stable ? Yes.....

[patience4711]

To prove that the system is stable, please see picture.

[swbouman]

Here too! Running 7 inverters on a wemos d1

[ErikKok]

After months of use without any problems or interaction I noticed the values did not update anymore in HA.

I could log in the webpage of the ESP which displayed the last values. And I noticed the last entry in the journal was a wake up about 6 days ago. The uptime was about 3 days. After a warm reboot, nothing changed, uptime was still 3 days, values stayed the same. I interrupted the power, ESP was not pingable, after a second power cycle everything worked again.

Don't know what could be the reason, but wanted to let you know. Uptime more then 99 days? Maybe I just hit some simple limit or a cosmic ray?

I had a check in HA running on the ESP being pingable, but added a check with this as a trigger (>sunrise && <sunset): {{ (now() | as_timestamp ) - as_timestamp(states.sensor.aps_esp_west1_power_mppt_1.last_changed) > 600 }}

[swbouman]

system uptime 31 d 21 h 39

[Hansiart2]

@ErikKok Stable doesn't mean forever it seems. The state in what it was is unclear to me. Not frozen as the webinterface was still responsive. But not polling. The uptime points to a reset, that can have many reasons. It could be that the zigbee was in a certain state that for the esp it seems it is up where it is not. Due to the powercycle it came out of this state. Also could be that there is still a small memory leak that i can't get rid of. So no cosmic ray involved here. I think it won't hurt to repower the EECU from time to time. If it worked for months then its impossible to figure what happened. So i think the good news is that it worked for months.

[ErikKok]

Happened again today. Could not log in, and had to disconnect the power 4-5 times before it would actually boot up. After that it still did not recieve info from the inverters, but after a reboot through the menu it did straight away. Now it's working again.

It not a big problem, but I do loose some data, espcially if it happens on a holiday. Maybe you could implement an automatic reboot each night / x days? I know, sounds like a lame way to fix, but if it works it works. Then again, maybe this crash has nothing to do with the uptime...

Again no real information to act on, but just FYI.

[patience4711]

This is very strange. When is was a memory leak, a simple reboot would solve that but in your case, you had to repower (reboot) 5 times before it was back online. So it must be something else that causes the pain, maybe the network connection? I had the automatic reboot every night implemented in an earlyer version but thought that it wasn't needed anymore. I guess there is nothing against it but the thing is that when it freezes, it would not reboot either so this wouldn't help much.

[joffito]

i can't seem to get long uptimes like 100d+ but it works for me without issues since 1 year.

maybe you could implement a watchdog timeout of 8s to deal with these kind of freezes?

[ErikKok]

I don't know why but the last 2 weeks it's been freezing every few days... Every time I need multiple hard resets, i actually wrote a little script with a Shelly to automate this.

wifi signalstrength atm is -59dB, should be pretty good right?