Hi everyone,

I’m having serious DHCP issues with my TP-Link Deco X55 PoE mesh setup and I’m hoping someone has seen this before.

Setup

• TP-Link Deco X55 PoE Mesh
• Deco is the only router/DHCP server
• around 40 active devices
• lots of IoT / smart home devices
• Sonos present
• Blink cameras, HomeKit, ESP devices, etc.
• IPv6 already disabled
•  

Problem

Devices randomly stop receiving DHCP leases even though the DHCP pool has plenty of free addresses.

Symptoms:

• clients stuck on “obtaining IP address”
• some devices reconnect only after rebooting
• others never receive an address at all

The issue affects random devices and happens intermittently.

Syslog findings

The logs are flooded with messages like:

daemon.err udhcpd: ip conflict, find a new ip
daemon.err udhcpd: checked a ip conflict node

This happens for many different IP addresses in sequence.

I also found this repeatedly:

imsnoop: mcManagerUpdateIFEntry: if node table is full

So it looks like the multicast/IGMP table may also be filling up.

Already tested

• rebooted all Deco units
• disabled IPv6
• confirmed there is no second DHCP server
• DHCP range is large enough
• no obvious static IP conflicts

My current theory

It feels like:

• the ARP/DHCP tables are filling up
• or the Deco firmware starts detecting phantom IP conflicts

As a result, the DHCP server refuses to hand out leases.

The multicast/mDNS/IGMP traffic from Sonos/HomeKit/IoT devices may also be overwhelming the Deco internally.

Questions

1. Has anyone seen this behavior on Deco X55 or other X-series models?
2. Are there known ARP/IGMP table limits on these devices?
3. Is there anywhere a clear ARP table option?
4. Would any of these help?
   • disabling IGMP snooping
   • reducing the DHCP pool size
   • moving static IPs outside the DHCP range
5. Is this a known firmware issue? (tested also 1.70 early access without success)

Thanks in advance 🙂