Critical Infrastructure Failure - Deco BE65/BE25 Mesh & Omada Managed Switch
To TP-Link Global Support / Engineering Team,
I am writing this email to express my extreme disappointment and frustration with the lack of support I have received. Despite numerous emails regarding the stability of my network after installing the provided beta firmware, I have received zero response.
This is my final attempt to resolve this through your support channels. If I do not receive an immediate technical intervention, I will be returning all products (Deco units and Omada switch) to the seller and replacing my entire infrastructure with a competing manufacturer's system.
Network Hardware Profile:
-
Main Router: 1 x Deco BE65 (Running Beta Firmware received on Dec 1st).
-
Satellites: 2 x Deco BE65 + 1 x Deco BE25 (Configuration: Ethernet Backhaul).
-
Switch: TP-Link Omada SG3218XP-M2 (2.5GbE Managed Switch).
-
Cabling: Certified Cat.6 copper.
The "16-Day Stability Leak" Observation: The beta firmware (MLO fix) provided on December 1st appeared stable for exactly 16 days. On December 17th, the system suffered a total collapse. This indicates a severe Memory Leak or MAC Table Corruption that accumulates over time until the "Network Resource Daemon" (nrd) fails.
Current Critical Failures:
-
Broadcast Storms: System logs show hundreds of DHCP DISCOVER/OFFER cycles per second, rendering the network unusable.
-
Hardware Hangs: Satellites frequently go "Offline" (Solid Red LED).
-
Total IoT Collapse: All 2.4GHz devices have lost connectivity.
-
Wired Latency: Severe packet loss (Request timed out) even on wired connections, affecting CCTV security cameras.
Technical Diagnostics (Based on Attached Logs v3, v4):
-
MAC Flapping Loop: The Deco system is incorrectly identifying the SG3218XP-M2 switch (MAC: E0:D3:62:0A:72:4B) as a roaming wireless station. This creates an internal loop between
ethandathinterfaces, which standard STP (currently set to Priority 32768 on the switch) cannot stop. -
Driver Errors: Constant
daemon.err nrd: send ioctl failedentries prove that the Deco OS is losing communication with its own network drivers.
Troubleshooting Already Conducted (with no success):
-
Assigned Static IP (10.5.236.250) to the switch outside the DHCP range.
-
Enabled/Disabled IGMP Snooping and Querier.
-
Verified physical layer (replaced SFP with RJ45).
-
Toggled Beamforming and Fast Roaming.
Expectations: I require an immediate response from a senior network engineer. I need a stable firmware that correctly handles Ethernet Backhaul on 2.5GbE managed switches without creating wireless loops for infrastructure MAC addresses.
As stated, if this is not addressed immediately, I will proceed with a full return of all TP-Link hardware and switch to another brand.
I am awaiting your urgent reply here or through the mail.
Best regards, Łukasz
