dhcp relay issue

Started by ggnfs000, February 28, 2017, 10:38:48 PM

Previous topic - Next topic

ggnfs000

i have set two routers with left one acting as DHCP relay and right one acting as DHCP server and got it working. However if I instead replace the right router's DHCP server with normal DHCP (not on router) behind that router, then DHCP request from DHCP relay is failing (I made sure ip helper-address on relay is adjusted accordingly). I can see the request is reaching the DHCP server by placing sniffer (boy this is a nice device: wireshark like thingy with two ethernet jacks and setup in seconds) and DHCP request from relay 172.16.1.1 (dhcp relay IP subnet from which DHCP request is coming from) to 172.16.2.11 (dhcp server).

Now the moral of the story is is DHCP server supposed to server DHCP relay request or is it just Cisco router specific implementation.

I know conventional router respond to 0.0.0.0 to 255.255.255.255 request but not sure to <relay IP> to <dhcp server unicast IP> type of request.

SimonV

Can you see the DHCP server responding on your sniffer?

wintermute000

DHCP relay is 100% standard operating practice to 'real' servers like active directory.

deanwebb

If the IP helper addresses are set up on both routers, then the fault lies with the DHCP server. Is it set up and configured correctly?
Take a baseball bat and trash all the routers, shout out "IT'S A NETWORK PROBLEM NOW, SUCKERS!" and then peel out of the parking lot in your Ferrari.
"The world could perish if people only worked on things that were easy to handle." -- Vladimir Savchenko
Вопросы есть? Вопросов нет! | BCEB: Belkin Certified Expert Baffler | "Plan B is Plan A with an element of panic." -- John Clarke
Accounting is architecture, remember that!
Air gaps are high-latency Internet connections.

Otanx

On the DHCP server do you have the scope defined correctly on the new DHCP server? That is where I would start looking. If the DHCP server does not have a scope defined for the request it will ignore it.

-Otanx

ggnfs000

Quote from: Otanx on March 01, 2017, 01:28:30 PM
On the DHCP server do you have the scope defined correctly on the new DHCP server? That is where I would start looking. If the DHCP server does not have a scope defined for the request it will ignore it.

-Otanx

i will look into that next few days. I was just about to post another ipv6 question but did not wanna flood this section of forum with basic questions (last two posts were mine). Well peeked at the book and found the answer.

icecream-guy

Quote from: ggnfs000 on March 20, 2017, 03:57:38 AM
Quote from: Otanx on March 01, 2017, 01:28:30 PM
On the DHCP server do you have the scope defined correctly on the new DHCP server? That is where I would start looking. If the DHCP server does not have a scope defined for the request it will ignore it.

-Otanx

i will look into that next few days. I was just about to post another ipv6 question but did not wanna flood this section of forum with basic questions (last two posts were mine). Well peeked at the book and found the answer.

Post em, it's not like were even close to our posting limit and overwhelming the servers,  there's usually only a handful or two of new messages per day..
glad you found the answers. but don't shy for posting. remember the only stupid question is the one you don't ask.
:professorcat:

My Moral Fibers have been cut.

ggnfs000

Quite an old topic but since long, I found the problem was I havent added the subnet scope that corresponds to that relay interface facing the client. Once i added the second subnet DHCP through relay worked.

server: 172.16.2.12 (works)
relay: 172.17.1.x (subnet facing relay client), ip helper: 172.16.2.12
added subnet scope: 172.17.1.x <- this worked.

Now I took it further today and yesterday, this prior work was done completely on packet tracer.

I setup real WinServer 2016 PXE server long with DHCP and used CS1000v as router acting as relay, with exactly same topology as in the packet tracer example (save for IP address difference).
Now DHCP in this configuration still works through relay, by looking at the DHCP pool.

However when I do, PXE server will not respond. PXE server will respond and install hosted image through direct PXE boot (not relay)
Through wireshark, when client does PXE through relay, I can see DHCP still assigns through 4-step protocol: 1. discover 2. offer 3. request 4. acknowledge between: relay and DHCP/pxe server then nothing happens. This makes me wondering if anything special needs to be done in terms of Windows Deployment Service.

ggnfs000

Lol further fune-tuning:  :o
The prior post regarding PXE boot does not work in UEFI BIOS mode. Switched the client to old legacy PXE boot mode and it worked. Now this means it further narrowed to UEFI boot only. To summarize:
DHCP relay DHCP: OK
DHCP PXE boot through relay (UEFI BIOS): not OK
DHCP PXE boot through relay (legacy BIOS): OK.

Isn't that pretty.

ggnfs000

So did the same thing with IPv6 PXE server. WDS failed miserably when the client boots through DHCP relay with IPv6.
Another option was to create multi-homed server for both ipv4,ipv6. It was erratic at best, it never worked reliably.

Also on linux red hat, I did some something similar: ipv4 with relay and multi-homed DHCP PXE server with 2 subnets. It failed. Both Windows and Linux was beyond terrible in that respect. At least i was sane enough not to follow the rabbit hole and repeat the similar fate that of this guy:

https://www.sudosatirical.com/articles/man-loses-will-to-live-during-gentoo-install/

deanwebb

Take a baseball bat and trash all the routers, shout out "IT'S A NETWORK PROBLEM NOW, SUCKERS!" and then peel out of the parking lot in your Ferrari.
"The world could perish if people only worked on things that were easy to handle." -- Vladimir Savchenko
Вопросы есть? Вопросов нет! | BCEB: Belkin Certified Expert Baffler | "Plan B is Plan A with an element of panic." -- John Clarke
Accounting is architecture, remember that!
Air gaps are high-latency Internet connections.

ggnfs000

it is hilariously tragic.