EAP655-In Wall 2.4GHz Radio Instability / Driver Hang Under IoT Load
Hi Everyone
I had a previous post about this, tried everything I could but I am still with the same problem. In short I have a reproducible 2.4 GHz radio instability affecting 2 EAP655-In Wall access points.
Environment
-
Devices: 2 × EAP655-In Wall
-
Firmware: 1.4.3 Build 20250714 Rel. 63955
-
Controller: Omada 6.0.0.25 (installed on LXC)
-
Backhaul: Wired Ethernet (no mesh)
-
Security: WPA2/WPA3 mixed
-
SSID: Single SSID (2.4 GHz + 5 GHz)
-
5 GHz operation is fully stable
Problem
The 2.4 GHz radio intermittently stalls, causing most or all 2.4 GHz clients (3 cameras, Thermostat, smart switches) to disconnect simultaneously.
The AP remains reachable via management IP, and 5 GHz clients continue to function normally.
To recover I do this:
-
Manual channel change on 2.4 GHz (change channel or anything that re-instantiate the connection), or
-
AP reboot
This occurs regardless of channel, power level, or radio mode.
What I tried so far:
-
Not PoE, power, or uplink related (AP stays connected, 5 GHz unaffected)
-
Not interference-based (I have Zigbee which I moved to channel 25)
-
Not roaming-related (802.11r disabled, roaming features removed)
-
Not AX-specific (802.11ax disabled on 2.4 GHz)
Logs:
I captured the AP support logs immediately after failures and show the following:
-
Repeated ieee80211_ioctl_siwscan: preempt_scan loops
-
wifiX interrupts didn’t increase watchdog messages
-
Forced station eviction (ieee80211_ioctl_kickmac)
-
Driver errors indicating peer/NSS context loss:
-
Unable to find peer object
-
peer NULL - Null nss context
-
Are you aware whether this is a known issue with the EAP655-In Wall 2.4 GHz radio/driver?
If needed I can provide additional logs, controller backups, etc
Thank you for your help and support.
