NT4 PDC locks the Domain Several Times a Day.

[cupper]

I have a customer running Win NT4.0 SP6a + latest updates (Knowledge base article q299444). 2 to 3 times a day the workstations lose connection to the shared drives on the PDC. If you go to Network Places, you cannot open the Domain to see all the computers on the network. The PDC is running WINS, no DNS or LMHOSTS lookup. Nothing is posting in the event log to reflect a problem. I have tried stopping and restarting services such as Browser and WINS but it doesn't help. The only way I have found to recover is to reboot the server. I have also uninstalled the WINS service as well as deleting the WINS folder and then re-installing it.
I am to the point where I am going to format the C: drive and reload NT4.

Any ideas would be greatly appreciated.

Sincerely, Cupper

 

[ShackDaddy]

I wouldn't expect the browser and WINS services to be a problem if you are getting dropped drive mappings. When a client loses a connection to a mapped drive, I would look at the networking and server services on the server, rather than the WINS and Browser services, since they are mainly helpful for initially finding the server.

Even if the WINS server is off-line, if the clients are in the same subnet as the server, they should be able to find the server to establish mapped drives after client reboots. Here are a few things I would look at:

1. Licenses.
2. Event log notices related to the Server service.
3. Event log notices related to the NetLogon service.
4. Whether the PDC was install from MSDN Subscription media, which has a hardcoded user limit in NT.
5. Whether your security/system event logs get full and prevent new connections.
6. When the PDC can't be reached, can it be pinged from a client? Have you ruled out actual network problems in your hubs and switches?
7. Are you seeing information in your event logs related to changes in settings on your network adapters? Perhaps speed and duplex is changing spontaneously due to a flaky connection to the switch.

Let us know what you find,

ShackDaddy

 

[cupper]

Turned licensing off. Don't beleive I am seeing anything of significance in the event logs. It was not installed from MSDN. The PDC can be pinged and if you ping the PDC name it is resolved and does ping. Cleaned up the network a fair amount. Installed two 48 port switches. Not one logging related to network adapters.
Thanks for the ideas. I will let you know what the problem was or what I did to fix it. My plan at this point is to reload the PDC from scratch.

Sincerely,
Cupper

 

[vanillawafer]

Same identical problems here, so when anyone has any information, please let me know asap.

It's almost like someone has written a script and is causing several of our "same" users to be locked out everyday, all at the same time.

This is driving me nuts. It's happening about 15-20 times a day.

 

[Winntguy]

The same thing is starting to happen to us. Right now it is about once a week. I have noticed that I will have a few or very many event 2019 errors. A Reboot is the only way I have been able to recover. I am able to ping the server, but I cannot start / stop services. I get an "RPC failed" error when trying to use either the services or server manager programs.

This is the event that has been present around the time the PDC stops.

Event ID 2019

The server was unable to allocate from the system nonpaged pool because the pool was empty.

Any help would be appreciated.

 

[cupper]

To Winntguy - Check Microsoft on this -

http://support.microsoft.com/default.aspx?scid=kb;en-us;294346

See if that helps. I think you have a different problem than what I am seeing. My problem is definitely associated with traffic to the PDC. This particular customer was using roaming profiles for approximately 120 users. I removed all the roaming profiles. My problem stopped for about a week and a half. It is now happening again intermittantly. Maybe once a day. I can duplicate by trying to copy a large file to a workstation from the PDC. The PDC needs to be rebuilt before I trouble shoot any further. They had RAID 0 sets with software mirroring running. Once I have a clean server I will continue to trouble shoot. (If that doesn't fix it of course).

I will keep you posted.

Cupper.