Clients not moving to alternate DNS server

[ROSS HANSON]

I have two DNS servers. All the clients, there are about 2000 of them, have WinME running with the two DNS servers listed in the correct place.
One DNS crashed but only about half the users correctly went to the alternate DNS server. The others had to be "moved by hand" to the alternate server. After checking the settings in
the clients that did not move I found nothing wrong. Is there a "secret" that I am unaware about?

[Steven E. Protter]

This is not a log to go on.

There could be issues with subnetting, routers, ports on firewalls, could be anything.

Is there a common factor in those clients that did not correctly make the transition? Are yo usure they have a secondary DNS server in their config?(run ipconfig).

Did you test this failover configuration in remote locations.

Check the event log or syslog(HP-UX) on the DNS server, there could be error messages. If the DNS server is HP-UX, you might be blocking the requests with inetd.sec.

The real secret is a thorough test and implementation plan. You should still be able to check the secondary server by bringing down the named daemon or DNS service.

You need to make sure this failover configuration works in test.

We control this behavior by having a common client image. Everyone gets the same basic client(W2K) with extra software added on with automation later. Still, we sometimes run into router configurations and subnetting issues that surprise us in testing.

You also need to tell us the OS and BIND version of the two DNS servers.

SEP

[Kevin Wright]

Obviously this is a WinME problem.. A DNS server does not control client failover. If you were able to move them 'by hand',then the networking is OK. Talk to your windows admin.