chore: remove explicit --vllm-port=8200 from sidecar configs

- llm-d-router disagg sidecar now defaults --vllm-port to 8200
(llm-d/llm-d-router#1723), making the explicit flag redundant

Signed-off-by: Wen Zhou <wenzhou@redhat.com>

Author: zdtsw

guides/agentic-serving/modelserver/gpu/vllm/nemotron-3-ultra/patch-decode.yaml

@@ -11,7 +11,6 @@ spec:
           image: ghcr.io/llm-d/llm-d-router-disagg-sidecar:main
           args:
             - --port=8000
-            - --vllm-port=8200
             - --kv-connector=nixlv2
             - --zap-log-level=4
             - --secure-proxy=false

guides/multimodal-serving/e-disaggregation/modelserver/gpu/vllm/e-p-d/base/patch-decode.yaml

@@ -10,7 +10,6 @@ spec:
         - name: routing-proxy
           args:
             - --port=8000
-            - --vllm-port=8200
             - --kv-connector=nixlv2
             - --ec-connector=ec-nixl
             - --zap-log-level=1

guides/multimodal-serving/e-disaggregation/modelserver/gpu/vllm/e-pd/base/patch-decode.yaml

@@ -11,7 +11,6 @@ spec:
           imagePullPolicy: Always
           args:
             - --port=8000
-            - --vllm-port=8200
             - --kv-connector=nixlv2
             - --ec-connector=ec-nixl
             - --zap-log-level=1

guides/recipes/modelserver/base/single-host/pd/sglang/patch-sidecar.yaml

@@ -9,7 +9,6 @@ spec:
         - name: routing-proxy
           args:
             - --port=8000
-            - --vllm-port=8200
             - --kv-connector=sglang
             - --zap-log-level=1
             - --secure-proxy=false

guides/recipes/modelserver/base/single-host/pd/vllm/patch-sidecar.yaml

@@ -9,7 +9,6 @@ spec:
         - name: routing-proxy
           args:
             - --port=8000
-            - --vllm-port=8200
             - --kv-connector=nixlv2
             - --zap-log-level=1
             - --secure-proxy=false

guides/wide-ep-lws/modelserver/gpu/vllm-deepseek-v4/base/decode.yaml

@@ -24,7 +24,6 @@ spec:
           - name: routing-proxy
             args:
               - --port=8000
-              - --vllm-port=8200
               - --kv-connector=nixlv2
               - --zap-log-level=1
               - --secure-proxy=false

guides/wide-ep-lws/modelserver/gpu/vllm/base/decode.yaml

@@ -27,11 +27,10 @@ spec:
         initContainers:
           - name: routing-proxy
             args:
-              # sidecar proxy is started with --port 8000, --vllm-port 8200, and --data-parallel-size 8,
+              # sidecar proxy is started with --port 8000 and --data-parallel-size 8,
               # so it listens on ports in the range [port:port+data-parallel-size-1] and forwards
               # requests each to the corresponding vllm port.
               - --port=8000
-              - --vllm-port=8200
               - --data-parallel-size=8
               - --kv-connector=nixlv2
               - --zap-log-level=1

guides/wide-ep-lws/modelserver/xpu/vllm/decode.yaml

@@ -31,7 +31,6 @@ spec:
           - name: routing-proxy
             args:
               - --port=8000
-              - --vllm-port=8200
               - --kv-connector=nixlv2
               - --zap-log-level=1
               - --secure-proxy=false