Multiple DHCP servers

I have 2 DHCP servers setup in my domain using the 80/20 rule that Microsoft
recommends. I have problems with the 2nd DHCP server however, because it
continues to run out of available address leases as it would appear clients
default to pulling their addresses from the DHCP server with fewer addresses.
Is there a way to default clients to pull their address from a specific DHCP
server first and when unavailable, from the second DHCP server?

 

In

Sounds like the faster server. Switch them around.

--
Regards,
Ace

This posting is provided "AS-IS" with no warranties or guarantees and
confers no rights.

Ace Fekay, MCSE 2003 & 2000, MCSA 2003 & 2000, MCSE+I, MCT, MVP
Microsoft MVP - Directory Services
Microsoft Certified Trainer

Infinite Diversities in Infinite Combinations

Having difficulty reading or finding responses to your post?
Instead of the website you're using, try using OEx (Outlook Express
or any other newsreader), and configure a news account, pointing to
news.microsoft.com. Anonymous access. It's free - no username or password
required nor do you need a Newsgroup Usenet account with your ISP. It
connects directly to the Microsoft Public Newsgroups. OEx allows you
o easily find, track threads, cross-post, sort by date, poster's name,
watched threads or subject. It's easy:

How to Configure OEx for Internet News
http://support.microsoft.com/?id=171164

"Quitting smoking is easy. I've done it a thousand times." - Mark Twain

 

Ace, If I wanted to set priority to one over the other, is there a way of
doing that? The faster of the 2 servers is indeed the one that keeps running
out of addresses, however, I want the other server to be the primary to
distribute addresses. Does this make sense?

 

<>anRkcnVtbWVyYm95

It does indeed make sense, but AFAIK there is no way to do this (not
if both servers are on the same subnet as the clients). In a routed
environment where you are using DHCP relays, this is probably
possible; however, we don't know if this would apply to your network.

Regards,
Scott

 

Hi Scott, thanks for your quick response. Both of these servers reside on
the same subnet so I agree with your theory that it would be difficult to
give priority to one over the other. I do also agree with Ace's response to
give the larger pool to the one that acknowledges more requests...I was just
trying to avoid this as this server is more heavily used.

 

Hi..! I have this message in my DHCP server.. event id 1020

Scope, 172.17.20.0, is 100 percent full with only 0 IP addresses remaining.

I have running Windows 2000 Server .. the case is that when I try to
put a new pc and it try to take dhcp automatally it present the error and I
have to put ip address manually..!
I have 10 pcs on my domain

How can I solve this warning?
please, send me a email.:
thanks

 

You should adapt you scope with a bigger range of IP addresses.
We do not email people, that's why this is a forum, to share knowledge.

 

Personally I don't like this rule because of issues like the one you are
facing.
Another option is to setup 2 different ranges in the same subnet with enough
addresses.

 

I mean, 2 # ranges, on in each server, of course.

 

<>anRkcnVtbWVyYm95

I wouldn't be too terribly worried about server utilization; the
overhead for DHCP is fairly minimal, and adding addresses to an
existing scope would be, I think, less of a worry than adding DHCP
from scratch to an existing workload.

Regards,
Scott

 

In

Sorry, unfortunately it doesn't work that way. Put the larger pool on the
faster box, as has been suggested.

Ace

 

jtdrummerboy,

When a DHCP client decides to look for a new address or refresh the one it
has (at 50% of the lease duration) it does a broadcast on the network looking
for available servers. In all cases, it will accept the first one that
replies back -- this is why ACE knew it was the faster server. Since
broadcasts don't cross routed boundaries by default only servers on the
subnet are in the running. You can use IP Helper (IP Helper-address in cisco)
to allow for DHCP requests to cross the boundary.

An 80/20 scenario is implemented to provide higher availability for DHCP in
the case that one server goes down. It should never be used to extend the
scope. You should use larger scopes or superscopes to grow capacity. When
the primary server goes down, you have the time that it takes the 20% to get
used to fix the problem before you start impacting the users.

It should also be noted that you shoudl authorize both servers to host the
ENTIRE range and use exclusions to mask the 80/20 scenario. Otherwise you
get NAK poisoning when a client goes to refresh and hits the "other" server
for some reason. You want it to respond with a differnt valid response
rather than issue a NAK telling the client that the range isn't valid. In
some cases, the client doesn't try again and the IP stack goes down.

Hope this helps.
--
Ryan Hanisco
MCSE, MCTS: SQL 2005, Project+
Chicago, IL

Remember: Marking helpful answers helps everyone find the info they need
quickly.

 

In

NAK poisoning, I like that.

Your suggestions of putting DHCP elsewhere such as on another subnet and
using IP helpers (similar to DHCP relay agent in Windows) is a great idea.

Ace