fix: add missing milvus deps to llama-server targets across all agents (#98)

The shared run_llama_server.yaml configures inline::milvus as the
vector_io provider, but 8 agent Makefiles only installed llama-stack
and ollama — missing pymilvus, milvus-lite, and a setuptools upper
bound (<82.0.0) needed to avoid the removed pkg_resources API.

Also moves milvus_data directory creation from the ollama target into
llama-server where it actually belongs.

Closes: RHAIENG-5038

Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: mpk-droid

agents/crewai/websearch_agent/Makefile

@@ -58,7 +58,8 @@ ollama: ## Install Ollama, start it, and pull models
 llama-server: ## Install llama-stack and start Llama Stack server
 	@source .venv/bin/activate && set -a && source .env && set +a && \
 	  echo "==> Installing llama-stack..." && \
-	  uv pip install "llama-stack==0.5.0" "llama-stack-api==0.5.0" ollama && \
+	  uv pip install "llama-stack==0.5.0" "llama-stack-api==0.5.0" ollama "pymilvus==2.6.9" "milvus-lite>=2.5.1" "setuptools>=80.9.0,<82.0.0" && \
+	  mkdir -p ../../../milvus_data && \
 	  echo "==> Killing any existing process on port 8321..." && \
 	  lsof -ti:8321 | xargs kill -9 2>/dev/null; true && \
 	  echo "==> Starting Llama Stack server on port 8321..." && \

agents/google/adk/Makefile

@@ -52,15 +52,14 @@ ollama: ## Install Ollama, start it, and pull models
 	  done && \
 	  echo "==> Pulling $(MODEL)..." && \
 	  ollama pull $(MODEL) && \
-	  echo "==> Creating Milvus dir..." && \
-	  mkdir -p ../../../milvus_data && \
 	  echo "" && \
 	  echo "Done. Run 'make llama-server' to start the Llama Stack server."
 
 llama-server: ## Install llama-stack and start Llama Stack server
 	@source .venv/bin/activate && set -a && source .env && set +a && \
 	  echo "==> Installing llama-stack..." && \
-	  uv pip install "llama-stack==0.5.0" "llama-stack-api==0.5.0" ollama && \
+	  uv pip install "llama-stack==0.5.0" "llama-stack-api==0.5.0" ollama "pymilvus==2.6.9" "milvus-lite>=2.5.1" "setuptools>=80.9.0,<82.0.0" && \
+	  mkdir -p ../../../milvus_data && \
 	  echo "==> Killing any existing process on port 8321..." && \
 	  lsof -ti:8321 | xargs kill -9 2>/dev/null; true && \
 	  echo "==> Starting Llama Stack server on port 8321..." && \

agents/langflow/simple_tool_calling_agent/Makefile

@@ -60,7 +60,8 @@ llama-server: ## Install llama-stack and start Llama Stack server
 	fi && \
 	source .venv/bin/activate && \
 	  echo "==> Installing llama-stack..." && \
-	  uv pip install "llama-stack==0.5.0" "llama-stack-api==0.5.0" ollama && \
+	  uv pip install "llama-stack==0.5.0" "llama-stack-api==0.5.0" ollama "pymilvus==2.6.9" "milvus-lite>=2.5.1" "setuptools>=80.9.0,<82.0.0" && \
+	  mkdir -p ../../../milvus_data && \
 	  echo "==> Killing any existing process on port 8321..." && \
 	  lsof -ti:8321 | xargs kill -9 2>/dev/null; true && \
 	  echo "==> Starting Llama Stack server on port 8321..." && \

agents/langgraph/agentic_rag/Makefile

@@ -55,15 +55,14 @@ ollama: ## Install Ollama, start it, and pull models
 	  ollama pull $(MODEL) && \
 	  echo "==> Pulling embedding model ($(OLLAMA_EMBEDDING_MODEL))..." && \
 	  ollama pull $(OLLAMA_EMBEDDING_MODEL) && \
-	  echo "==> Creating Milvus dir..." && \
-	  mkdir -p ../../../milvus_data && \
 	  echo "" && \
 	  echo "Done. Run 'make llama-server' to start the Llama Stack server."
 
 llama-server: ## Install llama-stack and start Llama Stack server
 	@source .venv/bin/activate && set -a && source .env && set +a && \
 	  echo "==> Installing llama-stack..." && \
-	  uv pip install "llama-stack==0.5.0" "llama-stack-api==0.5.0" ollama && \
+	  uv pip install "llama-stack==0.5.0" "llama-stack-api==0.5.0" ollama "pymilvus==2.6.9" "milvus-lite>=2.5.1" "setuptools>=80.9.0,<82.0.0" && \
+	  mkdir -p ../../../milvus_data && \
 	  echo "==> Killing any existing process on port 8321..." && \
 	  lsof -ti:8321 | xargs kill -9 2>/dev/null; true && \
 	  echo "==> Starting Llama Stack server on port 8321..." && \

agents/langgraph/human_in_the_loop/Makefile

@@ -52,15 +52,14 @@ ollama: ## Install Ollama, start it, and pull models
 	  done && \
 	  echo "==> Pulling $(MODEL)..." && \
 	  ollama pull $(MODEL) && \
-	  echo "==> Creating Milvus dir..." && \
-	  mkdir -p ../../../milvus_data && \
 	  echo "" && \
 	  echo "Done. Run 'make llama-server' to start the Llama Stack server."
 
