server to see if this is causing any problems (next time it crashs it should generate the even log even if any print job is creating this problem). Also checkout if any user is trying to print a large file...larger than the amount of RAM on the printer.

ABOVE ALL...when I was preparing for MCSE-accelerated (which i failed to pass) they were talking about installing the first processor and then installing the W2k server and later upgrading the machine to another processor.
OR install NT4.0 first and then upgrade OS to W2k.
You may want to consider this......

prasanna
 
I had a similar problem on a home pc running win2k prof. After a month of solid up-time, I would get the odd reboot. They became more frequent over the next month until it wouldn't even finish a boot. No indication in the log file of anything amiss. I suspected and tried many things, but the short of it was a bad DDR ram in the high addresses. Something to consider.
 
hello all,

the server is now up since 44 Days...

we think, that it was a problem with the mainboard. After we changed it and reinstalled the W2K SP2 it doesn't crash...

Thank you to all for your help...

many greetings
elwood
 
Can this be caused by fragmentation of the HDD? It seems to me that the more a system is run (this is any system), the more resources are temporarily ate up (RAM fragmented, etc). It might be wise (and more healthy in general) to run more than one server, so they can maintain themselves. Put them on direct cable connection, and make sure they log as asmin by default. Somewhere I read about a way to bypass the login screen in 2k (in 9x it's easy). You will need to do this. Have one server run a set ammount of time, then have it send a ring to the other (you will need a wake-on-ring capable board). The other comes up, online, and serving, then the first stops services, disconnects, and does a scandisk/defragment/shutdown. After a set ammount of time, it switches back to this server, as the other does the same thing. (hmm this gives me an idea...is it possible to send a ring to another box on LAN in the event of sys failure, like a backup?). I never have set anything up this way, but my old boss did. It proved to be a more reliable way of doing things.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top