disagg.yaml

# SPDX-FileCopyrightText: Copyright (c) 2025-2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
# SPDX-License-Identifier: Apache-2.0

apiVersion: nvidia.com/v1beta1
kind: DynamoGraphDeployment
metadata:
  name: vllm-disagg
spec:
  components:
  - name: Frontend
    podTemplate:
      spec:
        containers:
        - image: my-registry/vllm-runtime:my-tag
          name: main
    replicas: 1
    type: frontend
  - name: VllmDecodeWorker
    podTemplate:
      spec:
        containers:
        - args:
          - |
            export LD_LIBRARY_PATH=/usr/local/nvidia/lib64:$LD_LIBRARY_PATH
            export PATH=$PATH:/usr/local/nvidia/bin:/usr/local/nvidia/lib64
            /sbin/ldconfig
            python3 -m dynamo.vllm --model Qwen/Qwen3-0.6B --disaggregation-mode decode
          command:
          - /bin/sh
          - -c
          envFrom:
          - secretRef:
              name: hf-token-secret
          image: my-registry/vllm-runtime:my-tag
          name: main
          resources:
            limits:
              nvidia.com/gpu: "1"
          startupProbe:
            initialDelaySeconds: 180
          workingDir: /workspace/examples/backends/vllm
    replicas: 1
    type: decode
  - name: VllmPrefillWorker
    podTemplate:
      spec:
        containers:
        - args:
          - |
            export LD_LIBRARY_PATH=/usr/local/nvidia/lib64:$LD_LIBRARY_PATH
            export PATH=$PATH:/usr/local/nvidia/bin:/usr/local/nvidia/lib64
            /sbin/ldconfig
            python3 -m dynamo.vllm --model Qwen/Qwen3-0.6B --disaggregation-mode prefill
          command:
          - /bin/sh
          - -c
          envFrom:
          - secretRef:
              name: hf-token-secret
          image: my-registry/vllm-runtime:my-tag
          name: main
          resources:
            limits:
              nvidia.com/gpu: "1"
          workingDir: /workspace/examples/backends/vllm
    replicas: 1
    type: prefill