Description of problem:

It appears that when running https://github.com/redhat-ai-dev/llama-stack-agentic-sample if we restart llamastack after ingesting our docs into RAG, the app can no longer access the RAG content.

Restarts of the app, but leaving llamastack up, works OK

Prerequisites (if any, like setup, operators/versions):

Steps to Reproduce

•  bring up https://github.com/redhat-ai-dev/llama-stack-agentic-sample and populate RAG
• bring down app and llamastack, but do not clean up RAG db i.e. you ~/.llama folder
• bring it all back up

Actual results:

The app correctly detects the RAG is present, but then attempts to query the RAG DB fail, and an error log shows up in llamastack with keys for items in the vector store being unavailable.

Expected results:

it should work

Reproducibility (Always/Intermittent/Only Once):

Build Details:

Additional info (Such as Logs, Screenshots, etc):