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: Which component state in the DataScienceCluster CR instructs the operator to actively reconcile and keep the component r. 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 component state in the DataScienceCluster CR instructs the operator to actively reconcile and keep the…
- Which container registry is standard for downloading official Red Hat OpenShift AI workbench images?
- What format must a machine learning model be saved in to allow deployment using the default OpenShift AI Open…
- What is the primary benefit of using KServe over ModelMesh for advanced deep learning deployments?
- To connect an external Git repository to a Jupyter notebook cell securely inside OpenShift AI, which object s…
- When deploying a model via KServe, which component provides the ingress routing gateway functionality?
Answers and explanations
Tap a question to expand the answer and the exam reasoning. Try to commit to your own pick first.
Q1. Which component state in the DataScienceCluster CR instructs the operator to actively reconcile and keep the component running?
Answer: B. Managed
Setting a component to Managed in the DataScienceCluster custom resource tells the operator to actively reconcile and maintain that component. The Removed state disables it entirely.
Q2. Which container registry is standard for downloading official Red Hat OpenShift AI workbench images?
Answer: B. registry.redhat.io
Official Red Hat container images, including OpenShift AI workbenches, are hosted on registry.redhat.io and require proper pull secret authentication. Public registries like Quay community are not used for these enterprise images.
Q3. What format must a machine learning model be saved in to allow deployment using the default OpenShift AI OpenVINO runtime?
Answer: D. ONNX or IR format
The OpenVINO serving runtime requires models to be saved in ONNX or IR format. Pickled Python objects or raw script files are distractors for standard serving runtimes.
Q4. What is the primary benefit of using KServe over ModelMesh for advanced deep learning deployments?
Answer: C. Support for large single models, auto-scaling to zero, and serverless framework integration
KServe is chosen for supporting large models, serverless integration, and scaling to zero. ModelMesh is a distractor designed for high-density scikit-learn deployments.
Q5. To connect an external Git repository to a Jupyter notebook cell securely inside OpenShift AI, which object should be used?
Answer: D. Secret
Secrets securely hold Git credentials for OpenShift AI authentication. ConfigMaps are a distractor because they store non-sensitive configuration data rather than passwords.
Q6. When deploying a model via KServe, which component provides the ingress routing gateway functionality?
Answer: A. Istio Ingress Gateway
KServe relies on the Istio Ingress Gateway for routing inference traffic. Standard OpenShift Routers are distractors that do not natively handle KServe traffic.
Q7. How can an administrator inject a custom CA certificate into the OpenShift AI environment for secure corporate proxy communication?
Answer: D. By adding the certificate to the trusted CA bundle configuration in the DSCInitialization CR
Trusted CA certificates are injected by configuring the trusted CABundle in the DSCInitialization CR. Manual uploads to JupyterHub are a distractor.
Q8. What is the primary configuration object for setting up an acceleration profile for AI workloads using GPUs?
Answer: B. AcceleratorProfile custom resource
The AcceleratorProfile custom resource defines GPU settings for workloads. MachineConfig files are a distractor because they handle general node configuration rather than AI profiles.
Q9. Which OpenShift AI feature allows administrators to inject specific environment variables or volume mounts into user notebook pods automatically?
Answer: A. PodDefault CR custom resources
The PodDefault custom resource is applied to notebook pods to mount volumes or inject variables automatically. Routes are a distractor because they only expose external traffic.
Q10. What component handles the scaling of inference service pods to zero instances when no web requests are received by KServe?
Answer: D. Knative Serving component
Knative Serving provides the serverless capabilities that allow KServe inference services to scale to zero when idle. This reduces compute costs, whereas ModelMesh keeps models pre-loaded for high-throughput serving.
Q11. Which tool can be used to synchronize machine learning model deployments from a Git repository to OpenShift AI automatically?
Answer: C. Red Hat OpenShift GitOps (Argo CD)
Red Hat OpenShift GitOps uses Argo CD to automatically synchronize declarative Git repository states to the cluster. While Tekton handles continuous integration tasks, Argo CD provides the continuous deployment synchronization.
Q12. Which deployment strategy does KServe use to scale down model serving pods when no requests are active?
Answer: D. Knative Serving scale-to-zero
Standard KServe relies on Knative Serving capabilities to manage traffic and scale pods to zero when inactive. ModelMesh, in contrast, targets high-density serving and keeps models loaded in memory.
Q13. Which runtime configuration allows a pipeline to execute a step using an absolute, immutable container image reference?
Answer: B. Specifying the unique container image SHA digest
Specifying the container image SHA digest ensures that the pipeline executes an exact, immutable version of the code. Relying on mutable tags like 'latest' risks pulling a newly updated and potentially breaking image.
Q14. Which engine is default for serving traditional tabular ML models (like Scikit-Learn) within the OpenShift AI ModelMesh framework?
Answer: A. MLServer
MLServer is the default runtime used by ModelMesh to serve common tabular frameworks like Scikit-Learn and XGBoost. vLLM or Triton are generally targeted at large language models or other specialized inference tasks.
Q15. Which underlying Tekton resource is automatically created when a user triggers an immediate manual run of a Data Science Pipeline?
Answer: C. PipelineRun
Triggering a manual pipeline run in OpenShift AI generates a Tekton PipelineRun custom resource. While TaskRun objects are created for individual steps, the PipelineRun wraps the entire pipeline execution.
Q16. How can an administrator scale up the processing capabilities of an active distributed training run using Ray?
Answer: D. Increase the worker replicas count in the RayCluster CR
To scale a distributed Ray workload, an administrator increases the worker replica count in the RayCluster custom resource. The Ray operator then automatically provisions the requested worker pods to handle the load.
Q17. Which CLI command allows you to switch your terminal context to the specific Data Science Project namespace?
Answer: C. oc project
The oc project command changes the active terminal namespace to the specified data science project. Other options are either invalid kubectl syntax or fabricated commands, making C the clear choice for navigating projects during the exam.
Q18. What storage configuration prevents data loss when a workbench pod is deleted or recreated?
Answer: D. Attaching a Persistent Volume Claim (PVC) to the workbench
A Persistent Volume Claim ensures data persists independently of the workbench pod lifecycle. Storing files in the container root directory or temporary folders leads to immediate data deletion when the pod restarts or is recreated.
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.