WiFi Auto Reconnect Still Not Working - Is there a best practice to guarantee connection?

[vseven]

Board: ModeMCU ESP32 Dev Module
Core Installation/update date: 15/Sep/2017
IDE name: Arduino IDE

It seems there is no auto reconnect logic allowing the ESP32 to reconnect when the connection drops or if there is its not working properly. I have had completely random results with my ESP32 board, sometimes its up for as little as 30 minutes and at most 18 - 20 hours. WiFi is dropping and doesn't want to reconnect.

The library I am using for WiFi is watching the WiFi event and on a disconnect its trying to reconnect which helps but sometimes it fails the reconnect and then that's it...I have to reboot. There is no sleep/low power mode at all involved and the device is constantly powered by 3.3v in. There was a case opened regarding this, #353, but its closed without any real answer (says auto connect is kind of implemented...not sure what that means).

I guess the question comes down to is auto reconnect implemented in the code itself, reliably, and if not what is the best Arduino code way of making sure it is connected?

-Allan

[me-no-dev]

can you please enable debug in the board menu and see why it does not want to reconnect when you tell it to do so?

[vseven]

I can try that tonight. The Arduino code is pretty basic, watching the WiFi event and then has this:

case SYSTEM_EVENT_STA_DISCONNECTED:
	Serial.println("WiFi lost connection.  Attempting to reconnect...");
	WiFi.reconnect();
	break;

But shouldn't there be something that can be called to make this happen on it's own?

[nepeee]

Just for a quick fix:
Start a timer in the SYSTEM_EVENT_STA_DISCONNECTED that calls the WiFi.reconnect(); function and stop it in the SYSTEM_EVENT_STA_CONNECTED. The main problem is with your idea is if the device cant connect than the SYSTEM_EVENT_STA_DISCONNECTED event is not called so its never try to reconnect again.

[vseven]

Unfortunately I'm adapting someone else library and I'm not very familiar with C++ but I'll see what I can do about adding a timer (I know VB.net...I can at least figure out the logical parts and hopefully adapt).

But I think you are right about it not reconnecting or better yet only trying on a disconnect which then if it fails just once it give up. Enabling debug might show me that so I'll start there.

But back to the original question and the previously closed issue: Is this functionality not built-in? And if not why was the previous issue closed as "kind of" working? Also if not will it be?

[vseven]

I switched the debug level to debug and opened serial monitor and got this over and over:

[WiFi-event] event: 5
WiFi lost connection.  Attempting to reconnect...
[WiFi-event] event: 5
WiFi lost connection.  Attempting to reconnect...
[WiFi-event] event: 5
WiFi lost connection.  Attempting to reconnect...

I checked the last time it sent data to another controller and that was over 24 hours ago so I'm assuming its stuck in this mode. So I powered it down and back up, it connected to WiFi just fine and transmitted data how it should for about 2 hours this time. Then without any other message it disconnected and started with the exact same message over and over. I let it go for a hour, rebooted the controller, and its connected again but I'm sure its just a matter of time before it starts over.

So the first question is how do I fix it. Second this still leads me to my previous questions about auto reconnect

[me-no-dev]

Guys please enable debug so you can see more verbose WiFi state output:

[everslick]

I use the following code to check if my device is still connected to the STA (call this from loop()). You get the idea? wifi_is_connected is a global that gets set and reset in the wifi event callback.

static void poll_watchdog_sta(void) {
  static uint32_t ms = millis();
  static uint8_t watchdog = 0;

  // watch out for STA disconnects and reboot if
  // connection cannot be reestablished after
  // X minutes

  // this watchdog MUST NOT reboot the device if
  // wifi is not enabled or never was connected
  // since last reboot

  if (wifi_is_enabled && watchdog_enabled && ((millis() - ms) > 1000 * 6)) {
    ms = millis();

    if (!wifi_is_connected) {
      if (++watchdog < watchdog_timeout * 10) { // timeout in minutes ( * 10 )
        if (watchdog == 1) {
          log_print(F("WIFI: arming network watchdog (reboot in %i min.)"),
            watchdog_timeout
          );
        }
      } else {
        log_print(F("WIFI: still not connected, triggering reboot ..."));
        system_reboot();
      }
    } else {
      if (watchdog) {
        log_print(F("WIFI: network is back, disarming watchdog"));
        watchdog = 0;
      }
    }
  }
}

[vseven]

Yeah, I don't see a Loop() in the library I'm using for WiFi (again, third party) but I understand what you are doing and can run with it.

Going to turn on the verbose logging and see if I get any more info first before changing anything.

Also the WiFi Events that are being checked and acted on are SYSTEM_EVENT_STA_GOT_IP, SYSTEM_EVENT_STA_DISCONNECTED, SYSTEM_EVENT_STA_START, and SYSTEM_EVENT_STA_CONNECTED. There are more however at https://github.com/espressif/esp-idf/blob/master/components/esp32/include/esp_event.h so I added a default: to the switch that should print out the event to see if the library is getting something but just not acting on it.

I saw in the library I'm using a //WiFi.setAutoReconnect(true); so I'm assuming the original author tried to use auto reconnect but it wasn't working so commented it out?

