Scenario Overview - On-Prem Runtime Images and Storage Foundation

This scenario helps operators configure the image registry, runtime images, and block, file, or object storage after onboarding a local cluster, creating a reusable foundation for online IDEs, training jobs, and inference services.

Applicable Roles

• Platform Operators, on-prem compute administrators, and storage administrators

Scenario Goals

• Make the image service reachable from the platform and target cluster and keep synchronization healthy.
• Give runtime images explicit versions, hardware environments, and purpose labels.
• Connect block, file, and object storage to the correct region according to workload needs.
• Validate that a test workload can pull the image and create, mount, or access storage.

Scenario Flow

Main path: Confirm cluster and network -> Register image service -> Synchronize runtime images -> Register storage components -> Bind the region -> Validate with a test workload

Stage	Key Result
1. Confirm dependencies	Cluster, network, registry, and storage services are reachable
2. Prepare images	Image service is healthy and the versioned target image is visible
3. Prepare storage	Required storage components are registered and bound to the target region
4. Validate together	A workload can pull the image and use storage as intended

Before You Start

• Complete On-Prem Compute Onboarding.
• Prepare the image registry, storage services, network, and authentication materials.
• Define the target region, cluster, tenant isolation, capacity, and reclaim policies.
• Enter credentials only in the platform configuration page, never in documentation or tickets.

Recommended Reading Order

1. Register the image service and validate synchronization.
2. Synchronize or upload runtime images with explicit version tags.
3. Select block, file, or object storage according to workload needs.
4. Bind the region and run an end-to-end test workload.

Document Index

Document	Description
Runtime Image and Storage Foundation Workflow	Complete steps from registry onboarding through storage and test-workload validation

Related Scenarios

• On-Prem Development, Training, and Assets: use runtime images and shared storage for development and training.
• On-Prem Inference Template Building: reference available images, models, and resource specifications in templates.
• On-Prem Model Deployment and Status Check: validate image pulls, volume mounts, and instance state.

Completion Checklist

Purpose: These are the scenario exit criteria. They confirm that the target cluster can actually use the runtime environment and data storage, rather than merely showing configuration records in lists.

Check	Pass Criteria
1	Image service state is healthy and it is bound to the correct region or cluster.
2	The target runtime image is visible with a clear version, architecture, hardware environment, and purpose.
3	Required storage state, capacity, access policy, and region binding are correct.
4	A test workload can pull the image and create, mount, read/write, or access storage.
5	Owners for deletion, reclaim, backup, and credential rotation are identified.