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 tool should a data scientist use within Elyra to pass files generated in an upstream notebook to a downstream scrip. 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 tool should a data scientist use within Elyra to pass files generated in an upstream notebook to a downs…
- In Red Hat OpenShift AI, which component provides the user interface for launching Jupyter notebooks?
- Which environment variable can you configure to point an OpenShift AI pipeline to an alternative object stora…
- Which component orchestrates the deployment of the OpenShift AI software components across the entire cluster…
- What action should be taken if a workbench fails to launch due to an 'Insufficient gpus' scheduling constrain…
- Which OpenShift tool handles automated builds of container images directly from clean Git repository source c…
Answers and explanations
Tap a question to expand the answer and the exam reasoning. Try to commit to your own pick first.
Q1. What tool should a data scientist use within Elyra to pass files generated in an upstream notebook to a downstream script node?
Answer: D. Elyra Pipeline Output Files parameter tracking properties
Elyra relies on explicit output file declarations in the node properties to pass artifacts between pipeline steps. The other choices require external, manual intervention that breaks the automated execution flow expected during a pipeline run.
Q2. In Red Hat OpenShift AI, which component provides the user interface for launching Jupyter notebooks?
Answer: B. OpenShift AI Dashboard
The OpenShift AI dashboard provides the centralized interface where users launch and manage their Jupyter notebook workbenches. While the Operator handles backend installation, users interact directly with the dashboard to provision resources.
Q3. Which environment variable can you configure to point an OpenShift AI pipeline to an alternative object storage bucket?
Answer: B. PIPELINES_OBJECT_STORAGE_SECRET
The PIPELINES_OBJECT_STORAGE_SECRET environment variable allows you to specify an alternative bucket configuration. Generic endpoint variables lack the comprehensive credentials and connection details needed to access secure storage.
Q4. Which component orchestrates the deployment of the OpenShift AI software components across the entire cluster environment?
Answer: C. Red Hat OpenShift AI Operator
The Red Hat OpenShift AI Operator handles the full lifecycle and deployment of the platform components. Native Kubernetes controllers lack the custom resource logic required to properly install the specialized data science applications.
Q5. What action should be taken if a workbench fails to launch due to an 'Insufficient gpus' scheduling constraint?
Answer: B. Add more GPU-enabled compute nodes or modify the workbench hardware profile requirements
To resolve scheduling constraints, you must either add compute capacity or adjust the profile. Increasing replicas only increases the hardware demand, making the scheduling failure worse without adding available compute power.
Q6. Which OpenShift tool handles automated builds of container images directly from clean Git repository source code?
Answer: B. OpenShift BuildConfigs (Source-to-Image)
OpenShift BuildConfigs using Source-to-Image automate building container images directly from Git repositories. While Tekton handles complex CI/CD pipelines, BuildConfigs are the native OpenShift way to build images from source.
Q7. Which route optimization occurs automatically when an inference request is sent entirely within the same OpenShift Service Mesh?
Answer: B. Traffic completely bypasses the external cluster ingress router for direct pod-to-pod mesh routing
Service Mesh enables direct pod-to-pod routing, bypassing the external cluster ingress router for internal traffic. This internal routing lowers latency and reduces the load on cluster edge components.
Q8. What is the primary mechanism to isolate different data science experiments and projects from each other on a shared cluster?
Answer: B. Deploying them into distinct OpenShift Data Science Projects (Namespaces)
Deploying workloads into distinct Data Science Projects provides logical isolation using OpenShift namespaces. Building separate physical clusters is unnecessarily expensive, while splitting local host networks does not provide the required security and resource boundaries.
Q9. Where are the pipeline runtime logs permanently preserved after the executing pod is culled?
Answer: A. In the configured object storage bucket pipeline logging directory
Pipeline execution logs are permanently archived in the configured object storage bucket. Local pod logs are ephemeral and deleted when the pod is culled, meaning standard OpenShift local logging will not retain the historical run data.
Q10. Which system monitors custom application alert rules and triggers notifications based on model server performance drops?
Answer: D. Prometheus Alertmanager
Prometheus Alertmanager monitors custom application alert rules and triggers notifications based on performance drops. While Prometheus handles scraping and rule evaluation, Alertmanager manages the actual notification delivery.
Q11. Which container component manages model loading routing rules within a multi-model ModelMesh deployment?
Answer: C. ModelMesh container sidecar proxy
The ModelMesh sidecar proxy handles model loading and routing rules within the deployment. While ModelMesh can run alongside Istio for external traffic, the internal model loading logic is specifically handled by its container sidecar.
Q12. Which custom resource defines a target serving environment for machine learning runtimes inside OpenShift AI?
Answer: B. ServingRuntime
The ServingRuntime custom resource defines the target serving environment and its associated machine learning runtimes. In modern KServe architectures, ServingRuntime effectively replaces the deprecated and specific KServeRuntime resource.
Q13. Which serving runtime is native to ModelMesh within OpenShift AI for ONNX formatted models?
Answer: B. OVMS (OpenVINO Model Server)
The OpenVINO Model Server acts as a default runtime within ModelMesh to handle various formats including ONNX. Triton is another capable runtime but is not the default native option tested.
Q14. Which custom resource (CR) is used to configure global settings for the Red Hat OpenShift AI operator?
Answer: A. DSCInitialization
The DSCInitialization custom resource configures global settings required by the Red Hat OpenShift AI operator before components are deployed. The DataScienceCluster custom resource handles enabling specific individual components.
Q15. Which storage class property is highly recommended for storing stateful data science workbench volumes?
Answer: A. ReadWriteOnce
ReadWriteOnce is the standard access mode for binding persistent volumes to a single workbench pod. ReadWriteMany is only necessary if you are explicitly configuring shared filesystem storage across multiple pods.
Q16. What default namespace is created for the Red Hat OpenShift AI core control plane components?
Answer: D. redhat-ods-operator
The redhat-ods-operator namespace is automatically created during installation to host the core control plane and operator components. Remembering this prefix is crucial for debugging operator logs during the exam.
Q17. To expose a machine learning model securely outside the OpenShift cluster via KServe, which component is required for routing?
Answer: C. OpenShift ingress controller with OpenShift Serverless and Service Mesh
Exposing KServe models securely requires the OpenShift ingress controller alongside Serverless and Service Mesh components. Standard NodePort services bypass the Knative routing and advanced traffic management capabilities needed.
Q18. Which OpenShift AI feature enables multiple users to collaborate inside the same exact runtime workspace simultaneously?
Answer: B. OpenShift AI workbenches do not natively support live multi-user concurrent typing; collaboration is done via Git repositories
OpenShift AI workbenches lack native concurrent editing, so users collaborate using Git. Shared system passwords are a distractor and violate security best practices.
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.