Modbus freezes from time to time

[djfanatix]

Describe the bug

As discussed on the HA community forum, hereby some logs when Solaredge modbus crashes.
I have an inverter with display (upgraded with Setapp, latest FW), a meter and Storedge.

I use a modbus proxy, but the crashes also when I tested the integration with a direct connection without other modbus interface.
I know have an automation to restart the modbus proxy when the freezing occurs.

Expected behavior

work fine

Screenshots

No response

Diagnostic File

solaredgelog.log

Debug logs

see attached, please ignore the template errors. (not relevant)

Home Assistant Version

2024.10.2

solaredge-modbus-multi Version

3.0.0

Installation Type

HAOS

Read the Instructions

• My configuration follows https://github.com/WillCodeForCats/solaredge-modbus-multi/wiki/Configuration
• I have read the Known Issues https://github.com/WillCodeForCats/solaredge-modbus-multi/wiki/Known-Issues
• This issue is not about a template error (do not report template errors, read https://github.com/WillCodeForCats/solaredge-modbus-multi/wiki/Template-Design-Notes)

Additional Context

No response

[WillCodeForCats]

Turn off the "Keep Modbus Open" option.

[WillCodeForCats]

I also need you to upload your diagnostic file and turn on pymodbus debugging:

https://github.com/WillCodeForCats/solaredge-modbus-multi/wiki/Integration-Diagnostics

[WillCodeForCats]

The issue appears to be that pymodbus isn't connected after sending a connect call:

2024-10-12 07:02:40.521 DEBUG (MainThread) [custom_components.solaredge_modbus_multi.hub] Connecting to 192.168.1.120:1502 ...
2024-10-12 07:02:40.527 DEBUG (MainThread) [custom_components.solaredge_modbus_multi.hub] coordinator timeout is 10.8
2024-10-12 07:02:46.532 ERROR (MainThread) [pymodbus.logging] Cancel send, because not connected!
2024-10-12 07:02:46.532 DEBUG (MainThread) [custom_components.solaredge_modbus_multi.hub] Disconnectng from 192.168.1.120:1502 (clear_client=False).

[djfanatix]

home-assistant_solaredge_modbus_multi_2024-10-13T19-34-18.165Z.log

Hereby with Pymodbus logs attached.

• When it crashes I have an automation to restart the modbus, this usually fixes it again.
• Could it be that some automations are writing too many battery commands? or is this not relevant? (it used to work fine before)
• To be clear, your integration worked fine for a very long time with my very old ancient firmware, it only happened after the new SE firmware and Setapp instead of display control.

[drumtrich]

The issue appears to be that pymodbus isn't connected after sending a connect call:

2024-10-12 07:02:40.521 DEBUG (MainThread) [custom_components.solaredge_modbus_multi.hub] Connecting to 192.168.1.120:1502 ...
2024-10-12 07:02:40.527 DEBUG (MainThread) [custom_components.solaredge_modbus_multi.hub] coordinator timeout is 10.8
2024-10-12 07:02:46.532 ERROR (MainThread) [pymodbus.logging] Cancel send, because not connected!
2024-10-12 07:02:46.532 DEBUG (MainThread) [custom_components.solaredge_modbus_multi.hub] Disconnectng from 192.168.1.120:1502 (clear_client=False).

I think I have the same problem. Appeard first time a few days ago and today again.

Dieser Fehler wurde von einer benutzerdefinierten Integration verursacht

Logger: pymodbus.logging
Quelle: custom_components/solaredge_modbus_multi/hub.py:506
Integration: SolarEdge Modbus Multi (Dokumentation, Probleme)
Erstmals aufgetreten: 11:41:15 (152 Vorkommnisse)
Zuletzt protokolliert: 15:40:11

Cancel send, because not connected!

Here my diagnostic file and logs, maybe it can help you to find the problem

config_entry-solaredge_modbus_multi-01JBEQY3F3CQFE50B72DWABGRF (1).json
home-assistant_solaredge_modbus_multi_2024-11-05T15-12-34.479Z.log

[GpdVatter]

The update of the Inline Meter to 1.0.90 fix the bug within my homeassistsnt environment

[drumtrich]

How to update the inline Meter? My has Firmware: 1.0.81 Am So., 24. Nov. 2024 um 17:46 Uhr schrieb GpdVatter < ***@***.***>:

…

The update of the Inline Meter to 1.0.90 fix the bug within my homeassistsnt environment — Reply to this email directly, view it on GitHub <#687 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AHUEKP4KKQMKDTZQ5R3Z6N32CH7HNAVCNFSM6AAAAABP2EMCSKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDIOJWGEYDEOJTGU> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.*** com>

[GpdVatter]

How to update the inline Meter? My has Firmware: 1.0.81 Am So., 24. Nov. 2024 um 17:46 Uhr schrieb GpdVatter < @.***>:
…

with the SolarEdge Inverter SetApp application on mobile device

[FrankTub]

I also experience this behavior with a polling frequence of 5 seconds. @GpdVatter , does every Solaredge installation have an inline meter? If so, I'm struggling where to find it using the direct wifi connection as described here. I guess that has the same options as would be available in the setapp without the need to install that app and create an account?

EDIT: I also have option Keep Modbus Connection Open set to False

[WillCodeForCats]

I've not yet seen modbus freezing personally.

My setup is connected with wired Ethernet and a 30 second polling rate. One meter (hardwired on the inverter 485 bus), no batteries, no commands, no proxies.

This is my diagnostic file with my settings:
config_entry-solaredge_modbus_multi-e3b4ad891a831bb72b37a1247705753f(5).json

[GroteGehaktBal]

I have the same problem. For me it started happening after adding batteries to the system. I have two inverters and 1 is working fine and the other one (with the inline meter and batteries connected to it) has this problem. The inline meter has Firmware 1.0.104. The Inverter has firmware 0004.0023.0027 and it is a SE10K-RWB48BFN4. I get this problem with either Keep Modbus Connection Open True or False.

[RedWheiler80]

• When it crashes I have an automation to restart the modbus, this usually fixes it again.

Can you explain how you did this? I do not find anything modbus related in Home Assistant it seems.

So yeah, I also do have the problem from time to time it freezes. Just on my newest invertor (I do have 2). Most of the time everything works great, but suddenly the new one is not available anymore. Not after a restart of HA also.

Only solution I did find: log in to the inverter with its own wifi and go to the IP of it. Then change my modbus port and also in HA. Then it works again. Also this thing happens both if I use the same port on both devices or each it's own port.

[xperseguers]

I had a similar problem, being with batteries as well. But somehow I fixed it by blocking my inverter to connect to SolarEdge website to send data. I constantly had the problem +/- over noon so I created a firewall rule that just blocks everything from around 11 AM to (3-4 PM). This means the monitoring app is "frozen" during that period but HA keeps the access and the SolarEdge platform catches up when the inverter is "online again".