refactor: standardize e2e tests across 3 frameworks (#2827)

Signed-off-by: alec-flowers <aflowers@nvidia.com>
Signed-off-by: Indrajit Bhosale <iamindrajitb@gmail.com>

Author: alec-flowers

components/backends/vllm/launch/agg.sh

@@ -5,8 +5,9 @@ set -e
 trap 'echo Cleaning up...; kill 0' EXIT
 
 # run ingress
-python -m dynamo.frontend &
+python -m dynamo.frontend --http-port=8000 &
 
 # run worker
 # --enforce-eager is added for quick deployment. for production use, need to remove this flag
-python -m dynamo.vllm --model Qwen/Qwen3-0.6B --enforce-eager --connector none
+DYN_SYSTEM_ENABLED=true DYN_SYSTEM_PORT=8081 \
+    python -m dynamo.vllm --model Qwen/Qwen3-0.6B --enforce-eager --connector none

components/backends/vllm/launch/agg_lmcache.sh

@@ -5,7 +5,7 @@ set -e
 trap 'echo Cleaning up...; kill 0' EXIT
 
 # run ingress
-python -m dynamo.frontend &
+python -m dynamo.frontend --http-port=8000 &
 
 # run worker with LMCache enabled
 ENABLE_LMCACHE=1 \

components/backends/vllm/launch/agg_router.sh

@@ -5,7 +5,7 @@ set -e
 trap 'echo Cleaning up...; kill 0' EXIT
 
 # run ingress
-python -m dynamo.frontend --router-mode kv &
+python -m dynamo.frontend --router-mode kv --http-port=8000 &
 
 # run workers
 # --enforce-eager is added for quick deployment. for production use, need to remove this flag

components/backends/vllm/launch/dep.sh

@@ -5,7 +5,7 @@ set -e
 trap 'echo Cleaning up...; kill 0' EXIT
 
 # run ingress
-python -m dynamo.frontend --router-mode kv &
+python -m dynamo.frontend --router-mode kv --http-port=8000 &
 
 # Data Parallel Attention / Expert Parallelism
 # Routing to DP workers managed by Dynamo

components/backends/vllm/launch/disagg.sh

@@ -5,7 +5,7 @@ set -e
 trap 'echo Cleaning up...; kill 0' EXIT
 
 # run ingress
-python -m dynamo.frontend --router-mode kv &
+python -m dynamo.frontend --router-mode kv --http-port=8000 &
 
 # --enforce-eager is added for quick deployment. for production use, need to remove this flag
 CUDA_VISIBLE_DEVICES=0 python3 -m dynamo.vllm --model Qwen/Qwen3-0.6B --enforce-eager &

components/backends/vllm/launch/disagg_lmcache.sh

@@ -5,7 +5,7 @@ set -e
 trap 'echo Cleaning up...; kill 0' EXIT
 
 # run ingress with KV router
-python -m dynamo.frontend --router-mode kv &
+python -m dynamo.frontend --router-mode kv --http-port=8000 &
 
 # run decode worker on GPU 0, without enabling LMCache
 CUDA_VISIBLE_DEVICES=0 python3 -m dynamo.vllm --model Qwen/Qwen3-0.6B &

components/backends/vllm/launch/disagg_router.sh

@@ -6,7 +6,7 @@ set -e
 trap 'echo Cleaning up...; kill 0' EXIT
 
 # run ingress
-python -m dynamo.frontend --router-mode kv &
+python -m dynamo.frontend --router-mode kv --http-port=8000 &
 
 # routing will happen between the two decode workers
 # --enforce-eager is added for quick deployment. for production use, need to remove this flag

components/backends/vllm/launch/dsr1_dep.sh

@@ -83,7 +83,7 @@ trap 'echo Cleaning up...; kill 0' EXIT
 
 # run ingress if it's node 0
 if [ $NODE_RANK -eq 0 ]; then
-    DYN_LOG=debug python -m dynamo.frontend --router-mode kv 2>&1 | tee $LOG_DIR/dsr1_dep_ingress.log &
+    DYN_LOG=debug python -m dynamo.frontend --router-mode kv --http-port=8000 2>&1 | tee $LOG_DIR/dsr1_dep_ingress.log &
 fi
 
 mkdir -p $LOG_DIR

examples/multimodal/launch/agg.sh

@@ -53,7 +53,7 @@ else
 fi
 
 # run ingress
-python -m dynamo.frontend &
+python -m dynamo.frontend --http-port=8000 &
 
 # run processor
 python3 components/processor.py --model $MODEL_NAME --prompt-template "$PROMPT_TEMPLATE" &

examples/multimodal/launch/agg_llama.sh

@@ -8,7 +8,7 @@ trap 'echo Cleaning up...; kill 0' EXIT
 MODEL_NAME="meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8"
 
 # run ingress
-python -m dynamo.frontend &
+python -m dynamo.frontend --http-port=8000 &
 
 # run processor
 python3 components/processor.py --model $MODEL_NAME --prompt-template "<|image|>\n<prompt>" &