HA integration: Not connected errors #2905
-
|
Hello, In my integration's @dataclass
class FlointData:
"Keeps the AsyncModbusTcpClient instance."
client: AsyncModbusTcpClient
async def async_setup_entry(hass: HomeAssistant, entry: ConfigEntry) -> bool:
"""Set up from config entry."""
ip_address = entry.data["IP address"]
client = AsyncModbusTcpClient(ip_address, port=502)
await client.connect()
# Save the client instance to reuse it from entities
entry.runtime_data = FlointData(client=client)
await hass.config_entries.async_forward_entry_setups(entry, PLATFORMS)
return TrueIn the class TotalImportedEnergy(SensorEntity):
def __init__(self, entry: ConfigEntry):
self.client = self._entry.runtime_data.client
async def async_update(self) -> None:
"""Fetch new state data for the sensor."""
_LOGGER.debug("Client state before reading registers: %s", self.client)
_LOGGER.debug("Client %s", type(self.client))
_LOGGER.debug(
"is socket open? %s",
self.client.connected,
)
# Error happens in next line
result = await self.client.read_input_registers(13036, count=2, device_id=1)
if result.isError():
_LOGGER.error(
"Error reading registers for IP %s, register 13036: %s",
self._ip_address,
result,
)
return
value = self.client.convert_from_registers(
result.registers, data_type=self.client.DATATYPE.UINT32, word_order="little"
)
self._attr_native_value = value * 0.1the marked line gives me this backtrace: Claude told me, that the connection is silently dropped, but https://pymodbus.readthedocs.io/en/latest/source/client.html#asynchronous-example says that it should auto-reconnect. Also, there is no other instance using that client. What can be the problem here? Thanks! |
Beta Was this translation helpful? Give feedback.
Replies: 4 comments 7 replies
-
|
It is hard to see what happens since you have not enabled pymodbus debut log. Did the client connect to the device and then lost connection...or did it never connect ? The connection is never dropped in a client...but if the device drops the connection it will automatically reconnect. Please activate the pymodbus debug log, then you will get a lot more information about what actually happens. It would also be good to check the return code of the connect call. |
Beta Was this translation helpful? Give feedback.
-
|
a couple of notes:
|
Beta Was this translation helpful? Give feedback.
-
|
Good Morning!
That is not correct, it the output of the script as pasted here. I would guess the repeating patter comes from the reconnect attempts during the during the sleep.
The logging code is taken from Modbus Logging Example, not from an AI. When searching for "debug" this turns up first Async Asyncio Client Example, put I fail to see a special call to enable debugging. Would you mind giving a hint?
Yeah, it does make sense here in this example. This was to simulate my HA behavior, where I want to keep and re-use the connection (via the client object). But it looks like that it is dropped almost instantly, as even during 2 seconds sleep, there are multiple re-connect attempts.
But instantly, resp. after 2 seconds?
The modbus integration I am aware of allows to create sensor etc from modbus addresses, but is not usable from a integration, i.e., python code. Thanks for your help! |
Beta Was this translation helpful? Give feedback.
-
Yes there will be reconnect during the sleep, but not that many.
from pymodbus import pymodbus_apply_logging_config the modbus logging example do not exist as an example, I will investigate why it is in the documentation.
read your device manual.
That is not true, there are a number of integrations that uses the modbus integration. Just like your example uses the sensor integration. |
Beta Was this translation helpful? Give feedback.
-
|
The example you refer to is from another project !!! this is the pymodbus project. All our examples like e.g. https://pymodbus.readthedocs.io/en/latest/source/examples.html#ready-to-run-examples contains the correct call. |
Beta Was this translation helpful? Give feedback.
-
Ok, my apologies! I did not know there were two projects with a very similar name. What is pymodbus-n about? It is just an old version of that project (pymodbus-dev)? The "Edit on GitHub" link from the pymodbus-n docs links to that projects here...?
For completeness, the code now looks like import logging
from pymodbus import pymodbus_apply_logging_config
from pymodbus.client import AsyncModbusTcpClient
import asyncio
registers = {"version_1": 2581, "version_2": 2596, "load_power": 13007}
async def get_load_power(client):
result = await client.read_input_registers(
registers["load_power"], count=2, device_id=1
)
if result.isError():
print("Error reading registers")
return None
value = client.convert_from_registers(
result.registers, data_type=client.DATATYPE.INT32, word_order="little"
)
return value
async def main():
pymodbus_apply_logging_config("DEBUG")
client = AsyncModbusTcpClient("192.168.0.36")
print("Connecting to device...")
await client.connect()
print("Connected to device")
value = await get_load_power(client)
print("Load Power 1:", value)
await asyncio.sleep(2)
value = await get_load_power(client)
print("Load Power 2:", value)
client.close()
if __name__ == "__main__":
asyncio.run(main())producing log output: My conclusion so far
Can you agree with that?
I will investigate, thanks for the hint. Best, Florian |
Beta Was this translation helpful? Give feedback.
-
|
I have no idea what pymodbus-n is about, I suggest you ask the project ! Your conclusion seems correct. |
Beta Was this translation helpful? Give feedback.
-
|
Please be aware of the timing: Recv at 10:08:22,830 So it is very likely that the device did not like the request, are you sure the address is an input register, and not e.g. a holding register. Later you see that pymodbus connects successfully, but the device immediately closes, which typically signals the device is in a problematic state, and maybe needs a reset. |
Beta Was this translation helpful? Give feedback.
The example you refer to is from another project !!! this is the pymodbus project.
All our examples like e.g. https://pymodbus.readthedocs.io/en/latest/source/examples.html#ready-to-run-examples contains the correct call.