review - use VLLM_RENDER_URL env variable, placeholder for hf token

Signed-off-by: Maya Barnea <mayab@il.ibm.com>

Author: mayabar

Makefile

@@ -40,6 +40,8 @@ export VLLM_IMAGE ?= $(VLLM_SIMULATOR_TAG_BASE):$(VLLM_SIMULATOR_TAG)
 # CPU-only vLLM image that exposes `vllm launch render` for the token-producer
 # plugin's HTTP backend.
 export VLLM_RENDER_IMAGE ?= vllm/vllm-openai-cpu:v0.21.0
+export VLLM_RENDER_PORT ?= 8082
+export VLLM_RENDER_URL ?= http://vllm-render:$(VLLM_RENDER_PORT)
 
 BUILDER_TAG ?= dev
 BUILDER_TAG_BASE ?= $(IMAGE_REGISTRY)/$(BUILDER_IMAGE_NAME)

deploy/components/vllm-decode/deployment.yaml

@@ -67,7 +67,7 @@ spec:
         - "--port=8200"
         - "--model=${MODEL_NAME}"
         - "--data-parallel-size=${VLLM_DATA_PARALLEL_SIZE}"
-        - "--render-url=http://vllm-render:8082"
+        - "--render-url=${VLLM_RENDER_URL}"
         - "${VLLM_EXTRA_ARGS_D}"
         ports:
         - name: http

deploy/components/vllm-encode/deployment.yaml

@@ -30,7 +30,7 @@ spec:
         - "--data-parallel-size=${VLLM_DATA_PARALLEL_SIZE}"
         - "--no-enable-prefix-caching"
         - "--mm-encoder-only"
-        - "--render-url=http://vllm-render:8082"
+        - "--render-url=${VLLM_RENDER_URL}"
         - "${VLLM_EXTRA_ARGS_E}"
         ports:
         - name: encode-http

deploy/components/vllm-prefill/deployment.yaml

@@ -28,7 +28,7 @@ spec:
         - "--model=${MODEL_NAME}"
         - "--port=8000"
         - "--data-parallel-size=${VLLM_DATA_PARALLEL_SIZE}"
-        - "--render-url=http://vllm-render:8082"
+        - "--render-url=${VLLM_RENDER_URL}"
         - "${VLLM_EXTRA_ARGS_P}"
         ports:
         - name: prefill-http

deploy/environments/dev/README.md

@@ -70,6 +70,8 @@ Variables substituted at deploy time via `envsubst` or Go test `substituteMany`:
 | `VLLM_IMAGE` | vLLM container image (simulator or real) | `ghcr.io/llm-d/llm-d-inference-sim:v0.9.2` |
 | `SIDECAR_IMAGE` | Routing sidecar image | `ghcr.io/llm-d/llm-d-router-disagg-sidecar:dev` |
 | `VLLM_RENDER_IMAGE` | vLLM render sidecar image | `vllm/vllm-openai-cpu:v0.21.0` |
+| `VLLM_RENDER_PORT` | Port the vllm-render Service listens on | `8082` |
+| `VLLM_RENDER_URL` | URL of the vllm-render Service used by decode/encode/prefill pods | `http://vllm-render:${VLLM_RENDER_PORT}` |
 | `MODEL_NAME` | Model name passed to vLLM. Can be a real HuggingFace model (e.g. `TinyLlama/TinyLlama-1.1B-Chat-v1.0`, `Qwen/Qwen3-VL-2B-Instruct`) or an arbitrary name when using the simulator (e.g. `food-review`) | `food-review` |
 | `POOL_NAME` | InferencePool name | `food-review-inference-pool` |
 | `VLLM_REPLICA_COUNT_E` | Encode deployment replicas | `1` |

