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: What is the primary advantage of choosing Single-Model Serving over Multi-Model Serving configurations in OpenShift AI?. 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 primary advantage of choosing Single-Model Serving over Multi-Model Serving configurations in Ope…
- What component allows OpenShift AI to dynamically provision storage when a data scientist requests a new note…
- Which parameter defines the location of the model file within an S3 data connection during deployment?
- Which type of database is utilized internally by the pipeline server to track execution metadata?
- What user group is granted full administrative control over the OpenShift AI Operator parameters by default c…
- What is the default namespace where Red Hat OpenShift AI applications and deployments are installed?
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 primary advantage of choosing Single-Model Serving over Multi-Model Serving configurations in OpenShift AI?
Answer: A. Support for advanced serverless features like scale-to-zero and dedicated resource isolation
Single-model serving provides dedicated resource isolation and supports advanced serverless features like scale-to-zero. Multi-model serving optimizes memory by bundling models, but it lacks these specific serverless scaling capabilities.
Q2. What component allows OpenShift AI to dynamically provision storage when a data scientist requests a new notebook workspace?
Answer: C. Container Storage Interface (CSI) drivers via StorageClasses
Container Storage Interface drivers provision the persistent volumes requested through StorageClasses. Note that local ephemeral drivers only provide temporary storage that vanishes when the workspace stops.
Q3. Which parameter defines the location of the model file within an S3 data connection during deployment?
Answer: B. Path or folder prefix inside the bucket
The path or folder prefix inside the bucket specifies the exact storage location of the model file during deployment. Do not confuse this with IAM policies or checksums, which handle access and integrity rather than file location.
Q4. Which type of database is utilized internally by the pipeline server to track execution metadata?
Answer: D. MariaDB / MySQL or PostgreSQL
The OpenShift AI pipeline server, utilizing ArgoDB, uses a PostgreSQL database internally to track execution metadata. Options like Redis or MongoDB are incorrect because the underlying Data Science Pipelines engine expects a relational database.
Q5. What user group is granted full administrative control over the OpenShift AI Operator parameters by default configuration conventions?
Answer: C. rhoai-admins
By default, OpenShift AI assigns full administrative privileges to the dedicated admin group. While standard OpenShift cluster administrators naturally inherit permissions, the specific product configuration relies heavily on this dedicated group.
Q6. What is the default namespace where Red Hat OpenShift AI applications and deployments are installed?
Answer: D. redhat-ods-applications
The redhat-ods-applications namespace hosts deployed models and pipelines by default. The redhat-ods-operator namespace exists, but operators are managed separately from applications.
Q7. Which protocol is typically used to send inference requests to a model deployed via Multi-Model Serving?
Answer: B. gRPC or HTTP/REST
Models deployed via OpenShift AI serving typically accept standard HTTP or gRPC inference requests. Watch for older messaging protocols like WebSockets or SOAP, which are not used for standard REST model endpoints.
Q8. Which technology is used to define the operational workflow graph steps in an Elyra pipeline?
Answer: C. JSON-based visual node topologies compiled into Tekton pipeline execution schemas
Elyra uses a JSON-based visual node topology that gets compiled into Tekton pipeline schemas for execution. Distractors like XML workflows or bash scripts do not represent the underlying architecture of OpenShift Pipelines.
Q9. Which component provides serverless scaling capabilities, allowing single-model serving pods to scale down to zero when idle?
Answer: A. Knative Serving
Knative Serving provides serverless scaling capabilities, allowing single-model serving pods to scale to zero when idle. ModelMesh handles high-density multi-model serving and does not scale to zero natively.
Q10. Which tool does OpenShift AI use to coordinate distributed model training over multiple pods?
Answer: B. Ray / CodeFlare
OpenShift AI integrates the CodeFlare stack and leverages Ray to coordinate and scale distributed training or tuning tasks across multiple compute nodes. While Kubeflow handles TensorFlow and PyTorch job operators, CodeFlare orchestrates the actual distributed workloads.
Q11. Which network protocol does OpenShift AI use by default to secure traffic to the dashboard?
Answer: B. HTTPS
OpenShift AI secures dashboard traffic by default using HTTPS to encrypt client connections. HTTP lacks encryption, while SSH and raw TLS/SNI are not the protocols used by web browsers to access the OpenShift AI interface.
Q12. Which OpenShift mechanism scales down idle workbench pods to optimize resource usage?
Answer: B. OpenShift AI Hibernation / Idling feature
OpenShift AI includes a dedicated workbench hibernation feature to stop idle pods and optimize resource usage. Horizontal Pod Autoscalers distribute load but do not automatically hibernated user notebooks.
Q13. What file extension is used for pipelines created via the visual editor in OpenShift AI workbenches?
Answer: D. .pipeline
Pipelines created via the Elyra visual pipeline editor in workbenches use the .pipeline file extension to define their Directed Acyclic Graph. While execution generates YAML, the editor format itself relies on this proprietary structure.
Q14. Which token format does the Elyra pipeline editor use to securely connect and authenticate against the pipeline server API?
Answer: B. OpenShift login service account authentication token
Elyra uses an OpenShift service account authentication token to securely connect and authenticate against the pipeline server API. Hardcoded passwords or external PGP files are not supported mechanisms for runtime pipeline authentication.
Q15. Which namespace is the primary default location where the Red Hat OpenShift AI operator controller pods are deployed?
Answer: A. openshift-operators or redhat-ods-operator
The Red Hat OpenShift AI operator typically installs in the openshift-operators or redhat-ods-operator namespace. Core components are deployed here, ensuring operators avoid critical system namespaces like kube-system.
Q16. Which API version is commonly used in 2026 to define the DataScienceCluster custom resource?
Answer: C. datasciencecluster.opendatahub.io/v1
The DataScienceCluster resource uses the datasciencecluster.opendatahub.io/v1 API version in current OpenShift AI releases. Earlier beta or alpha API versions are no longer provisioned, so standardizing on the stable v1 endpoint ensures the operator reconciles components correctly.
Q17. Which component provides the web user interface for Red Hat OpenShift AI?
Answer: D. OdhDashboard
The OdhDashboard component provides the dedicated web user interface for Red Hat OpenShift AI. While OpenShift Web Console manages the underlying cluster resources, it is not the primary data science interface.
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.