My main site, and all my remote sites have their own internet access. In DHCP for all sites and in the NIC configurations of all the servers I have my internal DNS server configured as the primary DNS server

I've never liked the idea of putting an vulnerable internal/intranet server such as an AD box onto Internet. People do it, but it has to be done with extreme attention to firewall rules.

 and then I have the respective ISP's dns servers as the alternate.

Uou have that backwards. the ISP DNS should be the primary for the remote sites, and your AD/DNS as secondary. When then primary DNS is down, resolvers will usually take many seconds to timeout before trying the secondary DNS, making all DNS work very slow, even appears that DNS is completely down. Having all the sites use their ISP DNS gives you great redundancy vs having all the site come to your single DNS.

Here are my questions:

What triggers a machine, either server or workstation to give up trying to use the primary dns server and hit the alternate? Did this behaviour change from win2k to XP?

when a DNS query to the primary times out, the resolver will try the secondary DNS. But the applications making the queries may timeout quicker than the resolver.

The timing-out primary will be tried for every new DNS query. The resolver doesn't remember than the primary is down and go to the secondary first.

On my win2k server, even though it could contact the primary DNS server, it couldn't get an answer so it went to the alternate.

correct. no answer = timeout, so resolver tries the secondary.

 My workstations just gave up.

Only apparently. Actually, they were taking a long time to timeout and appeared not to try the secondary. And, the applications could have been giving up before the resolver gave up.

Is there a way to change this behaviour?

I understand that I can set up local internal DNS servers at each site with a dynamic DNS zone for my internal domain and then forward to the ISP's DNS server for external but I was really trying to avoid that.

the more infrastructure redundancy you have, the better. Setting up will be more work, but the maintenance is almost nil, and the benefits would be you would have avoided this problem.

Len


To Unsubscribe: http://www.ipswitch.com/support/mailing-lists.html
List Archive: http://www.mail-archive.com/imail_forum%40list.ipswitch.com/
Knowledge Base/FAQ: http://www.ipswitch.com/support/IMail/

Reply via email to