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 occurs when an OpenShift AI pipeline run finishes successfully?. 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 occurs when an OpenShift AI pipeline run finishes successfully?
- To assign dedicated cluster infrastructure to a workbench, which combination of parameters must be defined in…
- What type of OpenShift resource links a Data Science Project to an underlying Kubernetes namespace?
- To change the default timeout duration for all pipeline executions in a project, which component resource mus…
- What configuration parameter prevents unauthorized access to the pipeline engine API?
- When creating an external data connection, what happens if you omit the AWS_DEFAULT_REGION parameter for a st…
Answers and explanations
Tap a question to expand the answer and the exam reasoning. Try to commit to your own pick first.
Q1. What occurs when an OpenShift AI pipeline run finishes successfully?
Answer: D. The executing pod is typically terminated or kept according to retention policies, and artifacts are saved to object storage
Successful pipeline runs complete their container execution and persist artifacts to the configured object storage. The cluster itself remains unchanged, and no automatic notifications are sent.
Q2. To assign dedicated cluster infrastructure to a workbench, which combination of parameters must be defined in the container specification?
Answer: C. tolerations and nodeSelector
Combining node selectors with tolerations targets specific dedicated infrastructure while preventing standard workloads from interrupting the workbench. Standard resource limits alone cannot guarantee placement on specialized machines.
Q3. What type of OpenShift resource links a Data Science Project to an underlying Kubernetes namespace?
Answer: C. A dedicated OpenShift Namespace/Project resource
A Data Science Project directly provisions and links to a dedicated OpenShift Namespace. Network policies and role bindings manage access and traffic, but they do not provide the foundational namespace isolation required.
Q4. To change the default timeout duration for all pipeline executions in a project, which component resource must be edited?
Answer: B. The specific Pipeline Server configuration settings or custom resource
Pipeline execution limits are governed by the Pipeline Server configuration settings. Node manifests or operator subscriptions control cluster-wide scheduling and updates, but they do not manage project-specific pipeline timeouts.
Q5. What configuration parameter prevents unauthorized access to the pipeline engine API?
Answer: B. OpenShift RBAC and OAuth proxy integration on pipeline routes
OpenShift RBAC combined with an OAuth proxy properly secures the pipeline engine API by validating identities before granting access. Relying on plain text tokens, hidden ports, or disabled routing does not provide adequate security.
Q6. When creating an external data connection, what happens if you omit the AWS_DEFAULT_REGION parameter for a standard AWS S3 bucket?
Answer: B. The client application library defaults to the us-east-1 region or fails initialization
If the region is omitted, AWS client libraries typically default to the standard endpoint or fail. The workbench does not crash during creation, and OpenShift does not automatically provision a replacement bucket locally.
Q7. Which technology facilitates high-speed, point-to-point network communication among distributed training nodes in a Ray cluster?
Answer: B. Headless Kubernetes Services and gRPC channels
Ray clusters utilize headless Kubernetes Services and gRPC for high-speed, point-to-point network communication between distributed worker nodes. Other options like basic host networking do not provide the necessary internal routing and discovery mechanisms.
Q8. What mechanism protects the TrustyAI dashboard service endpoint from unauthorized user access within a project?
Answer: B. OpenShift OAuth proxy authentication integration wrappers
An OpenShift OAuth proxy sidecar wraps the endpoint to enforce authentication and protect against unauthorized access. This is the standard security mechanism for exposing dashboards safely within OpenShift.
Q9. Which storage class type is highly recommended for persistent volume claims used by Jupyter notebooks in OpenShift AI?
Answer: C. ReadWriteOnce (RWO)
ReadWriteOnce is the standard access mode for workbench PVCs because a notebook pod runs on a single node. While ReadWriteMany works in distributed storage, single-node access is the expected baseline for standard workbenches.
Q10. Which underlying technology manages secure cross-pod network communication rules within an OpenShift AI deployment?
Answer: C. OpenShift SDN or OVN-Kubernetes CNI plugins
OpenShift uses OVN-Kubernetes or SDN plugins to enforce network policies and secure pod communication. While CoreDNS handles name resolution, it does not manage traffic filtering or cross-pod security rules.
Q11. Which microservice handles the calculation of data drift metrics inside the TrustyAI architectural framework?
Answer: B. trustyai-service pod controller
The TrustyAI service handles explainability, fairness, and drift metrics for models deployed on the platform. For the exam, ignore the fake container names and focus on the official OpenShift AI components designed for model monitoring and TrustyAI operations.
Q12. What is the primary language environment used by data scientists within Jupyter workbenches for AI model development?
Answer: C. Python
Python serves as the primary programming environment for data scientists developing AI models inside Jupyter workbenches. While other languages exist, the exam focus remains heavily on Python-specific notebook images and dependencies.
Q13. Which component handles user interface operations and visual tracking of experiments within OpenShift AI pipelines?
Answer: C. Kubeflow Pipelines UI
OpenShift AI pipelines use the Kubeflow Pipelines UI for tracking experiments visually. Tekton is the underlying engine, but it does not provide the specialized data science experiment interface.
Q14. What happens if the DSCInitialization resource is completely deleted from an active OpenShift AI cluster?
Answer: C. The operator reinitializes it or the cluster-wide control plane core features break
Deleting the DSCInitialization resource disrupts the core cluster-wide control plane components in OpenShift AI. The operator typically attempts to reconcile and recreate it, but missing dependencies can break platform features, making the disruption answer the safest choice.
Q15. What role-based access control cluster role must be bound to a service account for it to modify InferenceService resources?
Answer: B. admin or edit cluster role profiles
Modifying custom resources like an InferenceService requires write permissions typically granted by the admin or edit role. Binding one of these roles to the service account allows it to make the necessary changes within the project namespace.
Q16. What is the maximum number of models that can run on a single ModelMesh pod before additional resource allocations occur?
Answer: A. It depends on memory availability and configured deployment density thresholds
The number of models running on a single pod is dictated by memory availability and configured deployment density thresholds. There is no fixed global ceiling, as ModelMesh dynamically loads and evicts models to stay within limits.
Q17. What parameter specifies the minimum compute capability needed by a Ray worker node to manage model matrix operations?
Answer: D. limits.cpu and limits.memory parameters
The limits.cpu and limits.memory parameters define the compute resources allocated to a Ray worker node for operations like matrix multiplication. Storage classes and service mesh routing do not configure compute capacity.
Q18. Which OpenShift AI component is responsible for authenticating users into their respective workbench workspaces?
Answer: C. OpenShift OAuth
OpenShift AI delegates workbench authentication to the native OpenShift OAuth provider. Options like Keycloak or htpasswd are not the direct component, and the remaining choices are unrelated navigation artifacts from a practice test.
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.