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: How can an administrator revoke a data scientist from spawning new workbench containers without deleting their account?. 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
- How can an administrator revoke a data scientist from spawning new workbench containers without deleting thei…
- Which OpenShift AI feature allows administrators to pre-define specific container images, environment variabl…
- Which node property prevents traditional non-AI workloads from apportioning CPU cores assigned to specialized…
- Which Operator expands OpenShift workbenches to leverage massively distributed machine learning computing?
- How does OpenShift AI manage user access control and permissions for launching dashboards and notebook server…
- Which Operator is used to install Red Hat OpenShift AI in a self-managed cluster?
Answers and explanations
Tap a question to expand the answer and the exam reasoning. Try to commit to your own pick first.
Q1. How can an administrator revoke a data scientist from spawning new workbench containers without deleting their account?
Answer: A. Remove the user from the rhods-users group
Removing the user from the rhods-users group revokes their OpenShift AI access without account deletion. Tainting nodes or scaling the dashboard would disrupt the entire environment, not just block one user from creating workbenches.
Q2. Which OpenShift AI feature allows administrators to pre-define specific container images, environment variables, and sizes for notebooks?
Answer: D. Workbench Templates
Workbench templates are the correct mechanism for administrators to pre-define container images, environment variables, and sizes. Notebook runtimes are generally deprecated in favor of these flexible templates in recent versions.
Q3. Which node property prevents traditional non-AI workloads from apportioning CPU cores assigned to specialized AI/ML workloads?
Answer: B. Node Taints and Tolerations properties
Applying node taints ensures only pods with matching tolerations, like your AI workloads, can run on those nodes. Network policies and pod disruption budgets do not restrict pod scheduling based on hardware allocation.
Q4. Which Operator expands OpenShift workbenches to leverage massively distributed machine learning computing?
Answer: B. CodeFlare Operator
The CodeFlare Operator integrates with OpenShift AI to manage distributed compute resources and scale out complex ML workloads. Advanced Cluster Management handles multiple cluster fleets rather than internal notebook scaling.
Q5. How does OpenShift AI manage user access control and permissions for launching dashboards and notebook servers?
Answer: C. OpenShift OAuth and RBAC groups
OpenShift AI relies directly on native OpenShift OAuth and RBAC groups to control dashboard and notebook access. There is no separate custom database, and AWS IAM policies only apply to cloud infrastructure.
Q6. Which Operator is used to install Red Hat OpenShift AI in a self-managed cluster?
Answer: C. Red Hat OpenShift AI Operator
The Red Hat OpenShift AI Operator is the official method for installing the platform on a self-managed cluster. The Open Data Hub Operator is the upstream community project and is not used for the certified exam.
Q7. Which property declares interactive dependency tasks constructing replicable pipeline phases sequentially in Python pipeline code?
Answer: C. KFP pipeline tasks decorators and dependencies declarations (.after() method)
Kubeflow Pipelines uses task dependencies and the .after() method in Python code to enforce sequential execution. Avoid Tekton or Bash workarounds on the exam, as OpenShift AI relies strictly on native KFP SDK constructs.
Q8. Which component serves as the intermediary artifact registry where intermediate datasets produced by pipeline tasks are retained?
Answer: A. S3 Object Storage Bucket configured for the pipeline server storage layer
OpenShift AI pipelines use the configured S3 object storage bucket to store intermediate datasets and artifacts. Git is only for source code, while node storage is ephemeral and unsuitable for persistent pipeline outputs.
Q9. Which CRD represents the instance to control single component upgrades separately within Red Hat OpenShift AI controllers?
Answer: A. DataScienceCluster COMPONENT property states
The DataScienceCluster custom resource allows administrators to manage individual OpenShift AI components independently. The ClusterServiceVersion handles the overall operator lifecycle but not granular component toggles.
Q10. Which resource management property ensures that JupyterLab workbenches automatically shut down if they remain inactive for extended periods?
Answer: A. Notebook Idleness Notebook Controller CRD configurations
The Notebook Controller CRD configurations manage idle workbenches, but the provided answer text reads awkwardly and overlaps conceptually with the specific idle culler settings. For the exam, be prepared to associate the culling timeouts directly with the Notebook Controller.
Q11. Which storage configuration represents the fastest data access mode rooted within the notebook server compute pools directly?
Answer: B. Local NVMe Persistent Volumes via RWO storage classes
Local NVMe Persistent Volumes offer the fastest physical disk I/O directly attached to the compute node. However, an emptyDir mapped to memory is technically faster from a pure latency standpoint, making this question slightly ambiguous depending on constraints.
Q12. What parameter declared in the Data Connection interface permits disabling certificate verification for self-signed testing object stores?
Answer: C. AWS_CA_BUNDLE configuration or custom dashboard environment override
AWS_CA_BUNDLE custom environment overrides allow disabling standard certificate verification for object stores. The wording is slightly ambiguous as Kubernetes often relies on insecure flags, but OpenShift AI workbenches heavily use AWS variables.
Q13. What specific object is used by pipeline runtime instances to arrange persistent isolated data channels across different runtime nodes?
Answer: A. Tekton Workspaces mapped to shared Persistent Volume Claims (PVC)
Pipelines use Tekton Workspaces backed by Persistent Volume Claims to share data across tasks. While emptyDir is used for temporary caching, it does not provide the persistent isolation required across different nodes.
Q14. Which tool enforces network isolation between individual user projects in Red Hat OpenShift AI?
Answer: A. OpenShift SDN or OVN-Kubernetes NetworkPolicies
OpenShift network policies using OVN-Kubernetes enforce isolation between user projects by restricting ingress and egress traffic. The other options are external networking components that do not natively secure OpenShift AI project isolation.
Q15. WHAT GROUP TYPE MUST AN ADMIN DEFINE IN OPENSHIFT OAUTH CONFIGURATION TO INTEGRATE CUSTOM OCP GROUPS WITH RHODS-USERS?
Answer: A. Group.v1.openshift.io object
A Group object from the OpenShift API is required to integrate custom groups with the default OpenShift AI user group. This allows administrators to manage OpenShift AI access natively using standard cluster role bindings.
Q16. Which tool should an administrator use to monitor real-time GPU utilization inside an OpenShift AI cluster?
Answer: C. NVIDIA Prometheus Exporter with OpenShift Monitoring
The NVIDIA Prometheus Exporter integrated with OpenShift Monitoring correctly exposes real-time GPU metrics. A basic command like oc get pods only shows resource status, lacking performance telemetry required for hardware utilization analysis.
Q17. What kind of OpenShift environment object prevents noisy neighbor issues among data science workbenches within a project?
Answer: C. ResourceQuotas
ResourceQuotas limit aggregate resource consumption per namespace, ensuring one workbench cannot exhaust cluster memory or CPU. LimitRanges restrict individual resource requests, while NetworkPolicies only handle East-West traffic isolation.
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.