Clients are randomly losing DHCP assigned IP address

[matkins_PL]

Hi,

I have a very odd issue on my network, clients (both wired and wireless, various NICs used) are losing IP addresses randomly (not very often, once every few hours I would say).

They have no problem obtaining an IP address at any given time, after they lose one (it becomes 169.x.x.x) all it takes is to click troubleshoot/repair connection and it will work again. (or disconnect/connect from the network, or ipconfig /renew etc.)

The environment is built with Aruba 5406R zl2 with 6x 8x10GB modules as core switch (which has 2x ip helpers conigured on both wired and wireless vlans), and Aruba 2930F Edge switches. DHCP servers are 2x Windows Server 2016 configured with DHCP failover.

Anyone seen anything like this before? Any idea how to check what is happening with Windows machine when it is losing IP address (I have checked Event Logs but nothing is there, maybe some extra auditing policies needs to be enabled?)

Many Thanks,

Mateusz Kordaszewski 

[MstWntd]

Hi @matkins_PL,

You may try the following to check is that issue is being caused by the switch here.

01. When a user reports that they are losing the IP address, make a note of the timings.

02. Identify the data path of that user to the DHCP Server.
Example : A user may be connected to Port 2 on 2930F for VLAN 3 (DATA) and uplink port 24 from that switch may be connected to Aruba 5406R switch. And DHCP Server (Failover) may be connected to port 20 on both the 2930F.

03. Check the switch (2930F and Aruba 5406R) logs, when this issue arises. Is there any logs passing through the switches? Command : "Show log -r"

04. Check for the MAC-Address of that Client on that particular VLAN. Command : show mac-address vlan 1 or respective VLAN where the PC communicates.

05. If you are good at analysing *.pcap, you may connect a sniffer at 2930F or Aruba 5406Rs switch and monitor the Inbound and Outbound traffic of the uplinks to the DHCP Server and check the network traces as well.

The above steps are something that I may try from the switch side. Maybe some of the above steps can give us some hints to identify the issue here.

If the above steps alleviates your query, please give kudos!

Regards,

[matkins_PL · Accepted solution]

Hi,

Already resolved. The Issue was Windows Bug affecting failover DHCP service...

https://social.technet.microsoft.com/Forums/en-US/eaf32999-7e6d-44c3-a601-ba6372e3f4b9/kb4284833-causing-dhcp-clients-to-lose-thier-addresses?forum=winserveripamdhcpdns

Cheers,

[brconflict]

This is still an open issue, and even with not using Microsoft AD as DHCP servers. The symptom is the same where the client seems to receive a faulty DHCP response, and moves to 169.254.x.x IP. The latest patches for Windows Surface 2 laptops proves to not address the issue as of October 12, 2019.