chore: fix flaky unit test and add proper shutdown for file batches (#3725)

# What does this PR do?
Have been running into flaky unit test failures:
5217035494
Fixing below
1. Shutting down properly by cancelling any stale file batches tasks
running in background.
2. Also, use unique_kvstore_config, so the test dont use same db path
and maintain test isolation
## Test Plan
Ran unit test locally and CI

llama_stack/providers/inline/vector_io/faiss/faiss.py

@@ -225,8 +225,8 @@ class FaissVectorIOAdapter(OpenAIVectorStoreMixin, VectorIO, VectorDBsProtocolPr
         await self.initialize_openai_vector_stores()
 
     async def shutdown(self) -> None:
-        # Cleanup if needed
-        pass
+        # Clean up mixin resources (file batch tasks)
+        await super().shutdown()
 
     async def health(self) -> HealthResponse:
         """

llama_stack/providers/inline/vector_io/sqlite_vec/sqlite_vec.py

@@ -434,8 +434,8 @@ class SQLiteVecVectorIOAdapter(OpenAIVectorStoreMixin, VectorIO, VectorDBsProtoc
         await self.initialize_openai_vector_stores()
 
     async def shutdown(self) -> None:
-        # nothing to do since we don't maintain a persistent connection
-        pass
+        # Clean up mixin resources (file batch tasks)
+        await super().shutdown()
 
     async def list_vector_dbs(self) -> list[VectorDB]:
         return [v.vector_db for v in self.cache.values()]

llama_stack/providers/remote/vector_io/chroma/chroma.py

@@ -167,7 +167,8 @@ class ChromaVectorIOAdapter(OpenAIVectorStoreMixin, VectorIO, VectorDBsProtocolP
         self.openai_vector_stores = await self._load_openai_vector_stores()
 
     async def shutdown(self) -> None:
-        pass
+        # Clean up mixin resources (file batch tasks)
+        await super().shutdown()
 
     async def register_vector_db(
         self,

llama_stack/providers/remote/vector_io/milvus/milvus.py

@@ -349,6 +349,8 @@ class MilvusVectorIOAdapter(OpenAIVectorStoreMixin, VectorIO, VectorDBsProtocolP
 
     async def shutdown(self) -> None:
         self.client.close()
+        # Clean up mixin resources (file batch tasks)
+        await super().shutdown()
 
     async def register_vector_db(
         self,

llama_stack/providers/remote/vector_io/pgvector/pgvector.py

@@ -390,6 +390,8 @@ class PGVectorVectorIOAdapter(OpenAIVectorStoreMixin, VectorIO, VectorDBsProtoco
         if self.conn is not None:
             self.conn.close()
             log.info("Connection to PGVector database server closed")
+        # Clean up mixin resources (file batch tasks)
+        await super().shutdown()
 
     async def register_vector_db(self, vector_db: VectorDB) -> None:
         # Persist vector DB metadata in the KV store

llama_stack/providers/remote/vector_io/qdrant/qdrant.py

@@ -191,6 +191,8 @@ class QdrantVectorIOAdapter(OpenAIVectorStoreMixin, VectorIO, VectorDBsProtocolP
 
     async def shutdown(self) -> None:
         await self.client.close()
+        # Clean up mixin resources (file batch tasks)
+        await super().shutdown()
 
     async def register_vector_db(
         self,

llama_stack/providers/remote/vector_io/weaviate/weaviate.py

@@ -347,6 +347,8 @@ class WeaviateVectorIOAdapter(
     async def shutdown(self) -> None:
         for client in self.client_cache.values():
             client.close()
+        # Clean up mixin resources (file batch tasks)
+        await super().shutdown()
 
     async def register_vector_db(
         self,

llama_stack/providers/utils/memory/openai_vector_store_mixin.py

@@ -293,6 +293,19 @@ class OpenAIVectorStoreMixin(ABC):
         await self._resume_incomplete_batches()
         self._last_file_batch_cleanup_time = 0
 
+    async def shutdown(self) -> None:
+        """Clean up mixin resources including background tasks."""
+        # Cancel any running file batch tasks gracefully
+        if hasattr(self, "_file_batch_tasks"):
+            tasks_to_cancel = list(self._file_batch_tasks.items())
+            for _, task in tasks_to_cancel:
+                if not task.done():
+                    task.cancel()
+                    try:
+                        await task
+                    except asyncio.CancelledError:
+                        pass
+
     @abstractmethod
     async def delete_chunks(self, store_id: str, chunks_for_deletion: list[ChunkForDeletion]) -> None:
         """Delete chunks from a vector store."""