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: When setting up data connections in OpenShift AI, which storage standard is predominantly used to reference object stora. 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
- When setting up data connections in OpenShift AI, which storage standard is predominantly used to reference o…
- What configuration item is required to authorize an OpenShift pipeline to read data from a protected object s…
- What mechanism does OpenShift AI use to ensure that model inference endpoints are resilient to node failures?
- Which Custom Resource manages the global configuration required for initialization before configuring compone…
- What parameter controls how many worker nodes are allowed to join a distributed Ray cluster during an elastic…
- Where can an administrator view the total GPU utilization across all OpenShift AI workbenches?
Answers and explanations
Tap a question to expand the answer and the exam reasoning. Try to commit to your own pick first.
Q1. When setting up data connections in OpenShift AI, which storage standard is predominantly used to reference object storage?
Answer: D. S3 Compatible Object Storage
S3 compatible object storage is the standard used for referencing storage in OpenShift AI data connections. Remember that the AWS secret and access keys map directly to your S3 provider, making the other block or file storage options incorrect.
Q2. What configuration item is required to authorize an OpenShift pipeline to read data from a protected object storage server?
Answer: C. An object storage credential secret mapped into a pipeline service account
To authorize a pipeline to read from protected object storage, you must provide a credential secret mapped to a service account. OpenShift pipelines consume these secrets directly rather than relying on broad cluster admin roles or open network bridges.
Q3. What mechanism does OpenShift AI use to ensure that model inference endpoints are resilient to node failures?
Answer: A. Kubernetes deployment replication and pod rescheduling across available worker nodes
Kubernetes deployment replication and pod rescheduling across worker nodes ensure inference endpoints survive node failures. If a node drops, the cluster detects the lost pods and recreates them elsewhere to maintain the desired replica state.
Q4. Which Custom Resource manages the global configuration required for initialization before configuring components?
Answer: C. DSCInitialization
The DSCInitialization custom resource manages the global configuration required before configuring other components. It handles prerequisite setup so that the DataScienceCluster resource can properly deploy the remaining operator managed components.
Q5. What parameter controls how many worker nodes are allowed to join a distributed Ray cluster during an elastic training cycle?
Answer: B. maxReplicas in the RayCluster worker group specification
The maxReplicas parameter within the RayCluster worker group specification sets the upper limit for autoscaling worker nodes in a distributed Ray cluster. Options like replicaCount in the DataScienceCluster CR do not control Ray worker autoscaling.
Q6. Where can an administrator view the total GPU utilization across all OpenShift AI workbenches?
Answer: C. OpenShift Web Console Administrator view under Observe or Metrics
Cluster-level resource metrics like GPU utilization are observed through the OpenShift Web Console under the Observe or Metrics dashboards. The OpenShift AI Dashboard settings lack cluster-wide hardware metrics, making the web console the required operational view.
Q7. Which metrics framework is natively integrated with OpenShift AI to collect operational data?
Answer: B. Prometheus and Grafana
OpenShift AI relies on the native OpenShift monitoring stack, which uses Prometheus to collect metrics and Grafana to visualize them. External platforms like Datadog or Splunk require custom integration and are not the default framework.
Q8. Which log stream should you analyze first if the Red Hat OpenShift AI Dashboard fails to load or shows a 500 error?
Answer: B. The rhods-dashboard pod logs within the redhat-ods-applications namespace
When the dashboard fails, the immediate troubleshooting step is reviewing the rhods-dashboard pod logs in the redhat-ods-applications namespace. API audit trails or worker node systemd logs are too broad and ignore the application layer failure.
Q9. Which custom resource coordinates specific component initialization parameters such as custom certificates across OpenShift AI?
Answer: A. DSCInitialization
The DSCInitialization custom resource coordinates specific component initialization parameters across OpenShift AI, such as custom certificates. The DataScienceCluster component configures which AI features are actually enabled.
Q10. When configuring an S3-compatible Data Connection, which credential key identifies the connection location?
Answer: A. AWS_S3_ENDPOINT
The AWS_S3_ENDPOINT specifies the S3-compatible connection location or URL. Access keys handle authentication, the bucket name targets the storage container, and the region defines geographical location for native AWS endpoints rather than custom S3 connections.
Q11. What custom resource status field indicates that all OpenShift AI control plane tools are successfully reconciled and ready?
Answer: C. status.conditions with status True
Standard Kubernetes controllers use status.conditions set to True to signal successful reconciliation and readiness. ManagementState only indicates the operator's administrative intent, while spec.phase and status.phase are not the standard condition checks here.
Q12. Which open-source project is used as the base dashboard interface engine for OpenShift AI?
Answer: C. Open Data Hub Dashboard
The Open Data Hub Dashboard serves as the underlying open-source interface engine for Red Hat OpenShift AI. Grafana handles metrics visualization and Kibana parses logs, but neither functions as the primary machine learning project and workload dashboard.
Q13. Which state must be defined in the DataScienceCluster CR component specification to disable a specific tool like JupyterHub?
Answer: D. Removed
Setting the component management state to Removed deletes the related operator and resources. Disabled is a strong distractor because it stops the component, but Removed is required for full cleanup.
Q14. Which component state in the DataScienceCluster configuration completely removes the related operator deployment and objects?
Answer: C. Removed
Setting the component management state to Removed deletes the related operator and resources. Deleted is a tempting distractor, but Removed is the specific keyword required by the OpenShift AI operator.
Q15. What is the primary advantage of deploying models using KServe over ModelMesh in OpenShift AI?
Answer: B. Advanced single-model deployment features, custom routing, and serverless scale-to-zero capabilities
KServe provides serverless inference with scale-to-zero capabilities and custom routing for individual models. ModelMesh is designed for high scale, multi-model fitting rather than standalone serverless deployments.
Q16. What role-based access control (RBAC) permission is required for a user to create a new Data Science Project?
Answer: A. self-provisioner or dedicated project creation rights
Users need the self-provisioner role to create a Data Science Project. On the exam, remember that OpenShift AI projects map directly to Kubernetes namespaces, so standard cluster RBAC governs creation.
Q17. Which runtime engine provides optimal inference processing speeds for deep learning models built using the PyTorch framework?
Answer: C. Triton Inference Server or TorchServe
The Triton Inference Server and TorchServe are optimized to serve deep learning PyTorch models. Standard web frameworks like Flask lack the necessary hardware acceleration capabilities for production inference.
Q18. Which configuration file must be modified to add custom environment variables to all workbenches globally?
Answer: D. The workbench settings or cluster-wide notebook configurations via the OpenShift AI dashboard admin panel
Global environment variables for workbenches are configured through the dashboard settings or cluster-wide notebook custom resources. You never modify core system files directly inside the operator pod.
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.