typxxi
commented
1 month ago All the mess seems to have a different cause. Growatt is overwriting our local IP adresses we changed in the stick settings to our local HA / Grott server IP while I was working on the AC side.
If you change the ip of the stick manually then it will work again - but only for the first packet and from the 2 nd on it will stop again cause growatt will have restored growatts own adress or so.
THIS IS MY STORY ABOUT A FULL BREAK DOWN OVER NIGHT which killed 8 growatt inverter "reportings"/monitoring
I had to deal with some flaws in our WLAN and network, I rewired the LAN cables, added a switch in the middle of a 50 m long LAN cable which connected a new MESH REPEATER (as LAN connected bridge) to the new ROUTER (Fritzbox 6590AX).
Nothing terrible, just usual maintenance even though a bit more this time :
Around that day I also started a rewiring of the central AC wiring box aka distribution box. You usually have only 1 in a house but due to new regulations here we needed a second one for a bigger solar installation to fullfill the newest requirements and safety whatsoever.
The former one has been from 2010 and has now been downgraded from the central distribution box incl. smartmeter to a kind of "floor" distribuition, but it stayed functional fully the same only the new additional one for 3500€ has become the "arrival point" where the new 3 phase Wallbox, the new heatpump and the solar inverters have been connected on 4 big "main line branch terminals" (each connected with 16 mm² / AWG 5 cables) .
Its like building a new central railway station in front of a big town as new hub. That whole rewiring in that central distribution box (1,5 x 1,2 m for just the heatpump, the wallbox, 2 smartmeter counters and the solar power systems) took about 2 or 3 days and during that period I did wire things up and had nothing to do with HomeAssistant, did not check anything. Summer and 25+ kWp in northern germany there is not much to fear cause the 14 kWh battery is usually full in the evening even on dark rainy days.
Long story but for a reason:
out of nowhere after all the LAN and WLAN and AC upgrades I noticed on a sunny morning that we had no more solar power data, nothing, no kWh produced, no Watt. None of the 8 inverters was there anymore even though I had not changed anything in HA for days. A week ago I had done a monthly HA upgrade.
But I was in hurry and had to do other stuff first. Later that day I came back and then got a warning from SPOOK of dead sensors , partly depreciated by the integrations so I started deleting them. These were not grott related sensors. After 20 or so I ended up somehow in deleting GROTT sensors like all the others before till I realised again that something must be totally wrong due to the sheer amount of grott sensors I had already and that this is not a usual depreciation of just 1 or 2 o3 sensors no longer supported (which equals for me in 8 or 16 or 24 sensors I have to delete).
At that point I was suddenly fully awake and checked how many more grott sensors were depreciated - over 390 which meant I had already deleted 60 or so.
I instantly decided to go back - and restore the backup from the evening before and did a FULL RESTORE OF HA. HA RESTORE Worked fine and within 5 minutes I was back in business.
But back means: only my deleted sensors were back again, but still without values.
It took me many hours and 2 sleepless nights to get behind the whole story cause the growatt web UI worked fine, the app worked fine only GROTT HA integration looks dead like hell. This night at 3 am I awoke and had a thought for the cause of all the issues and solved it and then started writing this down as an issue and warning.
========================================= Long story short: if your GROWATT INVERTER LOOSES POWER your DATALOGGER STICK looses your GROTT IP
I had to access every single inverter via growatt cloud ( I prefer this way over local IP cause it is faster and we still have the inverters in the CLOUD) and type in my server IP adress. If you have a power outage or used the circuitbreaker or whatever happens to your AC you have to be aware to change the datalogger IP adress to your grott server IP adress, here in my case HA Server.
I THINK THE DOCUMENTATION OF GROTT HA INTEGRATION AND GROTT NEED A WARNING regarding power outage cause you will forget about this topic and it might take a lot of work and time to figure it out. In my case many, many things have changes in about 3 or 4 or 5 days from new wallbox, LAN, WLAN, ROUTER and what not, but the key point were the inverters that had been disconnected from AC for minutes or maybe even hours and then lost the connection to HA or better GROTT.
You have no chance to identify that cause if you open the datalogger via Web the server IP is always shown empty in the settings dialog box and before on the web page with the internet IP growatt uses to adress and reach the "stick".
and here the "empty" dialog box
This is a screenshot from now after it is working again. The only way to find out is to connect locally to the stick once you have find out its IP adress.
I am running GROTT for about 4 or 5 months and I was not aware of this critical point every GROTT user should not on a big sticker on the HA machine or Growatt inverter not to forget to check the ip the datalogger stick is using in case of missing data in the grott integration.
Long story just to warn everyone cause it can happen after years and you would call everything else the cause but not the stick that has lost its IP adress - at least for me.
fun fact: the update intervall I had increased from 5 minutes to every minute was still active as you can see here:
And that makes it even more weird and strange that the datalogger stick looses the "home adress" but not the reporting intervall changes. Maybe looses is the wrong term rather it falls back to a factory default value.
Just to make it crystal clear for everyone: I bet this is key knowledge and lesson for everyone cause all our inverters have been hit the same way which are Min 4600, mic 3000, mic 2000, mic 600 and SPH inverters with in total at least 2 different kind of datalogger sticks, the shinewifi x and s !
I bet the same will happen to SPA or MOD or whatever growatt inverter users.
Might be important to upgrade the documentation of grott and grott ha integration and all others that grott requires to check the datalogger settings after any power outage locally or simply paste the grott server IP adress into that field.
The issue itself is solved for me, it might be most likely technicly no issue but the awareness will be lost or missing over time where to check what in such a case of every single power outage (maybe shorter ones might be buffered by the stick which also has a coin cell onboard if I remember right which can be an additional achilles wheel will tend to forget).
So make a big sticker inside your distribution box to check if grott sensors are still delivering data after a power outage and if not to paste the right Grott server adress into every single grott server. Why inside the distriution box: after power outage you usually check the RCD in the distribution box. A sticker on the HA server might not work nor seen if it happens.