Connection time out

[rzulian]

I'm getting a lot of connection time out using a polling interval below 10 secs. I was getting the same problems with other component for modbus-solaredge.
I think that this is related to the new HA version 7.2, and I was hoping that this component wasn't affected using a much more recent pymodbus library.

Do you have the same problems?

[WillCodeForCats]

Tested on 2022.7.3 working OK with 1 second interval for 2 inverters and 1 meter.

Video:
https://user-images.githubusercontent.com/48533968/178546814-91d76a4e-bb93-4e49-a188-7f911437a3c9.mp4

[rzulian]

I need to investigate what's wrong on my side then ...
Thanks

[rzulian]

I was suspecting that I had a performance problem in my HA in a VM. So I've tested a dev container with only this component, but I'm still unable to set the polling to less than 10 secs. @WillCodeForCats Any idea on what I can check?

[WillCodeForCats]

How long did you wait after changing the time? Let it wait for 2 minutes and see if it's just an issue with the inverter thinking the old connection is still open.

[rzulian]

I've tried your suggestion and I've also restarted the inverter, without success. I've restored HA Core 2022.6.7, which, to my knowledge, was working fine, but it isn't.
I've tried to use ```pymodbus==3.0.0dev3", which is actually working without problem. Only an annoying Warning

2022-07-14 13:23:22.667 WARNING (MainThread) [asyncio] Executing <Task finished name='Task-2317' coro=<SolarEdgeModbusMultiHub.async_refresh_modbus_data() done, defined at /workspaces/core/config/custom_components/solaredge_modbus_multi/hub.py:190> result=None created at /workspaces/core/homeassistant/core.py:452> took 0.316 seconds

[WillCodeForCats]

PR #59 might help with the immediate connection error after reconfigure.

[rzulian]

Are you going to merge also the data_update_coordinator. I would test the new main with also the battery support

[WillCodeForCats]

Are you going to merge also the data_update_coordinator. I would test the new main with also the battery support

No, I'm not.

[rzulian]

BTW, digging into a very slow trigger issue, I was discussing about the possibility that the solaredge_modbus can be athe problem. In the discord channel they suggested me to "It looks like that package does the same thing. It should be an easy fix, you can just change line 187 there to use an executor job like "update_result = await self._hass.async_add_executor_job(self._read_modbus_data)"
But the better way to do it is to use the asyncio methods which the pymodbus library seems to have".

[WillCodeForCats]

I have been looking at converting it to use the async client of pymodbus. The first step is to use the data update coordinator then probably async pymodbus later. The branch is there but it doesn't work yet. Once I get it to work with sync then I will work on async.

Because it currently uses sync if you have a fast polling interval it can halt all home assistant execution while reading or waiting for a modbus timeout, so fast polling is not advised.

[WillCodeForCats]

I will also consider the async_add_executor_job wrapper if it's faster to implement

[WillCodeForCats]

The data-update-coordinator branch is working now, you could try it out if you want. It has not been tested, I've only done enough work so that it loads and runs.