Merge pull request #749 from NVIDIA/am/dynamo-multinode

Update documentation on Dynamo k8s multi node

Author: amaslenn

conf/experimental/ai_dynamo/test/vllm.toml

@@ -19,7 +19,7 @@ description = "vLLM backend with Qwen3-0.6B model"
 test_template_name = "AIDynamo"
 
 [cmd_args]
-docker_image_url = "nvcr.io/nvidia/ai-dynamo/vllm-runtime:0.6.1.post1"
+docker_image_url = "nvcr.io/nvidia/ai-dynamo/vllm-runtime:0.7.0"
 
   [cmd_args.dynamo]
   backend = "vllm"

doc/workloads/ai_dynamo.rst

@@ -16,13 +16,15 @@ Before running the AI Dynamo workload on a Kubernetes cluster, ensure that the c
 .. code-block:: bash
 
    export NAMESPACE=dynamo-system
-   export RELEASE_VERSION=0.6.1  # replace with the desired release version
+   export RELEASE_VERSION=0.7.0  # replace with the desired release version
 
-   helm fetch https://helm.ngc.nvidia.com/nvidia/ai-dynamo/charts/dynamo-crds-${RELEASE_VERSION}.tgz
-   helm install dynamo-crds dynamo-crds-${RELEASE_VERSION}.tgz --namespace default
+   helm upgrade -n default -i dynamo-crds https://helm.ngc.nvidia.com/nvidia/ai-dynamo/charts/dynamo-crds-${RELEASE_VERSION}.tgz
+   helm upgrade -n default -i dynamo-platform https://helm.ngc.nvidia.com/nvidia/ai-dynamo/charts/dynamo-platform-${RELEASE_VERSION}.tgz
 
-   helm fetch https://helm.ngc.nvidia.com/nvidia/ai-dynamo/charts/dynamo-platform-${RELEASE_VERSION}.tgz
-   helm install dynamo-platform dynamo-platform-${RELEASE_VERSION}.tgz --namespace ${NAMESPACE} --create-namespace
+   # The following components are required for multi node only.
+   # Versions should be aligned with Dynamo version.
+   helm upgrade -n default -i grove oci://ghcr.io/ai-dynamo/grove/grove-charts:v0.0.0-gd462e65
+   helm upgrade -n default -i kai-scheduler oci://ghcr.io/nvidia/kai-scheduler/kai-scheduler:0.0.0-4c29820
 
 Launch and Monitor the Job
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~

src/cloudai/systems/kubernetes/kubernetes_system.py

@@ -328,7 +328,9 @@ def _get_frontend_pod_name(self) -> str:
         for pod in self.core_v1.list_namespaced_pod(namespace=self.default_namespace).items:
             labels = pod.metadata.labels
             logging.debug(f"Found pod: {pod.metadata.name} with labels: {labels}")
-            if labels and str(labels.get("nvidia.com/dynamo-component", "")).lower() == "frontend":
+            if labels and str(labels.get("nvidia.com/dynamo-component", "")).lower() == "frontend":  # v0.6.x
+                return pod.metadata.name
+            if labels and str(labels.get("nvidia.com/dynamo-component-type", "")).lower() == "frontend":  # v0.7.x
                 return pod.metadata.name
         raise RuntimeError("No frontend pod found for the job")