Inconsistent DHCP ACK Errors 1

[F7401]

First of all, let me give you some info on our network.. It is a /16 and has zero VLANs.

We have two primary DHCP servers that are using Server 2012's DHCP Failover in a load balancing configuration. Both servers have option 176 and 242 setup and are limited to give IPs from 10.0.1.1 - 10.0.7.255. We have two IPO 500v2s, one for Digital Phones and one for VoIP phones. The IPO that handles VoIP phones also acts as a DHCP server for Avaya Phones only and uses the range 10.0.10.25 - 10.0.10.175.

What happens is we are having very random inconsistent DHCP ACK errors for some phones. Let me describe a couple scenarios..

1. I'll plug in a 9608G into lets say port 25 of our primary Cisco 3650 PoE switch and it'll go through the boot process and produce an DCHP ACK error and reboot (after being on the main call screen and showing a valid IP from the IPO unit). I'll then plug a 9608 (non gigabit) phone into the same port and it'll boot without any issues. Both of these phones are on the same FW revision. If I take a 9621G over to that same port it'll boot without issues. However, if I take the 9608G over the port the 9621G has been working and still works on it'll throw the ACK error.

2. There's a 9608 that's plugged into the same switch on port 26 that's been working fine for the past month or two and one day it decides to start looping with the DHCP ACK error. Take that phone and plug it into port 27 and it'll work fine. Plug a 5610 into the original port 26 and it'll boot fine (we haven't had a single 5610 throw an ACK error)

This Cisco switch has all ports configured as:
switchport mode access
switchport access vlan 1 (Default)
switchport nonegotiate

Any Ideas?

 

[janni78]

I assume you have separate VLANs for DATA and phones and that option 176,242 on the W2012 DHCP server just tell the phones to switch VLAN so they get DHCP from the IPO?

"Trying is the first step to failure..." - Homer

 

[F7401]

We do not. It is a completely flat network. The scope option on the 2012 DHCP server (Mcipadd=10.0.10.19, mcport=1719, httpsrvr=10.0.10.3, l2qvlan=0) is just to tell the phone where the IPO unit is (10.0.10.19) as well as where the boot files are (10.0.10.3)

 

[janni78]

You can't have 2 different DHCP servers on the same subnet, then it will just be random who answers first.
Either just use the W2012 for DHCP or put the phones on a different VLAN.

"Trying is the first step to failure..." - Homer

 

[F7401]

Even if the IPOs DHCP is set to apply to only Avaya Phones? Why does it work for 90+% of our phones currently without any issues?

 

[janni78]

The IPO DHCP will only answer to Avaya endpoints, but the W2012 DHCP servers will answer to all clients on the network.

"Trying is the first step to failure..." - Homer

 

[mattKnight]

>Why does it work for 90+% of our phones currently without any issues

Possibly, due to the network conditions at the time or load on the IP office or a DHCP server, 1 DHCP server is or isn't able to reply fast enough...

A flat /21 network is pretty large... I'd really consider a new vlan for the handsets, but I get the feeling you don't want to hear that...

Take Care

Matt
I have always wished that my computer would be as easy to use as my telephone.
My wish has come true. I no longer know how to use my telephone.

 

[F7401]

Nah, I do want to hear that. We're actively working towards VLANing our network. We're trying to do it in phases and keep running into hurdles along the way.

Moving from a flat /16 to a bunch of /24s or /22s is quite a change.

 

[janni78]

If you wanna continue to use a flat network disable the DHCP on the IPO and let the W2012 DHCPs handle all the requests and your problems should go away.

"Trying is the first step to failure..." - Homer

 

[F7401]

Gotcha, thank y'all for the help!