What's the best approach towards determining if I have a rogue DHCP server inside my network?
I'm wondering how most admins approach these kinds of problems. I found DHCP Probe through searching, and thought about trying it out. Has anyone had experience with it? (I would like to know before taking the time to compile it and install).
Do you know any useful tools or best-practices towards finding rogue DHCP servers?
One simple method is to simply run a sniffer like tcpdump/wireshark on a computer and send out a DHCP request. If you see any offers other then from your real DHCP server then you know you have a problem.
I suggest starting two terminals, one for monitoring and another for sending a request. Terminal1 will show responses from all existing DHCP servers including MAC address. This example was run on Ubuntu:
Terminal1 (for monitoring):
sudo tcpdump -nelt udp port 68 | grep -i "boot.*reply"
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on enp2s0, link-type EN10MB (Ethernet), capture size 262144 bytes
20:a6:80:f9:12:2f > ff:ff:ff:ff:ff:ff, ethertype IPv4 (0x0800), length 332: 192.168.1.1.67 > 255.255.255.255.68: BOOTP/DHCP, Reply, length 290
00:23:cd:c3:83:8a > ff:ff:ff:ff:ff:ff, ethertype IPv4 (0x0800), length 590: 192.168.1.253.67 > 255.255.255.255.68: BOOTP/DHCP, Reply, length 548
Terminal2 (for sending a request):
sudo nmap --script broadcast-dhcp-discover -e eth0
Starting Nmap 7.01 ( https://nmap.org ) at 2019-10-13 21:21 EEST
Pre-scan script results:
| broadcast-dhcp-discover:
| Response 1 of 1:
| IP Offered: 192.168.1.228
| DHCP Message Type: DHCPOFFER
| IP Address Lease Time: 2h00m00s
| Server Identifier: 192.168.1.1
| Subnet Mask: 255.255.255.0
| Router: 192.168.1.1
|_ Domain Name Server: 8.8.8.8, 8.8.4.4
WARNING: No targets were specified, so 0 hosts scanned.
Nmap done: 0 IP addresses (0 hosts up) scanned in 0.94 seconds
That monitoring terminal is needed just to see all responses (nmap is able to show only the first response).
To recap and add to some of the other answers:
Temporarily disable your production DHCP server and see if other servers respond.
You can get the IP address of the server by running ipconfig /all on a windows machine, and then you can get the MAC address by looking for that IP address using arp -a.
On a Mac, run ipconfig getpacket en0 (or en1). See http://www.macosxhints.com/article.php?story=20060124152826491.
The DHCP server information is usually in /var/log/messages. sudo grep -i dhcp /var/log/messages*
Disabling your production DHCP server might not be a good option, of course.
Use a tool that specifically looks for rogue DHCP servers
See http://en.wikipedia.org/wiki/Rogue_DHCP for a list of tools (many of which were listed in other responses).
Configure switches to block DHCP offers
Most managed switches can be configured to prevent rogue DHCP servers:
The Wireshark / DHCP explorer / DHCP Probe approaches are good for a one time or periodic check. However, I'd recommend looking into DHCP Snooping support on your network. This feature will provide constant protection from rogue DHCP servers on the network, and is supported by many different hardware vendors.
Here's the feature set as indicated in the Cisco docs.
• Validates DHCP messages received from untrusted sources and filters out invalid messages.
• Rate-limits DHCP traffic from trusted and untrusted sources.
• Builds and maintains the DHCP snooping binding database, which contains information about untrusted hosts with leased IP addresses.
• Utilizes the DHCP snooping binding database to validate subsequent requests from untrusted hosts.
dhcploc.exe is the quickest and handiest way on Windows systems. It is available in the XP Support Tools. The Support Tools are on every OEM/retail XP disk, but may or may not be on "recovery disks" provided by some OEMs. You can also download them from MS.
It's a simple commandline tool. You run dhcploc {yourIPaddress} and then press the 'd' key to do a fake discovery. If you leave it running without pressing any keys, it will display every DHCP request and answer it hears. Press 'q' to quit.
Scapy is a python based packet crafting tool that is good for these sort tasks. There is an example of how to do exactly this here.
To expand on l0c0b0x's comment about using bootp.type == 2 as a filter. The bootp.type filter is only available in Wireshark/tshark. It is not available in tcpdump which the contextual location of his comment inclined me to believe.
Tshark works perfectly for this.
We have our network divided up into numerous broadcast domains, each with their own Linux-based probe with a point of presence on the "local" broadcast domain and on an administrative subnet in one fashion or another. Tshark combined with ClusterSSH allows me to easily look for DHCP traffic or (anything else for that matter) on the further flung corners of the network.
