Server with long running callback - send ACKNOWLEDGE to avoid timeout

[truschival]

I am writing a server using RunAsyncTCPServer and following the example server_callback.py I created a Datablock that overwrites getValues() and validate().
My problem is that getValues() takes longer than my ModbusTCP client expects, I get timeout errors. According to the Modbus Spec page 49 I think I could send a ACKNOWLEDGE message and later send the data using the same transaction id. I have no clue how to send a message from the callback methods.

I had a look at request_tracer or response_manipulator but it seems they don't really help here - or at least I have no idea how.

How would you deal with the situation of long-running callbacks for reading data?

[janiversen]

You cannot send and Ack at that level !

You are expected to return an error if the data is not ready, then the client can retry later.

[truschival]

Thanks. I think I wrote the question before thinking it thought.

I guess the better idea is to implement some sort of protocol on application layer i.e.

1. write a request,
2. poll for a finish-bit,
3. read result