Clients losing DNS server

[Guest_imported]

I have a 2k DNS server. Every now and again NT and 2K clients will lose their internal DNS connection to it . Using

ipconfig /registerdns

kicks them back off and everythihng is fine. Some 2k clients don't seem to have this problem at all! The message from DNS is "negative cache entry".

Anyone had this before?

Cheers

 

[tom11011]

Could the clients that loose the connection be on the same hub/switch?

 

[Guest_imported]

Nah, all other connectivity is fine. If you access via IP rather than DNS then everything works fine 24/7.

 

[brontosaurus]

did you do a /flushdns prior to the register? A negative cache entry means that the client currently believes that the DNS server is offline and will continue to do so until the TTL has expired.

 

[Guest_imported]

Hi

Never used /flushdns before a /registerdns. The actual message i get when i do a /displaydns on a machine that has just lost it's DNS is "Negative cache entry for no records".

Why would only some of the clients think the DNS server was offline?

 

[brontosaurus]

Why...I would need to know much more about your network/servers to answer that question, unfortunately. Are all your machines on SP2? Anyway, an IPCONFIG /FLUSHDNS will remove any negative cache entries (actually, it will remove all cached entries...). Do you have only one DNS server? If so, consider bringing up a secondary and configure your clients to query it as an alternative. Any server can do this, and it doesn't require much "power". Lastly, there is a way to turn off this negative caching which involves a registry edit. I am not a proponent of this method but you should know about it anyway. Open REGEDT32 and go to :
HKEY_Local_Machine\System\CurrentControlSet\Services\DNSCache\Parameters\NegativeCacheTime.
Set the value to 0 to disable negative caching.

 

[AndyMac]

I have had this problem before with the main side effect being the loss of group policy resolution and the opening of restricted desktops. This only happened on around 1% of logins though. In our case the issue was with the 3Com layer 2 switches that were installed. The DNS event log was showing that the effected workstations had temporarily lost contact with the DNS server. Changed kit for Cisco and no problems after.

 

[strai81]

I am having this same problem. It's with win2k workstations only. Left work one day and everythings fine and came in the next morning and the problem was there. We are using an as400 for dns and dhcp. The problem only seems to be with internal host. I can ping the hostname of an external host but not an internal. I can do an ipconfig /registerdns and everything is fine until that pc is shutdown.
When it's booted back up the problem is back.

Any suggestions??????

Thanks,

Steven Y.