[vseven]

@me-no-dev - Verbose gave a much better explanation. Well...maybe to you:

Everything was fine then got this over and over, maybe 150 times:

[E][WiFiClient.cpp:97] connect(): lwip_connect_r: 113

Followed by this:

[D][WiFiGeneric.cpp:182] _eventCallback(): Event: 5 - STA_DISCONNECTED
[W][WiFiGeneric.cpp:187] _eventCallback(): Reason: 7 - NOT_ASSOCED
[WiFi-event] event: 5
WiFi lost connection.  Attempting to reconnect...
[D][WiFiGeneric.cpp:182] _eventCallback(): Event: 5 - STA_DISCONNECTED
[W][WiFiGeneric.cpp:187] _eventCallback(): Reason: 202 - ASSOC_FAIL
[WiFi-event] event: 5
WiFi lost connection.  Attempting to reconnect...
[D][WiFiGeneric.cpp:182] _eventCallback(): Event: 5 - STA_DISCONNECTED
[W][WiFiGeneric.cpp:187] _eventCallback(): Reason: 2 - AUTH_EXPIRE
[WiFi-event] event: 5
WiFi lost connection.  Attempting to reconnect...
[D][WiFiGeneric.cpp:182] _eventCallback(): Event: 5 - STA_DISCONNECTED
[W][WiFiGeneric.cpp:187] _eventCallback(): Reason: 2 - AUTH_EXPIRE
[WiFi-event] event: 5

The the last part (reconnect, event 5, reason 2) just repeats every 4 - 5 seconds.

[vseven]

Based on that first error that repeats over and over I found this: #180. In it you said "Try flushing the client or reading all available data. ". How would I go about doing that? I'm looking at the library I'm using and this is what they have to catch the WiFi Event:

//**************************************************************************************
/// Event Handler for ESP32 WiFi Events (needed to implement reconnect logic for now...)
//**************************************************************************************
	void MyESP32WiFi::WiFiEvent(WiFiEvent_t event)
	{
		Serial.printf("[WiFi-event] event: %d\n", event);

		switch (event) {
		case SYSTEM_EVENT_STA_GOT_IP:
			Serial.println("WiFi connected");
			Serial.println("IP address: ");
			Serial.println(WiFi.localIP());
			break;
		case SYSTEM_EVENT_STA_DISCONNECTED:
			Serial.println("WiFi lost connection.  Attempting to reconnect...");
			WiFi.reconnect();
			break;
		case SYSTEM_EVENT_STA_START:
			Serial.println("ESP32 station start");
			break;
		case SYSTEM_EVENT_STA_CONNECTED:
			Serial.println("ESP32 station connected to AP");
			break;
		default:			
			Serial.println("Unhandled WiFi Event raised.");
			break;
		}

	}

Based on the errors in the previous post I'm getting the "5" which is translating to the WiFi connection lost and then the code is trying to execute "WiFi.reconnect();" but its obviously not working. Should I have something else there? Maybe a counter kinda like @everslick example but simpler (just a i = i + 1) and if it end up on the same place after 20 tries it reboots? Or should this not happen/be prevented in the first place?

[me-no-dev]

Try calling WiFi.begin() instead of reconnect and report back :)

[vseven]

@me-no-dev - I changed it to WiFi.begin() instead of WiFi.reconnect() but it didn't help. Worked for about a hour then got the lwip_connect_r:113 line, each one about a minute or so apart followed by the disconnect every 5 or so seconds:

[E][WiFiClient.cpp:97] connect(): lwip_connect_r: 113
[E][WiFiClient.cpp:97] connect(): lwip_connect_r: 113
[E][WiFiClient.cpp:97] connect(): lwip_connect_r: 113
[E][WiFiClient.cpp:97] connect(): lwip_connect_r: 113
[D][WiFiGeneric.cpp:182] _eventCallback(): Event: 5 - STA_DISCONNECTED
[W][WiFiGeneric.cpp:187] _eventCallback(): Reason: 7 - NOT_ASSOCED
[WiFi-event] event: 5
WiFi lost connection.  Attempting to reconnect...
[D][WiFiGeneric.cpp:182] _eventCallback(): Event: 5 - STA_DISCONNECTED
[W][WiFiGeneric.cpp:187] _eventCallback(): Reason: 2 - AUTH_EXPIRE
[WiFi-event] event: 5
WiFi lost connection.  Attempting to reconnect...
[D][WiFiGeneric.cpp:182] _eventCallback(): Event: 5 - STA_DISCONNECTED
[W][WiFiGeneric.cpp:187] _eventCallback(): Reason: 2 - AUTH_EXPIRE

And then the 4 lines just repeat indefinitely until the control is rebooted. What else can I try? The fact @everslick is rebooting the controller when its in the same disconnected state is worrisome....I would think rebooting is the last option and something can be done to keep it online.

As a side note two ESP8266's connected to the same router (Netgear R8000) have no issues and have been stable for over a month without disconnects.

[me-no-dev]

hmmm.... this seems like an issue either in the lower stack or in your router.
you can see the first reason being NOT_ASSOCED and then you get AUTH_EXPIRE.
could you maybe turn off the STA and then turn it back on? The disconnect cause makes me think that something in the re-assoc request is either not changed or your router does not accept, so it refuses you to connect back. I'll see if I can raise an internal issue about this. What is the router firmware version that you are using?

[vseven]

Its a Netgear Nighthawk R8000. Currently on the latest firmware V1.0.3.54_1.1.37 although it did it on the last version also. I have 20 - 25 devices connected at any given time without issues including the previously mentioned two ESP8266's.

I will attempt to replace the Wi-Fi begin/reconnect line with the STA off and on and see what happens.

Edit: Knowing not much about this and using someone elses library....how do I "turn off the STA and then turn it back on"?

Second Edit: It couldn't be related to espressif/esp-idf#499 (comment) could it? One of the only things I've seen with the same NOT_ASSOCED and AUTH_EXPIRE. I wouldn't think so only because I compiled myself without the BLE stuff but never know.

[me-no-dev]

I have also alerted the wifi team :) we will see if anything comes out of this