10-01-2007 11:44 PM - edited 03-05-2019 06:48 PM
Hi, everyone.
I have a dhcp-related problem in my cable network. When all cable modems are being reset at once and try to get an IP, things get stuck. I see DHCPDISCOVER message from cable modems and DHCPOFFER sent back to them. Now a cable modem should send DHCPREQUEST, but instead it hangs out there quiet and in a while repeats its DHCPDISCOVER message. A Cisco CMTS is a dhcp relay.
When I reset cable modems one by one all goes smooth (DISCOVER, OFFER, REQUEST and ACK).
So I'm thinking that the bottleneck is a cisco router.
If anyone has experienced anything like that, please let me know. Thanks in advance.
10-08-2007 06:13 AM
You can check if the MAC address of the client Client Identifier (CID) appears in a DHCPDISCOVER packet or not. Use a sniffer and the traceroute command to check indirectly the bottleneck. The DHCP request dropped by the switch is generated by the PC less than 2 ms after the DHCP ack.
10-09-2007 06:28 AM
The problem occurs in a cable network. There is no switch. DHCPDISCOVER from a cable modem is passed through a relay agent (CMTS) to a provisioning server (with dhcpd running). It looks like CMTS is passing the messages to dhcpd all right, but is having a problem to pass OFFERs and ACKs from dhcpd back to cable modems.
I've found a problem description exact to my on the internet:
http://www.gossamer-threads.com/lists/cisco/nsp/71181
However they have not found a solution either :-(
10-09-2007 06:30 AM
Forgot to add... The problem is exactly like in the link above. Cable modems have EMTAs built in that have to get their own IPs. It is those EMTA IPs that are being problematic. Cable modems get their IPs all right.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide