Socket timeouts and socket closed

[francinv]

Background

I have two Python programs that runs on a RPi:

1. A worker that runs every second and fetches data from the inverter with Modbus/TCP. I only initialize a connection one time.
2. A SellProgrammer that is initialized every hour. I call programmer.disconnect()after each update.

The problem

In the start this works, but after some time (over 24 hrs) I start to get socket.timeout. As the time goes by the amount of timeouts increases until it suddenly throws: Socket closed. When I stop the service and start it again it however throws: Connection closed.

• Has anyone experienced issues like this before?

Other questions:

• Should I disconnect more often when reading?
• Should I store the socket to ensure I use the same one each time?

[davidrapan]

Hi @francinv, do you instantiate PySolarmanV5/Async w/ auto_reconnect = True?

[francinv]

Hi @francinv, do you instantiate PySolarmanV5/Async w/ auto_reconnect = True?

Yes, I do.

[davidrapan]

Try to take some inspiration from how I've done it here. It's pretty resilient.

[githubDante]

@francinv any not properly closed sockets when this starting to happen ?!? Can you check/provide the output from netstat or ss

  ss -nt | grep <logger address>