Practice for the Red Hat Certified Specialist in OpenShift AI (EX267) exam with 17 exam-style practice questions, instant answer reveals, and concise explanations of every correct answer. Topics include: What configuration element assigns specific user groups administrative rights over OpenShift AI?. Follow @CertPunch and visit certpunch.com for more certification practice exams and study content.
Prefer hands-on? Take this round as an interactive practice test — answer every question, get instant feedback, and see your score: Start the Red Hat Certified Specialist in OpenShift AI (EX267) practice test →
What you will practice
- What configuration element assigns specific user groups administrative rights over OpenShift AI?
- Which node architecture attribute assures data science workbenches are scheduled exclusively on GPU-enabled w…
- Which parameter indicates the unique ID of an individual execution run in pipelines?
- What constitutes the default pipeline artifact store for completed pipeline executions in Red Hat OpenShift A…
- Which CRD registers a reusable serving infrastructure runtime envelope type capable of loading deserialized m…
- Which network resource exposes model serving endpoints securely to applications outside OpenShift?
Answers and explanations
Tap a question to expand the answer and the exam reasoning. Try to commit to your own pick first.
Q1. What configuration element assigns specific user groups administrative rights over OpenShift AI?
Answer: A. OdhDashboardConfig administration groups properties settings definitions
You configure administrative groups using the OdhDashboardConfig custom resource within the dashboard settings. ClusterRoleBinding grants cluster-admin, which is far too broad for OpenShift AI specific management.
Q2. Which node architecture attribute assures data science workbenches are scheduled exclusively on GPU-enabled worker nodes?
Answer: D. Taints and Tolerations
Taints and tolerations ensure that workbenches requiring GPUs are exclusively scheduled on GPU-enabled nodes. Topology spread constraints distribute pods, but they do not guarantee exclusive hardware scheduling.
Q3. Which parameter indicates the unique ID of an individual execution run in pipelines?
Answer: B. Run ID UUID tracking string parameter metric property values
The Run ID is the unique identifier used to track individual pipeline executions in the OpenShift AI dashboard. The other options represent lower-level infrastructure details rather than pipeline execution metrics.
Q4. What constitutes the default pipeline artifact store for completed pipeline executions in Red Hat OpenShift AI?
Answer: D. An S3-compatible Object Storage Bucket
Pipeline artifacts are stored in S3-compatible object storage buckets to ensure scalable persistence. Ephemeral pod storage fails the requirement for long-term persistence across multiple pipeline executions.
Q5. Which CRD registers a reusable serving infrastructure runtime envelope type capable of loading deserialized model binaries?
Answer: B. ServingRuntime CRD definitions manifest
The ServingRuntime custom resource defines the environment and container setup required to load and serve specific model frameworks. The InferencePredictor merely references these predefined runtimes during application deployment.
Q6. Which network resource exposes model serving endpoints securely to applications outside OpenShift?
Answer: B. Route or Ingress mapping resources rules definitions
OpenShift Routes securely expose model serving endpoints to external applications via HTTP and HTTPS. Internal ClusterIP services keep traffic confined within the cluster, preventing external client access entirely.
Q7. Which parameter in the DataScienceCluster CR determines whether a component like KServe is fully managed or disabled?
Answer: C. managementState
The managementState parameter dictates whether a specific component is actively managed or removed by the operator. Generic status fields reflect current conditions but cannot trigger operational state changes directly.
Q8. What validation step ensures a Data Connection can successfully connect to an object bucket?
Answer: C. The OpenShift AI dashboard validates credentials upon Data Connection addition
Adding the Data Connection directly in the dashboard triggers an immediate validation of the provided credentials. Options suggesting background operator checks or workbench scripts fail because they do not validate the configuration prior to resource creation.
Q9. What mechanism does OpenShift AI use to allocate specific GPU resources to a user workbench container?
Answer: A. Kubernetes resource limits and requests specifying nvidia.com
OpenShift AI relies on standard Kubernetes resource requests and limits to allocate hardware accelerators. For GPUs, you must specify the nvidia.com/gpu resource; hostPath mounts are restricted by default security policies.
Q10. Which core operator supports the serverless traffic layer driving independent model endpoint pod autoscaling?
Answer: C. OpenShift Serverless Operator
The OpenShift Serverless Operator provides Knative Serving, which handles automatic scaling and network routing for deployed models. Do not confuse this with the base AI operator, which only deploys the dashboard and components.
Q11. What CRD controls the user privileges and capabilities inside an isolated workbench container?
Answer: C. SecurityContextConstraints
Security Context Constraints govern the security isolation and Linux capabilities granted to pods. The fake options like UserAccessControl are designed to mimic older Microsoft acronyms but do not exist in Kubernetes.
Q12. When creating an object storage connection for a data science project, which information is a primary requirement?
Answer: D. S3 bucket URL and access/secret keys
An S3 bucket URL paired with valid access and secret keys is required to authenticate and mount object storage. Git tokens are strictly for version control integration and cannot establish generic data connections.
Q13. Which CLI command provides the real-time logs of deployed inference pod instances?
Answer: D. oc logs -l component=predictor -n [serving-namespace]
The oc logs command combined with a label selector filters the exact runtime logs for inference pods. The component predictor label specifically targets the underlying runtime generated by the model server.
Q14. What must be installed on the OpenShift cluster before the Red Hat OpenShift AI Operator can utilize NVIDIA GPUs?
Answer: B. NVIDIA GPU Operator
The NVIDIA GPU Operator must be installed first to automate the management of NVIDIA GPUs. On the exam, remember that OpenShift AI relies on this operator for hardware acceleration rather than using built-in drivers.
Q15. Which CRD deploys multi-model serving environments optimized for dense packing in Red Hat OpenShift AI?
Answer: A. ModelMesh
ModelMesh is the custom resource designed for dense packing and multi-model serving. KServe is typically used for single-model serving, making it the incorrect choice for maximizing resource efficiency.
Q16. WHICH COMPONENT IN OPENSHIFT AI IS PRIMARILY USED BY DATA SCIENTISTS TO SPAWN JUPYTERLAB ENVIRONMENTS?
Answer: A. Data Science Workbenches
Data Science Workbenches provide the interface for data scientists to spawn JupyterLab environments. Data Science Projects serve as organizational namespaces rather than interactive computing environments.
Q17. WHAT IS THE CORRECT CUSTOM RESOURCE TYPE USED TO CONFIGURE GLOBAL SETTINGS FOR THE OPENSHIFT AI DASHBOARD COMPONENTS?
Answer: B. DSCInitialization
The DSCInitialization custom resource is used to apply global configuration settings required by the dashboard components. DashboardConfig is deprecated and no longer the primary resource for global initialization.
More Red Hat Certified Specialist in OpenShift AI (EX267) drills and other practice exams are on @CertPunch. New rounds drop every few days at certpunch.com.