Chapter 8. Validate the Openshift Container Platform cluster

After deploying Distributed Inference with llm-d on Openshift Container Platform, verify that all Operators and components were installed correctly.

Prerequisites

You have the OpenShift CLI (oc) configured with cluster administrator privileges.
You have completed the deployment procedure.

Procedure

Verify that the RHOAI Operator CSV is installed and shows a status of Succeeded:
```
$ oc get csv -A | grep rhods
```
Expected output shows the CSV in a Succeeded phase.

Verify that the DataScienceCluster is ready:

$ oc get datasciencecluster default-dsc -o jsonpath='{.status.phase}'

Expected output: Ready

Verify that the LLMInferenceService CRD exists:
```
$ oc get crd | grep llm
```
Expected output includes llminferenceservices.serving.kserve.io.
Verify that the RHOAI Operator pod is running:
```
$ oc get pods -n redhat-ods-operator
```
Verify that the KServe LLMISvc controller is running:
```
$ oc get pods -n redhat-ods-applications
```
Verify that all managed dependency pods are running:
```
$ oc get pods -n cert-manager
```
```
$ oc get pods -n istio-system
```
All pods should be in a Running state.

Verify GPU nodes are available and GPUs are allocatable:

$ oc get nodes -l nvidia.com/gpu.present=true
$ oc describe nodes | grep -A 5 "nvidia.com/gpu"

Verification

Confirm that all checks pass:

Expand

Table 8.1. Openshift Container Platform validation checkpoints
Checkpoint	Command	Expected
Operator installed	`oc get csv -A \| grep rhods`	`Succeeded`
DSC ready	`oc get datasciencecluster default-dsc -o jsonpath='\{.status.phase}'`	`Ready`
LLMISVC CRD exists	`oc get crd \| grep llm`	`llminferenceservices.serving.kserve.io`
Operator pods running	`oc get pods -n redhat-ods-operator`	All `Running`
KServe pods running	`oc get pods -n redhat-ods-applications`	All `Running`
GPU nodes available	`oc get nodes -l nvidia.com/gpu.present=true`	Nodes listed