[Blood]

Hello again

Further to my question earlier, I am looking into setting up a second DHCP server on our network to provide redundancy. I intend to switch over from static IP addressing to dynamic addressing.

I have read a few MS Technet articles about DHCP configuration and, because we have a single-site small AD W2k3 network (2 DC's, 33 clients), I will configure both DHCP servers with split-scopes.

My question relates to how the servers will be accessed. Presumably, because both will be integrated into DNS, there will be no 'primary' DHCP server? I assume both will be 'equal' in the domain, especially as there is no option to configure a preference. If this is so, then fine. If not, will I need to specify a preference by other means?

I've never come across the term 'primary DHCP server' so assume that it does not exist per se.

Thanks

[L4ndy]

Blood

I think you are referring to the 80/20 rule.
If you only have 33 clients I wouldn't worry to much about setting up another DHCP server. Best thing for you to do is create a backup of your DHCP database and keep it handy in case you will ever need to restore it (I very much doubt you will ever use it though, DHCP is quite reliable).
You are going to waste your time and resources by trying to provide some kind of redudancy given the size of your network.
Yes, you are right there is no "primary" or "Secondary" DHCP. If you create and enable a scope in both servers and authorise both of them, then they all will respond to clients with a DHCPoffer packet the client then accepts the offer in a"first come first accept basis".
Have a look at this article for more info on DHCP: http://technet2.microsoft.com/window....mspx?mfr=true

Cheers

[]SK[]

I don't see anything wrong with adding a secondary DHCP server. Any downtime for any company can be a headache.
Your research looks positive to me.

[Blood]

@L4ndy: Thanks for the link. I've gone through most of the Deployment articles on Technet, so seeing that was great because it brings it all together.

@]SK[: Thanks for the comments.

The reason why I wanted to create redundancy was because 1. There are two domain controllers available and 2. I would rather be safe than sorry.

At present, FRS is working between the two DC's and there is also a primary and secondary DNS server available. Seemed logical to setup a fail-over DHCP server. The actual time spent setting it up is minimal, it's the research that takes the time.

Thanks also for the info about backing up the DHCP records - I had gone through this when I setup the first DHCP server several years ago and had forgotten all about that.

I appreciate the help.

Cheers!

[Blood]

I have another question about providing DHCP redundancy, if I may.

I will setup a second DHCP server with the same scope and opposite excusions. That's fine. However, the documentation I have read, and the link which L4ndy provided refer to APIPA (Automatic Private IP Address) which provides redundancy for a client and assigns an address if a DHCP server does not respond to a renewal or new lease request.

As far as I can tell from various Technet and MS KB articles, APIPA is fully automatic. The only configuration options appear to be limited to turning it on (which is the default setting), or off.

According to the documentation, if APIPA assigns an address, then the address assigned is 169.254.0.0/255.255.0.0 etc. I will have several static IP assignments on my network which denote printers, DC's (for DNS etc), an ancient 486, the data\mail server and the gateway. My present IP addressing scheme uses 192.168.0.xxx

If APIPA is activated because both DHCP servers are down, and some clients need to renew their leases then presumably they will get 169.254.xxx.xxx. Now, I've never had to add a subnet to a network, but am I right in assuming that if an APIPA address is used, then the affected clients will not be able to communicate with the printers and, more importantly, the server which hosts our data and mail?

I was wondering if it is possible to configure APIPA to assign an address within a user-defined range i.e. 192.168.0.xxx?



Overkill, perhaps, but I have learnt over the years that it pays to protect and preserve data and access to data via any means possible.

Cheers!

[Ossian]

APIPA cannot be changed from 169.254.x.x, also please note it does not give a default gateway, so no PC can access the internet with APIPA. Although it is "sold" as an emergency way of letting PCs talk to each other, in reality it is a warning that something is not working properly

[joeqwerty]

This is a summary from RFC 3927 (which is what APIPA is all about):

To participate in wide-area IP networking, a host needs to be
configured with IP addresses for its interfaces, either manually by
the user or automatically from a source on the network such as a
Dynamic Host Configuration Protocol (DHCP) server. Unfortunately,
such address configuration information may not always be available.
It is therefore beneficial for a host to be able to depend on a
useful subset of IP networking functions even when no address
configuration is available. This document describes how a host may
automatically configure an interface with an IPv4 address within the
169.254/16 prefix that is valid for communication with other devices
connected to the same physical (or logical) link.

IPv4 Link-Local addresses are not suitable for communication with
devices not directly connected to the same physical (or logical)
link, and are only used where stable, routable addresses are not
available (such as on ad hoc or isolated networks).

You can read the entire RFC here:

http://www.faqs.org/rfcs/rfc3927.html

[Blood]

Thanks a lot to both of you for your replies and for the link.

Thinking about it further, the question is redundant as both DHCP servers will be installed on domain controllers. If both DC's go down, then clients will not be able to access anything anyway because there will be no DNS servers available, no AD and, therefore, no authentication.

Cheers!

[Blood]

OK, I was going through the options to make sure I knew how to create reservations etc and under Server Options is No. 045 NetBIOS over TCP/IP NBDD | NetBIOS over TCP/IP NBDD address(es) in priority order. As this is central to my reason for moving over to DHCP from static IP addressing, I have been trying to find out how a NetBIOS Datagram Distribution Server is setup. Although many websites quote Microsoft parrot fashion, none that I have seen offer any practical advice or explanation with regard to NBDD.

So, at risk of making myself look daft, can anyone help me understand this please? Is the NBDD server automatically setup? Is any user interaction required? If I set one up, will it be OK to place it on the same server that hosts DHCP and DNS?

Thanks

[Blood]

If I specify the primary domain controller's address will it cause problems? I cannot find anything about setting up a NBDD server, so if ayone knows anything about this, I will be very grateful.

Thanks.