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: WHICH API GROUP IS SPECIFIC TO THE DATA SCIENCE CLUSTER CONFIGURATION 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
- WHICH API GROUP IS SPECIFIC TO THE DATA SCIENCE CLUSTER CONFIGURATION IN OPENSHIFT AI?
- WHICH RUNTIME PERMITS TENSORFLOW AND PYTORCH MODELS TO RUN ON OPTIMIZED HARDWARE VIA A SINGLE SERVER PLATFORM…
- WHICH BASE COMPONENT DEPLOYS THE SINGLE-MODEL INFERENCE WORKLOADS IN OPENSHIFT AI ADHERING TO SERVERLESS PATT…
- What is the primary role of the ServiceMesh component when installing OpenShift AI with advanced model servin…
- WHICH STORAGE COMPONENT MUST BE DEFINED TO RECOVER A WORKBENCH STATE LOCALLY IF A WORKER NODE FAILS?
- How can an administrator temporarily prevent users from launching new workbenches?
Answers and explanations
Tap a question to expand the answer and the exam reasoning. Try to commit to your own pick first.
Q1. WHICH API GROUP IS SPECIFIC TO THE DATA SCIENCE CLUSTER CONFIGURATION IN OPENSHIFT AI?
Answer: A. datasciencecluster.opendatahub.io
The DataScienceCluster custom resource belongs to the datasciencecluster.opendatahub.io API group in OpenShift AI. This component manages the underlying operators and services for the entire data science platform.
Q2. WHICH RUNTIME PERMITS TENSORFLOW AND PYTORCH MODELS TO RUN ON OPTIMIZED HARDWARE VIA A SINGLE SERVER PLATFORM INFRASTRUCTURE?
Answer: B. Triton Inference Server Runtime
The Triton Inference Server allows both TensorFlow and PyTorch models to run on optimized hardware within a single platform. The other runtimes are largely tailored to specific frameworks rather than unified multi-framework serving.
Q3. WHICH BASE COMPONENT DEPLOYS THE SINGLE-MODEL INFERENCE WORKLOADS IN OPENSHIFT AI ADHERING TO SERVERLESS PATTERNS?
Answer: B. KServe
KServe is the component that deploys single-model inference workloads adhering to serverless patterns. For the exam, remember that ModelMesh is used for scalable multi-model serving, whereas KServe handles serverless deployments.
Q4. What is the primary role of the ServiceMesh component when installing OpenShift AI with advanced model serving?
Answer: A. It provides ingress routing, mTLS, and traffic management for models.
Service Mesh handles ingress routing, mTLS, and traffic management crucial for KServe model serving. It does not schedule GPU pods or handle data ingestion, which are the jobs of Kubernetes schedulers and pipelines.
Q5. WHICH STORAGE COMPONENT MUST BE DEFINED TO RECOVER A WORKBENCH STATE LOCALLY IF A WORKER NODE FAILS?
Answer: C. Persistent Volume Claim (PVC)
A Persistent Volume Claim ensures workbench state is stored on networked storage, allowing recovery if a worker node fails. HostPath and EmptyDir volumes are tied to local node storage, meaning data is permanently lost upon node failure.
Q6. How can an administrator temporarily prevent users from launching new workbenches?
Answer: D. Scale down the notebook controller deployment or modify dashboard permissions
Scaling down the notebook controller deployment stops the system from reconciling and launching new workbenches. Destroying storage or namespaces is destructive, whereas scaling the controller is a clean administrative toggle.
Q7. Which base workbench container tool can compile Python runtimes into replicable Tekton task artifacts internally?
Answer: A. Elyra Pipeline Editor tool within JupyterLab
The Elyra Pipeline Editor within JupyterLab allows users to visually build pipelines and export them as Tekton resources. It does not compile Python runtimes, but rather packages notebooks and scripts into pipeline artifacts.
Q8. Which inference format enables heterogeneous backend hardware runtimes to interpret models uniformly regardless of training engine?
Answer: D. ONNX (Open Neural Network Exchange)
ONNX provides a standard model representation that allows various runtime engines and hardware accelerators to execute the model consistently. Pickle files or raw Python scripts are tightly coupled to their specific training environments and lack hardware interoperability.
Q9. Which inference service platform in OpenShift AI is preferred for multi-model predictions where lower resource footprints are critical?
Answer: C. ModelMesh Serving Platform
ModelMesh is designed specifically for high-scale, multi-model serving with efficient memory and compute footprint management. KServe is typically favored for single-model deployments requiring dedicated GPUs rather than optimizing shared resource pools.
Q10. Which object keeps S3 secure account keys encrypted within the cluster backend preserving compliance assurances?
Answer: B. Secret
Kubernetes Secrets are designed to store sensitive information like S3 credentials and can be encrypted at rest on the backend. ConfigMaps are plaintext and should never be used for secure account keys or authentication tokens.
Q11. To update the global cluster-wide proxy configuration for OpenShift AI components, which resource do administrators modify?
Answer: D. The proxy configuration in the DSCInitialization or cluster Proxy CR
Administrators modify the proxy configuration in the DSCInitialization custom resource or the cluster Proxy CR to update global proxy settings. Manually editing ConfigMaps is inefficient because the operator will eventually overwrite your changes.
Q12. WHICH API GROUP CONTROLS DATA SCIENCE CLUSTER CONFIGURATIONS IN AN OPENSHIFT ENVIRONMENT RUNNING RED HAT OPENSHIFT AI OPERATORS?
Answer: D. datasciencecluster.opendatahub.io
The datasciencecluster.opendatahub.io API group controls cluster configurations in an environment running Red Hat OpenShift AI operators. Kubeflow resources belong to a different ecosystem and do not manage the OpenShift AI control plane.
Q13. Which service account permission is required for KServe to pull models securely from a private S3-compatible object storage bucket?
Answer: A. A Secret linked to the service account or annotated for KServe
The correct option works because KServe requires explicit credentials to access private object storage. Linking a Kubernetes Secret to the service account injects these necessary keys. Privileged cluster roles do not grant S3 bucket access, making them invalid distractors.
Q14. WHICH WORKBENCH IMAGE OPTION COMES WITH PRE-INSTALLED PACKAGES TAILORED FOR OBJECT DETECTION AND COMPUTER VISION?
Answer: A. PyTorch Notebook Image
The PyTorch Notebook Image is correct because Red Hat builds this specific workbench image with pre-installed dependencies tailored for deep learning workflows like computer vision. TensorFlow is also valid, but PyTorch remains the standard computer vision default.
Q15. WHICH SETTING DETERMINES THE RESOURCE CONFIGURATION POOL OPTIONS SEEN BY DATA SCIENTISTS WHEN CONFIGURING WORKBENCHES?
Answer: B. Accelerator Profiles and Container Size settings in OpenShift AI settings panel
Accelerator profiles and container size settings are correct because these OpenShift AI settings directly control the compute options visible to users. Standard OpenShift quotas like LimitRanges restrict usage but do not populate the friendly selection menu.
Q16. What Red Hat OpenShift AI operator must be installed before configuring data science cluster-wide environments?
Answer: A. Red Hat OpenShift AI Operator
The Red Hat OpenShift AI Operator is required initially to install and configure the platform components. Other operators might be dependencies for specific features later, but this operator establishes the core environment.
Q17. What configuration handles automatic storage provisioning when creating a new workbench storage?
Answer: C. Dynamic Storage Classes matching requested volume specifications
Dynamic storage classes handle automatic volume provisioning when you create workbench storage in OpenShift AI. Manual PV creation defeats the purpose of dynamic cluster provisioning for data scientists.
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.