 llama-server: ## Install llama-stack and start Llama Stack server
 	@source .venv/bin/activate && set -a && source .env && set +a && \
 	  echo "==> Installing llama-stack..." && \
-	  uv pip install "llama-stack==0.5.0" "llama-stack-api==0.5.0" ollama && \
+	  uv pip install "llama-stack==0.5.0" "llama-stack-api==0.5.0" ollama "pymilvus==2.6.9" "milvus-lite>=2.5.1" "setuptools>=80.9.0,<82.0.0" && \
+	  mkdir -p ../../../milvus_data && \
 	  echo "==> Killing any existing process on port 8321..." && \
 	  lsof -ti:8321 | xargs kill -9 2>/dev/null; true && \
 	  echo "==> Starting Llama Stack server on port 8321..." && \

agents/langgraph/react_agent/Makefile

@@ -52,15 +52,14 @@ ollama: ## Install Ollama, start it, and pull models
 	  done && \
 	  echo "==> Pulling $(MODEL)..." && \
 	  ollama pull $(MODEL) && \
-	  echo "==> Creating Milvus dir..." && \
-	  mkdir -p ../../../milvus_data && \
 	  echo "" && \
 	  echo "Done. Run 'make llama-server' to start the Llama Stack server."
 
 llama-server: ## Install llama-stack and start Llama Stack server
 	@source .venv/bin/activate && set -a && source .env && set +a && \
 	  echo "==> Installing llama-stack..." && \
-	  uv pip install "llama-stack==0.5.0" "llama-stack-api==0.5.0" ollama && \
+	  uv pip install "llama-stack==0.5.0" "llama-stack-api==0.5.0" ollama "pymilvus==2.6.9" "milvus-lite>=2.5.1" "setuptools>=80.9.0,<82.0.0" && \
+	  mkdir -p ../../../milvus_data && \
 	  echo "==> Killing any existing process on port 8321..." && \
 	  lsof -ti:8321 | xargs kill -9 2>/dev/null; true && \
 	  echo "==> Starting Llama Stack server on port 8321..." && \

agents/langgraph/react_with_database_memory/Makefile

@@ -52,15 +52,14 @@ ollama: ## Install Ollama, start it, and pull models
 	  done && \
 	  echo "==> Pulling $(MODEL)..." && \
 	  ollama pull $(MODEL) && \
-	  echo "==> Creating Milvus dir..." && \
-	  mkdir -p ../../../milvus_data && \
 	  echo "" && \
 	  echo "Done. Run 'make llama-server' to start the Llama Stack server."
 
 llama-server: ## Install llama-stack and start Llama Stack server
 	@source .venv/bin/activate && set -a && source .env && set +a && \
 	  echo "==> Installing llama-stack..." && \
-	  uv pip install "llama-stack==0.5.0" "llama-stack-api==0.5.0" ollama && \
+	  uv pip install "llama-stack==0.5.0" "llama-stack-api==0.5.0" ollama "pymilvus==2.6.9" "milvus-lite>=2.5.1" "setuptools>=80.9.0,<82.0.0" && \
+	  mkdir -p ../../../milvus_data && \
 	  echo "==> Killing any existing process on port 8321..." && \
 	  lsof -ti:8321 | xargs kill -9 2>/dev/null; true && \
 	  echo "==> Starting Llama Stack server on port 8321..." && \

agents/llamaindex/websearch_agent/Makefile

@@ -52,15 +52,14 @@ ollama: ## Install Ollama, start it, and pull models
 	  done && \
 	  echo "==> Pulling $(MODEL)..." && \
 	  ollama pull $(MODEL) && \
-	  echo "==> Creating Milvus dir..." && \
-	  mkdir -p ../../../milvus_data && \
 	  echo "" && \
 	  echo "Done. Run 'make llama-server' to start the Llama Stack server."
 
 llama-server: ## Install llama-stack and start Llama Stack server
 	@source .venv/bin/activate && set -a && source .env && set +a && \
 	  echo "==> Installing llama-stack..." && \
-	  uv pip install "llama-stack==0.5.0" "llama-stack-api==0.5.0" ollama && \
+	  uv pip install "llama-stack==0.5.0" "llama-stack-api==0.5.0" ollama "pymilvus==2.6.9" "milvus-lite>=2.5.1" "setuptools>=80.9.0,<82.0.0" && \
+	  mkdir -p ../../../milvus_data && \
 	  echo "==> Killing any existing process on port 8321..." && \
 	  lsof -ti:8321 | xargs kill -9 2>/dev/null; true && \
 	  echo "==> Starting Llama Stack server on port 8321..." && \