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 happens if the model file specified in an InferenceService resource exceeds the remaining persistent disk capacity . 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 happens if the model file specified in an InferenceService resource exceeds the remaining persistent dis…
- Which property metadata allows the OpenShift AI Dashboard interfaces to accurately identify a custom model se…
- WHICH TOOL AUTOMATES CONTINUOUS RESOURCE SYNCHRONIZATION ALONG MULTI-CLUSTER OPENSHIFT PIPELINES NATIVELY ENF…
- What configuration object defines the external endpoint mapping for access to the OpenShift AI dashboard UI?
- What happens to a running model serving deployment if the underlying object storage connection suffers an ext…
- What is the role of a Pipeline Run Trigger within an automated data science architecture?
Answers and explanations
Tap a question to expand the answer and the exam reasoning. Try to commit to your own pick first.
Q1. What happens if the model file specified in an InferenceService resource exceeds the remaining persistent disk capacity of the model server?
Answer: B. The model fails to mount or initialize, throwing a disk capacity or storage exhaustion error
The model fails to initialize because the underlying storage cannot fulfill the capacity request. OpenShift does not automatically compress model weights or arbitrarily slice files across unrelated workbench pods.
Q2. Which property metadata allows the OpenShift AI Dashboard interfaces to accurately identify a custom model serving runtime ConfigMap?
Answer: D. opendatahub.io/dashboard: 'true' label applied onto the target ConfigMap definition
The opendatahub.io/dashboard true label exposes the custom runtime to the OpenShift AI user interface. Without this specific label, the dashboard simply ignores the custom serving runtime configuration.
Q3. WHICH TOOL AUTOMATES CONTINUOUS RESOURCE SYNCHRONIZATION ALONG MULTI-CLUSTER OPENSHIFT PIPELINES NATIVELY ENFORCING INFRASTRUCTURE STANDARDS?
Answer: B. Red Hat OpenShift GitOps Operator implementing continuous synchronization engines
Red Hat OpenShift GitOps provides continuous resource synchronization across clusters. Tekton handles task execution rather than enforcing baseline infrastructure standards across multi-cluster environments.
Q4. What configuration object defines the external endpoint mapping for access to the OpenShift AI dashboard UI?
Answer: C. OpenShift Route Custom Resource
An OpenShift Route exposes the service providing the dashboard UI to external network traffic. While Service Mesh may handle internal routing, the standard OpenShift Route provides the external endpoint mapping.
Q5. What happens to a running model serving deployment if the underlying object storage connection suffers an extended network blackout?
Answer: D. The model continues serving requests using its in-memory cached copy but cannot reload or update
The correct answer applies because the runtime caches the model in memory after the initial load, isolating active inference from transient storage disconnects. Serving pods rely on the local cache to keep responding to prediction requests, even though they cannot fetch updates until connectivity returns.
Q6. What is the role of a Pipeline Run Trigger within an automated data science architecture?
Answer: C. Automating pipeline executions based on periodic schedules or chronological cron settings
The correct answer works because triggers are used to schedule recurring runs or execute them via chronological cron patterns. Other options confuse orchestration with CI/CD syncing or raw container compilation, which are handled outside the pipeline server.
Q7. What parameter inside an Elyra pipeline layout defines the execution dependencies between individual processing steps?
Answer: B. The upstream/downstream connection links saved within the .pipeline JSON metadata definition
The correct answer applies because Elyra saves the visual dependencies as upstream and downstream links within the JSON metadata. Knowing the underlying pipeline file structure helps quickly eliminate distractors like storage timestamps or resource sizes.
Q8. What happens to scheduled pipeline runs if the internal pipeline database storage volume runs entirely out of space?
Answer: A. New pipeline runs fail to schedule or record metadata tracking logs
The correct answer applies because the pipeline database requires free space to schedule new runs and record metadata. The platform will not automatically purge old execution history, and it avoids needlessly terminating active model deployments to free space.
Q9. Which command retrieves the configuration properties of the global dashboard component to check active feature gates?
Answer: C. oc get dashboardconfigs.dashboard.opendatahub.io -n redhat-ods-applications rhoai-dashboard -o yaml
The correct answer works because the dashboard configuration is retrieved from the custom resource name using the OpenDataHub API group. Memorize the exact namespace and custom resource definitions, as operator logs do not display structured feature gates.
Q10. What strategy does ModelMesh use to handle requests for a model that is currently stored in object storage but not active in memory?
Answer: D. Lazy loading where the model is dynamically pulled into an available slot while caching others
ModelMesh uses lazy loading to dynamically pull models from storage into memory only when requested, maximizing resource efficiency. This intelligently handles memory limits by evicting less frequently used models rather than crashing or timing out.
Q11. When creating a custom dashboard configuration, which file format must be uploaded to define custom notebook images?
Answer: A. ConfigMap YAML
Custom notebook images are defined by applying a ConfigMap in YAML format to the cluster. This configuration allows the OpenShift AI dashboard to dynamically discover and display the new image options for users.
Q12. Which optimization framework is typically integrated into the TGIS (Text Generation Inference Server) engine to accelerate LLM generation speeds?
Answer: C. TensorRT-LLM or specialized compilation optimization backends
TGIS integrates TensorRT-LLM and other specialized compilation backends to significantly accelerate large language model generation speeds. This drastically reduces latency compared to standard Python execution during inference operations.
Q13. Which command confirms that the TrustyAI custom resource definitions are properly populated in the active cluster system?
Answer: C. oc get crd | grep trustyai
Running oc get crd piped to grep is the standard CLI method to verify that custom resource definitions are properly installed in the cluster. Operator logs only show installation attempts, whereas this confirms the actual resources exist.
Q14. How can you isolate the training workloads of a premium data science team from a standard data science team on an OpenShift cluster?
Answer: A. Configure Node Selectors and Taints targeting specific high-performance node pools
Configuring node selectors and tolerations targets specific高性能 node pools, effectively isolating premium data science workloads from standard ones. Deleting projects or sharing ModelMesh deployments does not provide hardware-level isolation.
Q15. Which API property under the ServingRuntime manifest determines the compatible model formats (e.g., ONNX, PyTorch)?
Answer: D. supportedModelFormats
The supportedModelFormats property in the ServingRuntime manifest declares which model frameworks the runtime can load. This is crucial when configuring KServe to match your specific model file types.
Q16. What is the primary architectural boundary used to isolate different multi-tenant environments inside OpenShift AI?
Answer: D. Native OpenShift Namespaces with Kubernetes RBAC and Network Policies
Namespaces with RBAC and network policies provide the native multitenancy isolation in OpenShift AI. Physical hardware or browser profiles are distractors because OpenShift AI relies entirely on Kubernetes constructs to separate tenant workloads logically.
Q17. What security exposure happens if an administrator marks an OpenShift AI route as insecure or unencrypted?
Answer: D. Traffic passes in plain text exposing tokens and credentials over the network
Disabling encryption forces all dashboard traffic to pass in plain text, exposing sensitive credentials. The route will still function normally, but attackers could intercept unencrypted payloads across the network.
Q18. Which component is responsible for authenticating user login tokens when accessing the OpenShift AI dashboard web application?
Answer: D. OpenShift Container Platform OAuth Cluster Service
The OpenShift Container Platform OAuth service authenticates users for the OpenShift AI dashboard. OpenShift AI delegates authentication directly to the underlying cluster rather than relying on internal Kubeflow databases.
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.