Posted: Sat Apr 20, 2019 3:08 Post subject: hostapd flooding logs with "signal too weak for auth&qu
I have three routers in my home with two set up as Access Points and one 'master' that does all the DHCP. To improve roaming, I've recently configured the "Drop Clients with Low Signal" settings under the Wireless Advanced Settings. This mostly works, however, sometimes when I have a device (usually one of our cell phones) that attempts to connect to a router that is too far away, it'll get rejected:
04-19-2019 20:03:37 Daemon.Info 192.168.88.1 Apr 20 03:03:35 hostapd: ath0: STA c8:12:34:56:78:91 MLME: assoc request, signal -68 (DENIED) (signal too weak for authentication)
(I've modified the MAC here.)
That's correct behavior. However, sometimes hostapd will FLOOD the log with these messages over and and over continuously. The timestamp on the message increments, so they are fresh messages, not the same one over and over. Once this happens, that router gets bogged down and no one can connect to its wifi, even if in close range.
I'm running recent builds. The one here is DD-WRT v3.0-r39538 std (04/16/19) running on a TP-Link TL-WDR4300 v1. This also happens with a 4/10 build on a separate router, a TP-Link TL-WR841ND v8.
Any ideas of how to prevent this from happening? Or is it a bug that we might see a fix for?
Thanks for the tip on the logging configuration. I never knew that was there (never thought to check under Security).
However, this doesn't address the underlying issue as to why it floods these messages and hangs in certain circumstances when it rejects a weak connection. Is this a known issue and do you know that, if I do disable these logs, that it will not still be spinning the CPU on whatever operation it's actually logging about (i.e., that it's _just_ the logging that is bogging down and it would actually continue normally [and not get stuck on this rejection] without logging)? The problem is, if I turn off the logging, then I can longer see when rejections are occurring, which makes it difficult to tell whether it is actually resolved or just happens to not be happening. It also means I lose the benefit of having the logs to debug other issues or understand behavior.
Under normal circumstances, I can see weak signal authentication requests getting rejected and things work as normal. But there are certain cases where a roaming device triggers this problem scenario where it floods with these messages. It's not a normal rejection. It really seems like it's a bug that it gets stuck in some sort of loop, as it'll repeat that message continuously until I kill hostapd (or reboot the router).
This doesn't sound like a bug to you?
I'm not sure, but I'm wondering if it has to do with an authentication request while the device is moving, such that it is strong enough at one point, but then it drifts to be too weak. That's pure speculation, but I'm thinking some situation like that to put it into the bad state.