I have three plugins (URCaps). I do not know the order of their initialization. All of them have their own daemon implemented in python. “Communication layer” is XML-RPC. In python, during initialization, connection with RTDE should be established. For updating the IN and all other pages, I’m not using the Java Timer, like you have it in examples, instead, I’m using a thread, which is created and started in Installation node constructor. During the controller bootup, it happens that I have urcap full initialized and ready (from the logs), with established connection with the RTDE, but after the second urcap (installation node instance) starts with initialization, the connection with RTDE is lost, errors:
- lost connection with controller → __trigger_disconnected() when I try to receive data, e.g. speed slider or runtime state
- or socket is closed
The re-connection is periodically, after few seconds. I even postponed connection establishment with RTDE but I have no influence how the OS or UR SDK will create difference urcaps instances.
The controller is with 4 cores and 1.8GB of memory. Polyscope version is 5.11.0.something
Seems like not enough resources during the peak, but I’m not sure. Sound like well know problem with the “No controller”.
Is any way (api, …) how to control different urcaps initialization, so once they are all up and running to initiate connection with RTDE instead to having lot of retries.
I’m aware that this is workaround because controller issue, but do not have other choice.
Any help or answer appreciated.