ems-esp device does not respond after few hours

[maciejka104]

Strarting from version 3.6.3 my bbqkees ems-esp S32 after few hours is unavalivable.
Currently I am running 3.6.3-dev.1 and it works fors up to 6-8 hurs and then it becomes dead.
No ping no mqtt,

I have checked syslog and mqtt connected disconnected ....

<28>1 2023-10-09T18:02:52.085523+02:00 ems-esp mqtt - - - mqtt- 000+00:02:36.323 W 172: MQTT disconnected: TCP
<30>1 2023-10-09T18:02:56.770770+02:00 ems-esp mqtt - - - mqtt- 000+00:02:41.005 I 183: MQTT connected
<28>1 2023-10-09T18:32:57.233403+02:00 ems-esp mqtt - - - mqtt- 000+00:32:41.509 W 1407: MQTT disconnected: TCP
<30>1 2023-10-09T18:33:02.576616+02:00 ems-esp mqtt - - - mqtt- 000+00:32:46.851 I 1414: MQTT connected
<28>1 2023-10-09T18:37:59.882780+02:00 ems-esp mqtt - - - mqtt- 000+00:37:44.089 W 1617: MQTT disconnected: TCP
<30>1 2023-10-09T18:38:42.374151+02:00 ems-esp mqtt - - - mqtt- 000+00:38:26.580 I 1655: MQTT connected

<28>1 2023-10-09T04:24:54.310904+02:00 ems-esp mqtt - - - mqtt- 000+01:24:38.175 W 4586: MQTT disconnected: TCP
<30>1 2023-10-09T04:24:58.280836+02:00 ems-esp mqtt - - - mqtt- 000+01:24:42.143 I 4592: MQTT connected
<28>1 2023-10-09T04:54:58.504756+02:00 ems-esp mqtt - - - mqtt- 000+01:54:42.352 W 5886: MQTT disconnected: TCP
<30>1 2023-10-09T04:55:04.300629+02:00 ems-esp mqtt - - - mqtt- 000+01:54:48.147 I 5897: MQTT connected

My setup is EMS-ESP bbqkees connected to Junkers Cerapur + FW100.

What could be the problem? I have checked mqtt (on HA) and looks/works fine.

[proddy]

Damn, I was hoping we had fixed this issue. It looks like EMS-ESP is going into a MQTT re-connect loop. Perhaps we need to add more delay on the retry @MichaelDvP ?

@maciejka104 did you notice a WiFi glitch or anything that time?

[maciejka104]

@proddy - wifi looks ok
I will double check on WLC logs
wifi setup - cisco 9800 wlc + 3 x C9105

[MichaelDvP]

The mqtt reconnects are not very often and in one case it took 49sec to reconnect. I found mqtt reconnects with IPv6 enabled on every change of IP (dynamic IP from provider) and fixed this in my testbuild here.

I'm always missing the support info in most new issues. Without knowing the config it's difficult to analyse possible reasons.

Since we have a lot of stable ems-esp and only a few with "freezes" it's mainly something in the environment of the users.
#1320 reminds me that the full network scan takes more power for wifi and low/unstable power supply could have these effects.

Users with good rssi to next AP could test with reduced tx-power.

[maciejka104]

IPv6 disabled on network and on the ems-esp
wifi coverage it is really good (one of the APs is 5-6 meters away and thin wall)
before raising a new issue I have checked usb cable and power supply.

I have external syslog server so I can provide full log for both wifi and ems-esp.

[proddy]

let's continue this discussion in #1321

[proddy]

let's continue this discussion in #1321

[proddy]

let's continue this discussion in #1321