Reboot ESP32 with ESPEASY from remote?

[Alpensepp]

Hi!
I have an ESP32 dev with ESPEASY running in a remote place.
Version: ESP_Easy_mega-20191116_esp32test_4M316k

The ESPs webserver gets stuck irregularly, I cant open it any more.

But I still can ping the ESP and it sends data via MQTT.
Is there a way to reboot it without use of its website?

[TD-er]

If you portscan it, does it react on port 80?
If so, you could try to send just the reboot url: http://ip/?cmd=reboot

Do you have any idea at what version this started to happen, as you state it happens regularly?

[TD-er]

I fired up my test ESP32 and it can be reproduced quite easily by clicking the disconnect button on the web interface.
I will make an issue for it.

port 80 of the web interface does not respond anymore, so the web server is not running so it seems.

[Alpensepp]

ESP_Easy_mega-20191116_esp32test_4M316k

When I do a portscan, port 80 is open. But it does not react to any command.
I just started with ESP32s and therefore started with ESP_Easy_mega-20191116_esp32test_4M316k .

When the ESP32 freezed this night, it had to do nothing special. It sent temperature data from a bme280 via MQTT every hour. Suddenly it stopped sending data.

It is quite important for me, that the ESP32 and ESPEASY run stable, because I have 1 1/2h drive to the place, where it is running to restart it.
Is it a common issue for an ESP32? Should I go for an ESP8266, instead? The ESP8266 I run at home seem to run stable.

[TD-er]

Is it a common issue for an ESP32? Should I go for an ESP8266, instead? The ESP8266 I run at home seem to run stable.

For now it may be better I guess.
I've been testing it here now and it does not seem to act well on WiFi reconnects.
I also see a lot of reconnect issues to the MQTT broker if it does have webserver issues.

And to be honest, I have not tested the ESP32 builds very well.

[Alpensepp]

OK, thank you for the quick reaction. I will go for the 8266.
Are there plans to fully support the ESP32 in future?

[grovkillen]

OK, thank you for the quick reaction. I will go for the 8266.
Are there plans to fully support the ESP32 in future?

Yes but first the stable release (very soon) and then the new GUI. We're only two guys and it's been taking longer than expected to get a grip of the project, but now we're gaining momentum.

[TD-er]

I've been running tests on ESP32 since your post this morning, but I'm quite worried about the WiFi issues's I'm seeing.
Sure there are plans to support ESP32 and I was under the impression that it was quite OK-ish for the ESP32 even though I have not tested a lot with it.

But now I'm not convinced we should include ESP32 as officially "released" on the upcoming release of ESPEasy.
The WiFi stability is abominable right now and I truly cannot imagine why I haven't seen more issues with it.

[Misiu]

@TD-ed reboot command should work?
According to the docs (https://www.letscontrolit.com/wiki/inde ... _Reference) reboot is internal.
There is an old topic about the reboot (viewtopic.php?t=5447), so I'm confused.

I'd like to be able to reboot ESP from my internal network if something happens. I have Home Assistant and I'd like to add button on the interface that will allow me to restart a specific ESPEasy node, that's why I need that http request for.

if ?cmd=reboot works then the docs should be updated.

[TD-er]

Have you tried if it works?
There is a PR pending for updating the command origin: https://github.com/letscontrolit/ESPEasy/pull/3109

[Misiu]

@TD-ed, yes I've tried that yesterday, but without success. Probably I screwed something.
I'll try in the following days, also with method @grovkillen suggested.
So if PR#3109 get merged this will be possible (send cmd=reboot)? Sadly I can't read that out from code changes

BTW what does restart command do (compared to reboot)? It isn't described in the docs, which one should we use? In my case I'd like to restart(or reboot) the unit after my mqtt server got down.
Ref: https://github.com/letscontrolit/ESPEas ... pp#L27-L40

[TD-er]

Well I don't know if it is the best idea to do it like this.

If your node is unable to keep a connection to the broker, there is very likely something wrong with the network connection.
So sending a HTTP GET request to the ESP node will probably also not work.

What you can do, is act on the MQTT disconnected event and start a timer in the rules.
If you get reconnected, you reset the timer to 0. (disable it)
When the timer reaches its end, it will fire a new event, in which you can call for a reboot.

This way the ESP node itself will perform the reboot, purely based on the MQTT disconnect/connect events.

[Misiu]

@TD-ed,

the connection to MQTT is restored automatically after some time.

I noticed that when I restart my Home Assistant instance (with MQTT server on it) my Wemos needs about a minute or two to reconnect.
If I reset them (unplug and plug power) they connect much faster.

My idea was to do a script in Home Assistant that will reboot all Wemos units after the system starts, but I think this can be done on ESP as you described.
I need to search on the forum, I'm sure I find something useful.

[TD-er]

Instead of a reboot, you can also call for a wifi disconnect.
This should perform a rather quick reconnect, which is a lot faster compared to a reboot.