I have all the workstations at one site unable to pull an IP address from DHCP suddenly this morning and no idea why. This is a small site running a single ICX 7150-48-POE unit with Layer-3 code for 08.0.90e.
The users are only getting 169.254.x.x APIPA IPs, which is strange because all of my other 90+ sits are doing perfectly fine.
We have NOT made any configuration changes.
Here is the VRI:
interface ve 1 port-name Default Gateway ip address 10.81.0.1 255.255.0.0 ip helper-address 1 10.1.2.3 ip helper-address 2 10.4.5.6 !
If I do a debug of the ip dhcp relay, this is what I get. Any ideas?
Debug: Mar 2 10:56:33 DHCP_RELAY : Relay agent rxed message on port v1 Debug: Mar 2 10:56:33 DHCP_RELAY : Received Bootp request msg on port v1 Debug: Mar 2 10:56:33 DHCP_RELAY : Bootp request packet send with GiAddr 10.81.0.1 Debug: Mar 2 10:56:33 DHCP_RELAY : Bootp request packet sent on port INVALID,With helper-address 10.1.2.3 Debug: Mar 2 10:56:33 DHCP_RELAY : Bootp request packet send with GiAddr 10.81.0.1 Debug: Mar 2 10:56:33 DHCP_RELAY : Bootp request packet sent on port INVALID,With helper-address 10.4.5.6 Debug: Mar 2 10:56:33 DHCP_RELAY :restored gi_l2_action_required: 0
Yes, DHCP is showing operational for those subnets, I can reach the LAN network, I can ping the DHCP servers from the switch, have rebooted the switch, etc. For some reason everything works but DHCP and just for that one site.
Sometimes the best bet is to just pcap along the path see where it is breaking in the DORA (discover, offer, request, ack) process. A quick check would be to just pcap on your dhcp server and see if discovers are even making it there. If you don't see discovers at the dhcp server, then I would check egress on ICX to see if we are forwarding the discover in the first place. There are some additional things we can check on the ICX side via dm raw, etc but I'd recommend opening a case if you need to go that route.
Ben Beck, RCNA, RCNI, Principal Technical Support Engineer support.ruckuswireless.com/contact-us
I was communicating with our end-user-support group (desktop technicians) , and it is suddenly clear to me the scope and severity of this issue was overstated because I have power-cycled (err warm-reloaded) the device and that would clearly cycle ALL interfaces.
That said, I do a #sh arp on it and can clearly see a LOT of DHCP assigned IP addresses.
My supposition therefore goes that it is working and there may be a problem with a select few workstations OR perhaps they have a small dumb-switch or something that is not authorized to be on the network. None the less, I see both data and voice DHCP provided IPs.