This will find DHCP replies using Linux:
# ifconfig ethX promisc
# tshark -i ethX -n port 68 -R 'bootp.type == 2'
once you've established that there's a rogue dhcp server on the network I found the quickest way to resolve it was...
Send an email round to the whole company saying:
"which one of you has added a wireless router into the LAN, you've killed the internet for everyone else"
expect a sheepish response, or the conflicting device to disappear, quickly :)
You can use RogueChecker from Microsoft:
original link: http://blogs.technet.com/b/teamdhcp/archive/2009/07/03/rogue-dhcp-server-detection.aspx
link with download: https://web.archive.org/web/20141022013752/http://blogs.technet.com/b/teamdhcp/archive/2009/07/03/rogue-dhcp-server-detection.aspx
also read this information:
Disable the main DHCP server and (re)configure a connection.
If you get an IP address, you've got a rogue.
If you have a Linux handy, the standard dhcpclient tells you the IP address of the DHCP server (else you can sniff the traffic to see where the DHCP response came from).
There are several ways, if your running a small network the simplest way is to turn off / disable / un-plug your dhcp server and then run ipconfig /renew or similar on a client and if you obtain and IP you have something rougue on your network.
Another way would be to use Wireshark packet capturer/analyser to look at your network traffic and find DHCP connections, there is a lab worksheet on how do do this avaliable from here.
There are also a number of utilies avaliable which proport to do this one is DHCP explorer another is DHCP probe which you mentioned in your original post.
on debian/ubuntu one also has the options to use dhcpdump and/or tcpdump with the help of e.g. dhclient
Use dhcpdump:
dhcpdump -i eth0 in one shell/shell (eth0 or the name of your interface)dhclient in another shell (it doesn't have to run successfully)dhcpdump for information (it should be a nice formatted, informative list of the most details)Option 2 if you dont like to use dhcpdump:
tcpdump -i eth0 -t -n > /tmp/my_file.txt in one shell/window -t = disable timestamp // -n = disable name-resolution, just IP-address, no servernames (for RHEL/centos use -nn) )dhclient in another shell (it doesn't have to run successfully)*sidenote: tcpdump and dhcpdump probably have to be installed (e.g.: sudo apt get install tcpdump dhcpdump); dhcpdump depends on tcpdump
You could do a ping sweep of your networks and then compare that to the number of DHCP leases handed out by your DHCP server.
You need to have a general idea of the number of static devices (router interfaces and printers perhaps) which will skew this number slightly, but this should be a quick and accurate way of identifying them across multiple networks.
Under OSX (and I suspect Linux), this is what I did:
tcpdump -i en0 -n port 67 or port 68
You will need to change the en0 to be the name of your interface. Once I made a DHCP request (by clicking Renew DHCP Lease in the advanced section of the Network System Preferences panel), I saw something like the follwing:
15:59:17.843 IP 0.0.0.0.68 > 255.255.255.255.67: BOOTP/DHCP, Request from 6:b:5:9:f:3, \
length 300
15:59:17.847 IP 10.1.2.1.67 > 10.1.2.155.68: BOOTP/DHCP, Reply, length 309
Port 67 is listed in my /etc/services as "bootps" Bootstrap Protocol Server.
Turning off your primary DHCP server may be needed to see any other network devices that are inappropriately responding to DHCP requests.
On linux, after playing with many tools, following is the only thing that worked reliably (On a RHEL8 based system)
In this case:
eth0.eth0-static is my static IP configuration, active by defaulteth0-dhcp is the dynamic IP configuration#!/bin/bash
Define the servers already known in your network
known_servers="10.0.0.1|192.168.0.1"
On the background, dump all DHCP calls other then requests, ignore known servers
tcpdump -i eth0 -n port 67 or 68 2> dhcp.err|grep -v Request| grep -v "$known_servers" > dhcp.log &
Disable static IP
nmcli con down eth0-static 2>&1 > /dev/null
sleep 2
Enable the connection that will do dhcp request
nmcli con up eth0-dhcp 2>&1 > /dev/null
sleep 2
Re-enable the static IP, to go back to original IP.
nmcli con up eth0-static 2>&1 > /dev/null
Kill tcpdump
pkill tcpdump
Print results
cat dhcp.log
Say there is another rouge server with IP 10.1.1.61 on your network, output file will look like:
11:51:53.583149 IP 10.1.1.61.bootps > 10.1.1.23 .bootpc: BOOTP/DHCP, Reply, length 317 11:51:53.583271 IP 10.1.1.61.bootps > 10.1.1.23 .bootpc: BOOTP/DHCP, Reply, length 307
Also, do not disable your main DHCP server just to find other ones ;)
