-
Type: Bug
-
Status: Resolved (View Workflow)
-
Priority: Medium
-
Resolution: Done
-
Affects Version/s: None
-
Fix Version/s: None
-
Component/s: Deprecated - VOLTHA, ofagent-py
-
Labels:
-
Story Points:3
-
Epic Link:
-
Sprint:VOLTHA Sprint 5
As part of the effort to support ofagent in a swarm deployment, a streaming GRPC call was implemented to monitor connectivity between OFAgent and Voltha.
The GRPC call is currently implemented as a bidirectional streaming call, meaning that a generator on the client side continuously sends requests to the server. The server takes these requests and continuously sends responses.
To avoid CPU intensive loops, delays have to be integrated. Unfortunately, the asynchronous sleep that is used throughout VOLTHA cannot be used within a generator. Therefore, a plain time.sleep had to be used. As a result, the ofagent will block for a period of time within the generator at every pass of the loop.
We need to find a better approach to avoid any blocking.