deploy/environments/dev/base-kind-istio/vllm-render.yaml

@@ -19,36 +19,43 @@ spec:
         image: ${VLLM_RENDER_IMAGE}
         imagePullPolicy: IfNotPresent
         command: ["vllm", "launch", "render"]
-        args: ["${MODEL_NAME}", "--port=8082"]
+        args: ["${MODEL_NAME}", "--port=${VLLM_RENDER_PORT}"]
+        # For gated models, inject HF_TOKEN via a Secret reference:
+        # env:
+        # - name: HF_TOKEN
+        #   valueFrom:
+        #     secretKeyRef:
+        #       name: hf-token
+        #       key: token
         ports:
         - name: render-http
-          containerPort: 8082
+          containerPort: ${VLLM_RENDER_PORT}
         startupProbe:
           httpGet:
             path: /health
-            port: 8082
+            port: ${VLLM_RENDER_PORT}
           initialDelaySeconds: 2
           periodSeconds: 1
           failureThreshold: 600
           timeoutSeconds: 1
         livenessProbe:
           httpGet:
             path: /health
-            port: 8082
+            port: ${VLLM_RENDER_PORT}
           periodSeconds: 10
           failureThreshold: 5
           timeoutSeconds: 1
         readinessProbe:
           httpGet:
             path: /health
-            port: 8082
+            port: ${VLLM_RENDER_PORT}
           periodSeconds: 5
           failureThreshold: 1
           timeoutSeconds: 1
         volumeMounts:
         - name: model-cache
           mountPath: /root/.cache/huggingface
-      volumes: 
+      volumes:
       - name: model-cache
         emptyDir: {}
 ---
@@ -64,6 +71,6 @@ spec:
   ports:
   - name: http
     protocol: TCP
-    port: 8082
-    targetPort: 8082
+    port: ${VLLM_RENDER_PORT}
+    targetPort: ${VLLM_RENDER_PORT}
   type: ClusterIP

deploy/environments/dev/e2e-infra/vllm-render.yaml

@@ -20,35 +20,34 @@ spec:
         imagePullPolicy: IfNotPresent
         args:
         - "--model=${MODEL_NAME}"
-        - "--port=8082"
+        - "--port=${VLLM_RENDER_PORT}"
         - "--mode=echo"
         - "--force-dummy-tokenizer"
         ports:
         - name: http
-          containerPort: 8082
+          containerPort: ${VLLM_RENDER_PORT}
         startupProbe:
           httpGet:
             path: /health
-            port: 8082
+            port: ${VLLM_RENDER_PORT}
           initialDelaySeconds: 2
           periodSeconds: 1
           failureThreshold: 600
           timeoutSeconds: 1
         livenessProbe:
           httpGet:
             path: /health
-            port: 8082
+            port: ${VLLM_RENDER_PORT}
           periodSeconds: 10
           failureThreshold: 5
           timeoutSeconds: 1
         readinessProbe:
           httpGet:
             path: /health
-            port: 8082
+            port: ${VLLM_RENDER_PORT}
           periodSeconds: 5
           failureThreshold: 1
           timeoutSeconds: 1
-      volumes: []
 ---
 apiVersion: v1
 kind: Service
@@ -62,6 +61,6 @@ spec:
   ports:
   - name: http
     protocol: TCP
-    port: 8082
-    targetPort: 8082
+    port: ${VLLM_RENDER_PORT}
+    targetPort: ${VLLM_RENDER_PORT}
   type: ClusterIP

deploy/environments/dev/epd/patch-decode.yaml

@@ -22,7 +22,7 @@ spec:
         - "--block-size=16"
         - "--data-parallel-size=${VLLM_DATA_PARALLEL_SIZE}"
         - "--zmq-endpoint=tcp://${EPP_NAME}.${NAMESPACE}.svc.cluster.local:5557"
-        - "--render-url=http://vllm-render:8082"
+        - "--render-url=${VLLM_RENDER_URL}"
         - "${VLLM_EXTRA_ARGS_D}"
         ports:
         - $patch: replace

scripts/kind-dev-env.sh

@@ -65,6 +65,8 @@ export SIDECAR_IMAGE
 # Set a default VLLM_RENDER_IMAGE if not provided (CPU-only vLLM image that
 # runs `vllm launch render` for the token-producer plugin's HTTP backend).
 export VLLM_RENDER_IMAGE="${VLLM_RENDER_IMAGE:-vllm/vllm-openai-cpu:v0.21.0}"
+export VLLM_RENDER_PORT="${VLLM_RENDER_PORT:-8082}"
+export VLLM_RENDER_URL="${VLLM_RENDER_URL:-http://vllm-render:${VLLM_RENDER_PORT}}"
 
