WINS/DNS/DHCP redundancy

[elcephus]

Hello All,

I am looking for information on the best practice on how to setup a server with the above mentioned services (WINS/DNS/DHCP/RAS on it as a backup/disaster recovery/fail over server in a remote location. Any suggestions?

 

[ShackDaddy]

Well, first the remote location has to be reachable by all the clients who normally use those services. If you have a WAN link, then DNS and WINS aren't that hard to set up. DHCP, on the other hand, will be quite difficult, since it usually is highly dependent on broadcasts on the local subnets. If you want to have a DHCP server waiting over a WAN link, you need to configure a system to act as a DHCP relay server or configure your network devices to forward DHCP requests across the WAN link to your waiting server.

For RAS you will just need to have a separate phone number at the remote location for users to dial if the main RAS server is unavailable.

As far as actually configuring the first three services, you should do the following:

For WINS, you would just install WINS on the remote server and configure it as a replication pull partner of your main server. You would also configure it as the secondary WINS server for all your clients.

For DNS, you would configure the remote server as a secondary DNS server and when you add zones, you would pick 'secondary' zones and put in the address of the main server. It would periodically be alerted to new records on the main server and would pull copies over. You would also configure this server to be the second server in the list of client DNS servers.

For DHCP, assuming that you've gotten the relay/forwarding of DHCP requests squared away, you would want to configure a scope on the remote server that has addresses in the same subnet as the existing DHCP server. You would even use the exact same scope, but you would mark a large swath of that address range as excluded addresses. Then you would exclude a small range of addresses in the current scope on your existing server. That small range of excluded addresses on the existing server would be the range that you would have live on the remote server. That way each server would respect the reservations given out by the other server but would be ready to hand out addresses in the same range when needed.

That's a very brief overview. When/if you set all this up, I'd get more specific details about the implementation from TechNet or any of the good books out there. Or from the helpful people here.

ShackDaddy

 

[GlenJohnson]

If it's a backup server, why do you need it at a remote location? Just install a secondary server at you're location and let it do it's thing. ShackDaddy is correct on the how to, but a remote local can be a pain. Plus, you'll get you should see better network performance, (Except when the servers are synching, and you can have them do this at night), and rest assured if one goes down, the network keeps on going without you doing a thing. Just a thought. (If you're worried about something like a fire, do a second server first, then the remote server. You'll have gotten practice from the first.) Good luck.

Glen A. Johnson
If you're a kid at heart, check out Tek-Tips Disney
"Grasp the subject, the words will follow."
Cato the Elder (234 BC - 149 BC); Roman orator & politician.

 

[elcephus]

Thanks for the advice. The resaons for the remote location is for redundancy/Disaster recovery purposes. Can any of you guys offer any good books that should help me when settings this up? Much appreciated. Thanks.