Merge pull request #178439 from nikitapitliya/patch-19

Update develop-deploy-grpc-inference-srv.md

Author: PRMerger18

articles/azure-video-analyzer/video-analyzer-docs/edge/develop-deploy-grpc-inference-srv.md

@@ -35,9 +35,6 @@ To create a gRPC inference server and implement it as an extension with Video An
 
 Perform the necessary steps to have Video Analyzer module deployed and working on an IoT Edge device.
 
-> [!NOTE]
-> The deployment manifests use a tag of 1 (video-analyzer:1). So simply redeploying such manifests should update the module on your edge  devices when newer tags are released.
-
 ### High level implementation steps
 
 1. Choose one of the many languages that are supported by gRPC: C#, C++, Dart, Go, Java, Node, Objective-C, PHP, Python, Ruby.
@@ -67,6 +64,9 @@ When collocating on the same node, `shared memory` can be used for best performa
 1. Upon receiving of a frame, access the address offset within the shared memory.
 1. Acknowledge the frame processing completion so its memory can be reclaimed by Video Analyzer.
 
+> [!NOTE]
+> When using a gRPC extension module for inferencing with shared memory, both the Video Analyzer edge module and the extension module should run in the same [user and group](https://docs.docker.com/engine/reference/builder/#user)
+
 ## Create a gRPC inference server
 
 Now you will build an IoT Edge module (External AI) that accepts video frames from Video Analyzer using [protobuf](https://github.com/Azure/video-analyzer/tree/main/contracts/grpc) messages via shared memory, classify the frames as **dark** or **light** and return inference results back to the IoT Hub Message Sink in Video Analyzer using the inference metadata schema.