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 open source workflow tool does Elyra leverage to output valid runtime tasks onto an OpenShift Pipelines backend ar. 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 open source workflow tool does Elyra leverage to output valid runtime tasks onto an OpenShift Pipelines…
- For multi-model serving where high-density deployment on shared resources is required, which component does O…
- What is the function of the DSCInitialization custom resource in an OpenShift AI deployment?
- Which OpenShift AI component is responsible for managing distributed training workloads using frameworks like…
- Which dashboard component displays real-time performance indicators for served models?
- Which project template should a developer use to set up automated pipelines using Git repository events?
Answers and explanations
Tap a question to expand the answer and the exam reasoning. Try to commit to your own pick first.
Q1. Which open source workflow tool does Elyra leverage to output valid runtime tasks onto an OpenShift Pipelines backend architecture?
Answer: D. Tekton pipelines templates compiled configurations
Elyra leverages Tekton pipelines templates to output valid runtime tasks onto an OpenShift Pipelines backend. Ignore distractors like Jenkins or Airflow, as OpenShift Pipelines natively relies on Tekton.
Q2. For multi-model serving where high-density deployment on shared resources is required, which component does OpenShift AI utilize?
Answer: D. ModelMesh
ModelMesh is purpose-built for high-density, multi-model serving on shared resources in OpenShift AI. Knative Serving is strictly for single-model serverless deployments rather than multi-model use cases.
Q3. What is the function of the DSCInitialization custom resource in an OpenShift AI deployment?
Answer: B. It configures cluster-wide prerequisites and global settings for the RHOAI operator
The DSCInitialization resource sets up cluster-wide prerequisites and global settings for the RHOAI operator. It does not define individual notebook quotas or manage GPU drivers, which are handled elsewhere.
Q4. Which OpenShift AI component is responsible for managing distributed training workloads using frameworks like PyTorch or TensorFlow?
Answer: C. Training Operator
The Kubeflow Training Operator is responsible for managing distributed training operators using frameworks like PyTorch and TensorFlow. GitOps handles cluster state synchronization, while CodeFlare focuses specifically on orchestrating Ray-based workloads.
Q5. Which dashboard component displays real-time performance indicators for served models?
Answer: B. OpenShift AI Metrics and Monitoring tab dashboards
Real-time performance indicators for served models are displayed in the OpenShift AI Metrics and Monitoring tab dashboards. The user's terminal is isolated to a notebook, and Tekton pipeline logs do not show live inference metrics.
Q6. Which project template should a developer use to set up automated pipelines using Git repository events?
Answer: B. OpenShift Pipelines Triggers
OpenShift Pipelines Triggers is the correct technology used to instantiate pipelines automatically based on Git repository events. GitOps webhooks manage continuous deployment synchronization, not pipeline task execution.
Q7. If a workbench fails to start due to an "Insufficient memory" error, what is the best immediate administrative solution?
Answer: C. Increase node count or choose a smaller workbench size configuration
Fixing insufficient memory involves adding cluster nodes or selecting a smaller workbench size configuration. Changing the image or restarting the API server will not resolve underlying hardware allocation limits.
Q8. What configuration defines the container registry pull secrets for custom notebook images?
Answer: B. Image Pull Secrets linked to the namespace ServiceAccount
Custom notebook images require Image Pull Secrets linked directly to the namespace ServiceAccount to authenticate properly. Modifying global cluster pull secrets is generally discouraged and plaintext environment variables are insecure.
Q9. What file format is typically used to store the metadata and weights of a model designed to run on the ONNX runtime inside OpenShift AI?
Answer: C. .onnx
Models intended for the ONNX runtime use the .onnx file format to package metadata and weights for cross-platform inference. The other formats represent framework-specific storage, such as PyTorch or TensorFlow models.
Q10. What must be included in an external custom runtime image to allow the OpenShift AI notebook controller to properly manage its launch behavior?
Answer: C. A pre-configured Jupyter notebook or JupyterLab server operating on port 8888
Custom notebook images must run a Jupyter or JupyterLab server on port 8888 for the OpenShift AI controller to attach and manage the pod properly. If the image lacks this endpoint, the dashboard cannot route the OAuth proxy session.
Q11. What cluster logging configuration is recommended to retain historical inference request data for auditing purposes?
Answer: B. Integrating OpenShift Logging to forward container stdout logs to a secure external SIEM or Elasticsearch cluster
OpenShift AI inference traffic is logged to standard output, so forwarding these stdout logs via OpenShift Logging to an external system preserves them securely. Browser caches or ephemeral node storage are volatile and will not satisfy auditing requirements.
Q12. What configuration element prevents data science pipelines from accessing unauthorized external networks?
Answer: B. Egress NetworkPolicies applied to the data science namespace
Applying egress NetworkPolicies to the data science namespace restricts outgoing traffic, preventing pipelines from reaching unauthorized external networks. Ingress rules only manage incoming traffic and do not secure outbound connections from the pipeline workloads.
Q13. Which environment variable holds the access token for authenticating internal pipeline API calls?
Answer: D. KUBERNETES_SERVICE_ACCOUNT_TOKEN
Internal pipeline API calls authenticate automatically via the mounted service account token stored in KUBERNETES_SERVICE_ACCOUNT_TOKEN. For the exam, remember that OpenShift AI injects this standard Kubernetes volume mount rather than relying on manually configured static tokens.
Q14. What role must a user possess within an OpenShift project to create a Data Science Workbench?
Answer: B. admin or edit role
Users need the admin or edit role within a project to create a Data Science Workbench. These roles grant the necessary Kubernetes RBAC permissions to manage project resources, whereas view is strictly read-only.
Q15. Which protocol is standard for sending inference requests to models deployed using the KServe v2 Data Plane standard?
Answer: A. v2 REST and gRPC API protocols
The KServe v2 Data Plane standard uses the v2 REST and gRPC API protocols to handle inference requests. For the exam, remember that standard OpenShift AI model serving relies on these modern API formats, discarding legacy enterprise messaging protocols.
Q16. Which custom resource definition tracks the status of a single model deployment under KServe?
Answer: B. InferenceService
The InferenceService custom resource definition tracks the status of a single model deployment under KServe. Expect to query this resource using oc or kubectl to verify endpoint readiness or troubleshoot deployment failures.
Q17. What parameters define the minimum and maximum scaling boundaries of a KServe single-model predictor deployment configuration?
Answer: D. minReplicas and maxReplicas annotations or spec fields
The minReplicas and maxReplicas spec fields or annotations define the scaling boundaries for KServe predictor deployments. Rely on these standard Kubernetes scaling parameters rather than custom environment variables.
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.