1

I’m trying to change the IP address range of my network using DHCP. My current range is 192.168.1.0 /24 and I’m trying to change it to 10.20.1.0 /24. I have a scope in DHCP setup for my current range which is working fine. I’ve created a new scope with the new range, activated it, deactivated the old one, and when I try to get a workstation to renew it’s IP, it just hangs. As soon as I activate the old scope, it instantly gets an IP from that.

It’s worth mentioning I currently have my server running 2x IP’s of 192.168.1.1 and 10.20.1.1 as I’m trying to limit disruptions and allowing my users to continue working while I test this. Not sure if this is an issue?

4 Spice ups
2

I’d try to releasing your lease and then get a new one. OR just reboot.

3

What happens during a DHCP renewal is slightly different than what happens during a discover. When the computer has an active lease and you ask it to renew, it will contact the DHCP server that last issued that lease. By deactivating the old scope, it stops responding to those requests so a renewal will time out.

As Jim alluded to, releasing and then subsequently requesting a new lease works different in that it will stop using its current IP address (it will send a DHCP Release to the old server which may or may not respond to it at this point) and then rather than doing a renewal it will send a DHCP Discover broadcast to start the process from scratch.

While you use the same /renew option with the ipconfig command, it just behaves slightly differently based on the current state.

4

Even if I do a release and then a renew, it’s still the same.

5

Logging is going to come in handy here. If you run DHCP on Windows Server, go to the server’s IPv4 properties, Advanced tab and check where the Audit logs are stored. Take a look at the files there. Each files covers a day and the top of the file contains a handy reference list explaining what each status code means on each line. See if you can spot any errors here that give a clue.

6

How do you have 2 IPs on the server? How are you doing the new subnet on your switches, same VLAN or different VLAN? You can only have one active default gateway on your nic. If the default gateway is the 192.168 set of addresses and your 10.x subnet is in a different VLAN then it’s not going to get there because your server won’t have a default gateway pointing at that subnet.

1 Spice up
7

I was thinking what Patrick was thinking.

What network setup do you have?

8

I’ve just set an additional IP in the IPv4 properties, no VLAN. Another thing I’ve noticed is that when setting the DNS address, clients arn’t able to see that DNS is running on the 10.x.x.x address of the server.

9

The reason the 10. and 192. ranges cannot see each other (and explains you DNS observation) is they are in different network segment’s, this is dictated by the Subnet mask, and you need a router to pass traffic from one address range to the other

10

When you installed DHCP role initially on the server , you must have selected the IP (all existing ips appear there) . if you add additional IP later on in NIC setting, that won’t work .
You can check in adsiedit if that ip is showing as authorized,ideally it will not be reflecting there. Path will be adsiedit.msc > connect to configuration schema>CN-services> CN= netservices. On right pane you can see all authorized IP.

Also why can’t you deploy a new server ,plan for cutover -Update Ip helper list at switch end,it should not be much complex.

11

DNS issue solved, in the DNS settings it was set to only accept responses from the one IP, just needed to check the other. DHCP server has been authorised using new IP.

12

I guess you mean to say in DHCP binding option ,not in NIC

13

please check for the bindings of the DHCP server. are both the ip’s listed and activated?external-content.duckduckgo.png

14

@c-j-r ​ Just checked this and it only has the 192.x.x.x address listed. I’m guessing in your screenshot you have multiple NICs? Compared to my 1 NIC with 2x IP’s, I’ll test setting the 10.x.x.x address as the primary.

15

Ah, my bad. The binding is to a NIC, not to an IP.
You have one NIC that has two IP’s. I don’t know if that is the problem, but I think it can be.

Are you able to test with 2 NIC’s. One for 192.x.x.x. and one for 10.x.x.x ?

16

That was going to be my next suggestion.
I’m interested in the outcome of this.

17

Haven’t managed to implement this in the live environment yet, but just been doing some testing using a server and worksation in Hyper-V. I recreated the issue and found that the DHCP server needs to have that 10.x.x.x as it’s primary IP to be able to lease out address of that range in the same subnet. DHCP isn’t my strongest area, I tend to set it up and let it run, never had to change it in anyway. Struggled to find any documentation on the necessary steps but found some helpful information regarding working with multiple DHCP scopes and how the server determines which scope to get an address for the client from is based on the subnet of the network card it’s communicating to the client with, and that the scope and the servers network card IP’s need to be in the same subnet.

1 Spice up
18

That is one of the ways DHCP servers determine this yes. You can also have a setup where you have multiple scopes defined without actually needing any IP addresses assigned from those subnets on the DHCP server itself. This often happens in situations where you have DHCP relay/helpers setup. The DHCP server then determines it by looking at the IP address of the relay/helper (ie your router or switch – whichever device is doing the relaying). If the helper’s IP address matches a scope’s subnet, it will use that.

You’ll want to ensure this scope hands out a proper default gateway so once those clients receive the IP address, they can contact a router. A DHCP server does not necessarily need to reside on the same subnet as the client.