 # Set the inference pool name for the deployment
 export POOL_NAME="${POOL_NAME:-${MODEL_NAME_SAFE}-inference-pool}"
@@ -385,14 +387,14 @@ kubectl --context ${KUBE_CONTEXT} create configmap epp-config --from-file=epp-co
 # Deploy Istio base (shared infrastructure)
 kubectl kustomize --enable-helm deploy/environments/dev/base-kind-istio \
   | envsubst '${POOL_NAME} ${MODEL_NAME} ${MODEL_NAME_SAFE} ${EPP_NAME} ${EPP_IMAGE} ${VLLM_IMAGE} \
-  ${SIDECAR_IMAGE} ${VLLM_RENDER_IMAGE} ${TARGET_PORTS} ${NAMESPACE} ${METRICS_ENDPOINT_AUTH} \
+  ${SIDECAR_IMAGE} ${VLLM_RENDER_IMAGE} ${VLLM_RENDER_PORT} ${VLLM_RENDER_URL} ${TARGET_PORTS} ${NAMESPACE} ${METRICS_ENDPOINT_AUTH} \
 ${VLLM_REPLICA_COUNT_E} ${VLLM_REPLICA_COUNT_P} ${VLLM_REPLICA_COUNT_D} ${VLLM_DATA_PARALLEL_SIZE}' \
   | kubectl --context ${KUBE_CONTEXT} apply -f -
 
 # Deploy scenario-specific vLLM components
 kubectl kustomize --enable-helm ${KUSTOMIZE_DIR} \
   | envsubst '${POOL_NAME} ${MODEL_NAME} ${MODEL_NAME_SAFE} ${EPP_NAME} ${EPP_IMAGE} ${VLLM_IMAGE} \
-  ${SIDECAR_IMAGE} ${VLLM_RENDER_IMAGE} ${TARGET_PORTS} ${NAMESPACE} \
+  ${SIDECAR_IMAGE} ${VLLM_RENDER_IMAGE} ${VLLM_RENDER_PORT} ${VLLM_RENDER_URL} ${TARGET_PORTS} ${NAMESPACE} \
   ${VLLM_REPLICA_COUNT_E} ${VLLM_REPLICA_COUNT_P} ${VLLM_REPLICA_COUNT_D} ${VLLM_DATA_PARALLEL_SIZE} \
   ${KV_CONNECTOR_TYPE} ${EC_CONNECTOR_TYPE} ${CONNECTOR_TYPE} ${KV_CACHE_ENABLED} ${HF_TOKEN} ${VLLM_SIM_MODE} \
   ${DECODE_ROLE} ${VLLM_EXTRA_ARGS_E} ${VLLM_EXTRA_ARGS_P} ${VLLM_EXTRA_ARGS_D}' \

test/e2e/e2e_suite_test.go

@@ -60,6 +60,9 @@ const (
 	metricsTestLabel       = "Metrics"
 	deprecatedPDTestLabel  = "DeprecatedPD"
 	disaggTestLabel        = "Disagg"
+
+	// images
+	simulatorImg = "ghcr.io/llm-d/llm-d-inference-sim:v0.9.2"
 )
 
 var (
@@ -74,9 +77,10 @@ var (
 
 	containerRuntime = env.GetEnvString("CONTAINER_RUNTIME", "docker", ginkgo.GinkgoLogr)
 	eppImage         = env.GetEnvString("EPP_IMAGE", "ghcr.io/llm-d/llm-d-router-endpoint-picker:dev", ginkgo.GinkgoLogr)
-	vllmSimImage     = env.GetEnvString("VLLM_IMAGE", "ghcr.io/llm-d/llm-d-inference-sim:v0.9.2", ginkgo.GinkgoLogr)
+	vllmSimImage     = env.GetEnvString("VLLM_IMAGE", simulatorImg, ginkgo.GinkgoLogr)
 	sideCarImage     = env.GetEnvString("SIDECAR_IMAGE", "ghcr.io/llm-d/llm-d-router-disagg-sidecar:dev", ginkgo.GinkgoLogr)
-	vllmRenderImage  = env.GetEnvString("VLLM_RENDER_IMAGE", "ghcr.io/llm-d/llm-d-inference-sim:v0.9.2", ginkgo.GinkgoLogr)
+	vllmRenderImage  = env.GetEnvString("VLLM_RENDER_IMAGE", simulatorImg, ginkgo.GinkgoLogr)
+	vllmRenderPort   = env.GetEnvString("VLLM_RENDER_PORT", "8082", ginkgo.GinkgoLogr)
 	loadRenderImage  = env.GetEnvBool("LOAD_VLLM_RENDER_IMAGE", true, ginkgo.GinkgoLogr)
 	// nsName is the namespace in which the K8S objects will be created
 	nsName = env.GetEnvString("NAMESPACE", "default", ginkgo.GinkgoLogr)