Feature Request: Re scan and change automaticaly to better RSSI is available every few minutes

[Ricardo-Miguel-Caldeira]

With the implentation in v16.0.1 of
-- WiFi access point selection fix so now WLED connects to the strongest signal not simply the first detected (by @DedeHai

Can we please add an automatic retry/check every other few minutes to connect to the highest RSSI available again? If for any reason wled looses it WIFI connectivity it stays in limbo and ends up always needing a reboot...

And thank you all for your great work and dedication to this project!

[netmindz]

Have you verified that it isn't doing this already?

[DedeHai]

wifi is handled by the core, the "full scan" is set. if it is not respected, that is an arduino bug (which would not surprise me ;) )

[willmmiles]

I'm pretty sure there is a legitimate issue here - I think I ran in to this one myself with units dual-configured for my home test SSID and my phone's hotspot I use on the road. They'd latch on the strongest one they found after bootup; if it went offline, they'd never try to connect to the other one.

I sketched a lame patch, but never PR'd - I'd intended to rework it along with the rest of the network management subsystem (cf. #5650, also src/dependencies/network cleanup).

[DedeHai]

@willmmiles so arduino does not auto-rescan and we need to enforce it "manually"?

[willmmiles]

It will reconnect to the original SSID should it come back online. (I don't know offhand if it scans channels - I'm pretty sure it does.) When disconnected, the Arduino core definitely does not auto scan for other SSIDs, though. Our code was supposed to restart the scan process but it didn't seem to work reliably.

Basically what I saw was:

• Configure "MyCoolWifi" and "EvenCoolerWifi" in the config page
• WLED scans, finds "MyCoolWifi" is stronger at boot, connects
• "MyCoolWifi" goes down
• WLED hangs forever looking for "MyCoolWifi", even though "EvenCoolerWifi" is available

There were also a big pile of possible race conditions in dispatching scan events from both the event handler (runs on the wifi task) and the loop() task. The patch moved everything to loop() to ensure it never raced with itself.

[netmindz]

I think we are confusing two different things here. I would expect that if we give multiple SSID it would connect to the strongest, then change between different base stations on this SSID, but I would not expect it to jump between totally different SSIDs - which typically are different networks, just because one has a stronger signal

(Yes i am aware there are shitty companies that produce wifi extenders where "_ETX" is actually the same network, just under a different SSID so no technical users can see which they are connected to)