Retry timin clarification for client.connect, client.read_holding_register, etc (ModbusClient.AsyncModbusTcpClient pymodbus v3.68)

[Ducati748ss]

Hi
Is my following understanding of the retry timings for client.connect client.read_holding_register, etc (ModbusClient.AsyncModbusTcpClient pymodbus v3.68) correct?

As my application must rapidly provide a control system setpoint, I need to make sure I understand exactly the retry process and the exact timings I can expect.

1. The “reconnect_delay” doubles on each retry?
2. Once the “reconnect_delay_max” is reached, any future retries will be for its duration?
3. The “timeout” value will always take precedence, even if the number of retries (“retries”) is not met?
4. There is an initial try to communicate then N retries, as set by the value of “retries”?

Therefore, with these values:
timeout = 5s
retries = 3
reconnect_delay = 0.4s
reconnect_delay_max = 1.0s

Then approximate timings for the waits would be:
Try 0 = 0.4s
Retry 1 = 0.4s
Retry 2 = 0.8s
Retry 3 = 1.0s (capped at “reconnect_delay_max”)
wait 5.0s - 2.6s = 2.4s
return

That is, the total time trying the 4 tries is approximately 2.6 seconds, so now wait 2.4 seconds for a response from Retry 3 until the “timeout” of 5 seconds expires, then return.

Thank you very much
Mark Norman

[janiversen]

Sound just about right.