Weird DHCP Issue on MX67

[beta-389-user]

I have a created static route for subnet 20.20.20.0/26 and pointed it to next hop 10.10.10.2 (IP on HP core switch). That IP belongs to Interconnect VLAN 10.10.10.0/29 (VLAN 11) between Meraki MX67 (10.10.10.1) and Core switch (10.10.10.2).

Meraki is acting as DHCP Server for 20.20.20.0/26 subnet, settings are as follows :

Client addressing : Run a DHCP server
Gateway IP : 20.20.20.1 (IP On HP Core Switch) - I can reach this IP
Lease time : 1 day
DNS nameservers : Specify nameservers...
Custom nameservers
1.1.1.1
2.2.2.2
Boot options disabled
No DHCP Option
No Reserved Range
No Fixed IP assignment

Configured Switchport as Trunk and configured ICN VLAN 11  as Native

Now there are two machines in Subnet 20.20.20.0/26 which are not able to receive IP from Meraki. In the event log, I see below error.

DHCP DHCP problem extra: no_offers_received, vap: 0, vlan: 11

Above error only appears for first machine...

Whereas for second machine, it requests IP from whole different range and can be seen under DHCP Option 50 in packet capture. (Will run this with client first thing tomorrow)

For first machine, it does not list any DHCP option 50

So far whenever I run packet capture, I Only see DHCP Discovery packets but nothing for DHCP OFFERS and all.

I have similar setup at other sites where it is working flawlessly but not here. Firmware (18.107.2) is also up to date, same firmware is also present on working sites. Client says switch has same config as other sites, so far CISCO TAC not able to conclude.. Any help ?

[alemabrahao]

You need to create the IP helper address in the SVI of the HP Switch pointing to the IP of the MX.

https://techhub.hpe.com/eginfolib/networking/docs/switches/K-KA-KB/15-18/5998-8164_mrg/content/ch12s...

[alemabrahao]

More specifically for the SVI of this subnet 20.20.20.0/26

[beta-389-user]

IP Helper Address is configured pointing to Meraki IP of Interconnect VLAN 10.10.10.1 but still no luck.

[alemabrahao]

From what I understand, only two machines are not receiving IP, right? Are you sure the addresses haven't already run out?

[alemabrahao]

https://documentation.meraki.com/MR/Monitoring_and_Reporting/DHCP_%22no_offers_received%22_Errors_in...

[beta-389-user]

1. Is your DHCP server online, operational, and configured correctly with an active DHCP scope? Yes, correctly configured.
2. Have you exhausted all of your available IP addresses in your DHCP address pool? Do you need to increase the size of your DHCP address pool? Has 0 utilization.
3. If the DHCP scope for your local LAN is configured on a DHCP server located in a different broadcast domain, do you have IP helper or a DHCP relay agent configured on the local LAN to intercept DHCP discovery broadcasts and forward them to the appropriate DHCP server? Switch VLAN is having IP Helper address pointing to Meraki IP. 
4. If you are using a DHCP reservation, is the IP address assigned to your potential clients MAC address part of your DHCP address pool and not excluded? No DHCP reservations are done.
5. Are DHCP discovers reaching your DHCP server successfully? MX receives DISCOVERY Packet and seems Source is VLAN IP (20.20.20.1) from Switch and Destination IP reaching to MX IP of ICN VLAN.

[alemabrahao]

I'm almost 100% sure that it's a problem with the HP switch's DHCP relay, I've seen this happen in the past with HP switches.

See this link, maybe it will be useful.

https://support.hpe.com/hpesc/public/docDisplay?docId=sf000057137en_us&docLocale=en_US&page=index.ht...

[alemabrahao]

Another question is, are you sure that on both the Switch and the MX the network mask (255.255.255.192) and subnet are configured the same?

[beta-389-user]

My bad if I am not being clear there. There are only 2 machines in that Subnet so pool is completely free. 

[alemabrahao · Accepted solution]

A very quick test is to connect the machine directly to the port on the MX to validate whether the machine will receive IP. I still believe the problem is not the MX.

[beta-389-user]

After replacing HP Core switch, issue got resolved. Thanks for all the suggestions and you were right. 

[kydave]

Out of curiosity, what IP address is/was the MX using for itself in the 20.20.20.0 subnet? Could it have been using 20.20.20.1, the same address as the HP core SVI? Perhaps the HP hardware swap just forced a readdressing that no longer conflicts?

Dave H