Practice for the Red Hat Certified Specialist in OpenShift AI (EX267) exam with 18 exam-style practice questions, instant answer reveals, and concise explanations of every correct answer. Topics include: What is the function of the serving.kserve.io/v1beta1 API group in 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 is the function of the serving.kserve.io/v1beta1 API group in OpenShift AI?
- How do you update the resource limits (CPU/Memory) of an existing OpenShift AI workbench?
- Which cloud-storage API standard is used when creating a Data Connection in OpenShift AI?
- Where are pipeline artifacts, such as output datasets or metadata, stored by default?
- Which framework can be explicitly selected in OpenShift AI to host Large Language Models (LLMs) efficiently?
- Which object handles cluster routing and exposes the OpenShift AI dashboard outside the cluster?
Answers and explanations
Tap a question to expand the answer and the exam reasoning. Try to commit to your own pick first.
Q1. What is the function of the serving.kserve.io/v1beta1 API group in OpenShift AI?
Answer: D. It defines the schema and management endpoints for deploying InferenceServices via KServe
The KServe API group defines the schema and endpoints for deploying model servers as InferenceServices. It manages model inference deployments rather than authentication, storage, or networking rules.
Q2. How do you update the resource limits (CPU/Memory) of an existing OpenShift AI workbench?
Answer: A. Stop the workbench, edit its configuration in the dashboard, and restart it
You stop the workbench, adjust resource limits in the dashboard, and restart it. Directly modifying the deployment manually is discouraged because the dashboard manages the underlying resources.
Q3. Which cloud-storage API standard is used when creating a Data Connection in OpenShift AI?
Answer: B. AWS S3 API
Data Connections in OpenShift AI use the Amazon S3 API standard for object storage integration. This mechanism securely injects endpoint, bucket, and credential references into your workbenches or pipelines as standard Kubernetes secrets.
Q4. Where are pipeline artifacts, such as output datasets or metadata, stored by default?
Answer: D. In the configured S3-compatible Object Storage data connection
Pipeline artifacts and outputs are stored in the S3-compatible object storage bucket configured via your data connection. Ephemeral workbench storage is strictly local and temporary, so relying on it risks losing pipeline outputs when a pod restarts.
Q5. Which framework can be explicitly selected in OpenShift AI to host Large Language Models (LLMs) efficiently?
Answer: C. vLLM
vLLM is specifically integrated into OpenShift AI as a highly optimized runtime for efficient large language model serving. Standard Jupyter images and base PyTorch runtimes lack the specialized memory management and inference optimizations required for production LLM hosting.
Q6. Which object handles cluster routing and exposes the OpenShift AI dashboard outside the cluster?
Answer: D. Route
OpenShift uses Route objects to expose services like the AI dashboard externally, providing hostname and TLS termination. While Services handle internal traffic, Routes are the specific OpenShift abstraction for external ingress.
Q7. In OpenShift AI, where are the execution artifacts (such as trained models or datasets) from a pipeline run physically stored?
Answer: D. In the configured object storage bucket via the data connection
Pipeline execution artifacts are written to the object storage bucket configured through the pipeline data connection. This ensures persistent tracking and prevents data loss, since ephemeral container storage vanishes after completion.
Q8. What capability does the Node Feature Discovery (NFD) Operator provide to the OpenShift AI scheduling system?
Answer: A. It discovers and labels specific hardware capabilities on worker nodes, such as GPUs or AVX instructions
The Node Feature Discovery Operator tags worker nodes with specific hardware capabilities like GPUs, enabling targeted workload scheduling. Without these labels, OpenShift AI cannot automatically pin machine learning tasks to specialized hardware.
Q9. Which custom resource (CR) is modified to initialize and configure Red Hat OpenShift AI?
Answer: C. DataScienceCluster
The DataScienceCluster custom resource initializes the operator and manages all OpenShift AI components. Modifying this resource toggles features on or off, ensuring the cluster aligns with your desired configuration.
Q10. What type of storage is recommended for hosting a Multi-Model Serving platform model repository in OpenShift AI?
Answer: A. S3-Compatible Object Storage
S3-compatible object storage is the standard for hosting model repositories because it efficiently handles large unstructured data. Block storage or HostPath volumes are inappropriate for scalable, distributed model access.
Q11. Which command can a developer use to interactively log into an OpenShift cluster from their workbench terminal environment?
Answer: A. oc login
The oc login command authenticates users and establishes a session with the OpenShift API server from any terminal. Standard kubectl commands lack the built-in authentication shortcuts provided by the OpenShift CLI.
Q12. What custom resource (CR) is modified by an administrator to enable or disable specific components like dashboards or training operators in RHOAI?
Answer: C. DataScienceCluster
Administrators modify the DataScienceCluster custom resource to enable or disable specific OpenShift AI components like the dashboard or model serving. The DSCInitialization resource only handles bootstrap configurations and should not be edited directly.
Q13. What field is required in an S3 Data Connection to specify the regional endpoint location of a self-hosted MinIO object storage instance?
Answer: A. AWS_S3_ENDPOINT
The AWS_S3_ENDPOINT field specifies the URL for self-hosted object storage like MinIO, overriding default AWS endpoints. Expect to map this key when creating a workbench data connection to a custom S3 provider.
Q14. What is the purpose of configuring "Tolerations" on a workbench or model server deployment within OpenShift AI?
Answer: B. To permit pods to schedule on dedicated nodes with specific taints, such as GPU nodes
Tolerations allow pods to schedule onto tainted nodes reserved for specialized hardware. If your workbench cannot access GPU nodes, check whether the deployment includes a matching toleration for the hardware taint.
Q15. What is the purpose of a PVC in the context of an OpenShift AI workbench?
Answer: C. It provides persistent file system storage for notebooks and code files across pod restarts
A PersistentVolumeClaim mounts durable storage so notebook files survive pod restarts. Without a PVC, any code or datasets saved inside the container are lost when the workbench stops.
Q16. Which operator must be present in the OpenShift cluster to enable serverless scaling behaviors for single-model deployment?
Answer: A. Red Hat OpenShift Serverless Operator
The Red Hat OpenShift Serverless Operator provides Knative Serving, which handles scale-to-zero functionality for model deployments. You must install it before enabling KServe single-model serving platforms.
Q17. What is the primary role of the ServiceMeshControlPlane resource in relation to OpenShift AI single-model serving?
Answer: C. It manages ingress routing, mutual TLS, and network policies for KServe
The ServiceMeshControlPlane configures the Service Mesh components required by KServe for routing and security. KServe relies on this mesh to properly manage external ingress and mutual TLS.
Q18. What storage access type is required for multi-user read-write data sharing across workbenches?
Answer: B. ReadWriteMany (RWX) Persistent Volume storage solutions
ReadWriteMany access mode allows multiple pods to mount and write to the volume simultaneously. You must select this storage type when configuring shared datasets across multiple workbenches.
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.