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 is the purpose of the trust-bundle ConfigMap in an enterprise OpenShift AI deployment?. 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 is the purpose of the trust-bundle ConfigMap in an enterprise OpenShift AI deployment?
- Which metric value for Disparate Impact indicates that an AI model exhibits zero bias between a protected and…
- What is the consequence of configuring a workbench image with an incorrect ImageStream reference path?
- Which specific configuration within an Accelerator Profile maps directly to an NVIDIA GPU resource type?
- Which storage access mode is required when multiple workbenches need to read and write to the same training d…
- Which metric helps administrators identify underutilized workbench instances across the OpenShift AI platform?
Answers and explanations
Tap a question to expand the answer and the exam reasoning. Try to commit to your own pick first.
Q1. What is the purpose of the trust-bundle ConfigMap in an enterprise OpenShift AI deployment?
Answer: A. Injecting custom CA certificates for secure corporate proxy communication
The trust-bundle ConfigMap is used to inject custom certificate authority certificates into workbenches, ensuring secure connections through corporate proxies. It does not store user credentials, authenticate inference requests, or handle PVC encryption operations.
Q2. Which metric value for Disparate Impact indicates that an AI model exhibits zero bias between a protected and unprotected group?
Answer: B. 1
A Disparate Impact value of one indicates perfect parity, meaning the AI model exhibits zero bias between protected and unprotected groups. Values below or above one signify disproportionate outcomes, with zero representing complete exclusion.
Q3. What is the consequence of configuring a workbench image with an incorrect ImageStream reference path?
Answer: C. The workbench fails to start and displays an ImagePullBackOff status error
An incorrect ImageStream reference prevents the cluster from pulling the correct workbench image, resulting in an ImagePullBackOff error. The platform will not substitute default images or reboot nodes, as it simply fails the pod deployment.
Q4. Which specific configuration within an Accelerator Profile maps directly to an NVIDIA GPU resource type?
Answer: A. k8s.io/nvidia-gpu
The identifier k8s.io/nvidia-gpu correctly maps the accelerator profile to the NVIDIA GPU hardware resource. Other options are either invalid syntax or represent Node Feature Discovery labels.
Q5. Which storage access mode is required when multiple workbenches need to read and write to the same training data volume?
Answer: D. ReadWriteMany
ReadWriteMany (RWX) is the required access mode when multiple workbenches need simultaneous read and write access to the same volume. ReadWriteOnce restricts access to a single node, which fails for concurrent shared writes.
Q6. Which metric helps administrators identify underutilized workbench instances across the OpenShift AI platform?
Answer: B. Idle notebook duration tracking metrics available via Prometheus dashboards
Tracking idle notebook duration metrics through Prometheus helps administrators spot underutilized workbenches for cost optimization. Focus on user activity metrics rather than infrastructure rates like network packets or fragmentation.
Q7. What is the default monitoring stack component used to scrape metrics from Red Hat OpenShift AI operators?
Answer: D. Prometheus
Prometheus is the default monitoring stack component used to scrape metrics from Red Hat OpenShift AI operators. The platform relies on the built-in OpenShift monitoring stack rather than third-party logging tools.
Q8. What happens to active model inference endpoints if the TrustyAI service pod experiences a brief internal application crash?
Answer: C. The model continues serving predictions normally but payload logging metrics collection pauses temporarily
The model continues serving predictions normally, but payload logging pauses. TrustyAI operates alongside the runtime rather than controlling it, meaning inference pods are unaffected by its downtime, avoiding total prediction failure.
Q9. WHAT COMMAND VALIDATES THE LIVE STATUS OF THE OPENSHIFT SERVERLESS DEPENDENCY SYSTEMS CRD?
Answer: A. oc get knativeserving –all-namespaces
The oc get knativeserving –all-namespaces command validates the live status of the required Serverless dependency. You must verify KnativeServing deployments before troubleshooting model serving issues, since KServe relies on this completely.
Q10. What component in the stack intercepts incoming traffic routing for single-model requests before passing data forward to model infrastructures?
Answer: C. Istio Ingress Gateway managed by Service Mesh Operator
The Istio Ingress Gateway managed by the Service Mesh Operator intercepts incoming routing for single-model requests. KServe relies on this gateway to handle traffic before passing it to the model runtime infrastructure.
Q11. WHICH METHOD BEST CONFIGURATES PERSISTENT REPOSITORIES THAT CAN RE-USE NOTEBOOK CONTENT UPON POD RESTARTS?
Answer: B. Attaching a PersistentVolumeClaim to the Workbench profile
Attaching a PersistentVolumeClaim to the workbench profile configures persistent storage that retains notebook content upon pod restarts. Memory-cached volumes like emptyDir are erased when pods stop, making them unsuitable for persistence.
Q12. Which system deploys the Knative Serving controller for single-model serving?
Answer: C. OpenShift Serverless Operator
The OpenShift Serverless Operator is responsible for installing and managing Knative Serving, which provides the scalable infrastructure required for single-model serving. While the OpenShift AI operator configures the dashboards and custom resources, Serverless actually runs the Knative controllers.
Q13. What metric is used by the Knative Pod Autoscaler to trigger scaling actions for single-model serving deployments?
Answer: C. Concurrency / number of concurrent requests per pod
The Knative Pod Autoscaler primarily uses concurrency, meaning the number of concurrent requests received per pod, to determine when to scale up or down. Standard infrastructure metrics like storage space or git commits are not used by Knative for application scaling.
Q14. Which user group configuration is the standard administrative choice to grant a group of users complete access to manage all RHOAI settings?
Answer: A. rhoai-admins or dedicated cluster administrator group bindings
The rhoai-admins group grants complete administrative access to manage all OpenShift AI settings. Viewers only have read-only access, while standard authenticated users lack the necessary privileges to modify dashboard configurations or manage platform components.
Q15. Which custom resource must be edited to add cluster-wide cluster-admin equivalents for OpenShift AI dashboard administration?
Answer: A. OdhDashboardConfig
The OdhDashboardConfig custom resource manages dashboard administrative settings, including cluster administrator group configurations. DataScienceCluster controls component installations, while the other options do not exist as standard custom resources in the current OpenShift AI environment.
Q16. What configuration allows an administrative operator to prevent users from requesting more than 4 GPUs for any single workbench?
Answer: B. OpenShift Cluster ClusterResourceQuota or Namespace-level ResourceQuota specifications
ClusterResourceQuota or Namespace-level ResourceQuota specifications limit hardware usage like GPUs per namespace or project. Changing operator subscriptions or deleting operators breaks functionality, whereas quotas safely enforce resource boundaries for user workbench requests.
Q17. Which configuration strategy enables teams to integrate pre-baked enterprise packages into local workbench image option menus?
Answer: B. Adding Custom Notebook Images into Dashboard settings configurations targeting ImageStreams
Custom workbench environments are integrated by creating custom notebook images and registering them via the dashboard settings. The dashboard relies on OpenShift ImageStreams to populate the notebook launcher menu for users.
Q18. Which resource type tracks the cluster-wide operational availability of the RHOAI dependencies?
Answer: C. ClusterServiceVersion custom resource tracking status index
ClusterServiceVersion resources represent the installed operators and track their operational status and health. KServe or ServiceMesh resources define specific serving or routing configurations rather than tracking the global operator dependencies.
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.