For a StorageGRID node running as a VMware virtual machine, NetApp requires the node to be gracefully shut down before hypervisor migration or maintenance. The supported migration approach is therefore a cold migration, in which the StorageGRID node is stopped before it is moved to another ESXi host. NetApp documentation explicitly states that cold migration is supported.

Option A is specifically unsupported. NetApp warns that VMware live vMotion can cause the virtual machine clock to jump. StorageGRID depends heavily on synchronized time for distributed database operations, object metadata, configuration updates, and inter-node coordination. A time discontinuity can therefore lead to data or configuration integrity problems, which is why live vMotion must not be used for StorageGRID grid nodes.

Option B is incorrect because StorageGRID appliance maintenance mode is intended for specific appliance maintenance procedures; it is not the prescribed migration workflow for a VMware-hosted software Storage Node. Option C is also not the required host-maintenance procedure and unnecessarily introduces datastore-level movement considerations.

Thus, the correct sequence is to gracefully shut down the StorageGRID VM and perform a cold migration to another ESXi host, preserving grid integrity while host maintenance is performed.

Reference topics: VMware Deployment → Hypervisor Maintenance → StorageGRID Node Shutdown → VMware Migration Restrictions → Cold Migration

The appropriate consistency level is strong-global because the requirement explicitly applies to clients and object metadata across multiple StorageGRID sites. NetApp defines strong-global as providing read-after-write consistency for all client requests across all sites. Therefore, after an object is created or its metadata is changed, subsequent requests from any site receive a consistent view of that change.

strong-site is insufficient because its consistency guarantee applies only to client requests within a single site. A client accessing the bucket from another StorageGRID site could therefore fall outside the required site-wide consistency guarantee. read-after-new-write, which is the default and generally recommended setting, guarantees consistency for newly created objects but provides eventual consistency for object updates. Because the scenario requires immediate consistency following both new writes and changes, it does not meet the stated requirement. available provides eventual consistency for both new objects and updates and prioritizes availability rather than cross-site consistency.

Thus, among the available options, strong-global is specifically designed for workloads requiring globally consistent object and metadata visibility across a multisite StorageGRID deployment.

Reference topics: S3 Buckets → Bucket Consistency → Strong-global → Strong-site → Read-after-new-write → Metadata Distribution Across Sites

StorageGRID Platform Services traffic, including Apache Kafka event notifications, originates from Storage Nodes that run the Administrative Domain Controller (ADC) service. NetApp explicitly requires these ADC Storage Nodes to have network connectivity to external Platform Services endpoints. For Kafka, StorageGRID normally establishes outbound TCP connectivity from ADC Storage Nodes to the Kafka cluster, using port 9092 by default unless another port is configured.

In this topology, the Storage Nodes do not have Client Network interfaces, so they cannot use a Client Network default route to reach the Kafka endpoint. NetApp's supported alternative is to configure the Storage Nodes with Admin Networks containing appropriate Admin External Subnet List (AESL) entries that provide routes to the external endpoint network. Without the required AESL route on the ADC nodes, StorageGRID cannot validate connectivity to the Kafka endpoint.

Option B incorrectly identifies LDR as the service responsible for sending Platform Services messages. Option D is irrelevant because Gateway Nodes do not originate these Kafka Platform Services notifications. Although Platform Services must be enabled for the tenant, the networking details in the scenario specifically identify missing ADC-node endpoint routing as the intended failure.

Reference topics: Platform Services → Kafka Event Notifications → ADC Service → Admin Network → AESL → External Endpoint Connectivity

The correct approach is to create three separate S3 tenant accounts, one for each business unit. StorageGRID tenant accounts are designed as administrative and data-isolation boundaries. NetApp states that each tenant account has its own federated or local groups, users, S3 buckets, and objects, allowing different organizational entities to manage their own resources independently. NetApp specifically cites an enterprise use case in which separate tenant accounts are created for different departments such as Marketing, Human Resources, and Customer Support.

This design also provides delegated administration. After the grid administrator creates each tenant, authorized users for that tenant can use Tenant Manager to manage their own groups, users, S3 access keys, and buckets without gaining administrative control over resources belonging to another tenant.

Option A can segregate S3 resources with bucket policies, and NetApp documents that as a valid alternative where separate tenants are unnecessary. However, it does not provide the same strict administrative separation required here. Options B and C similarly place all business units within one tenant and therefore share the same tenant-level administrative boundary.

Therefore, D best satisfies both strict isolation and independent management.

Reference topics: Tenant Accounts → Enterprise Multitenancy → Tenant Manager → Groups and Users → S3 Buckets → Administrative Isolation

The bank-records bucket requires indefinite immutable retention, so its ILM rule must comply with StorageGRID S3 Object Lock requirements. NetApp specifies that a compliant ILM rule must create at least two replicated copies or one erasure-coded copy, those copies must reside on Storage Nodes, at least one placement must begin at day 0 using Ingest time, and at least one placement must continue forever. Critically, copies governed by a compliant S3 Object Lock rule cannot be stored in a Cloud Storage Pool.

This requirement directly eliminates A, B, and C because every one of those choices introduces a Cloud Storage Pool. The organization's general preference to use cloud storage for long-term data does not override StorageGRID compliance restrictions for Object Lock-protected content.

NetApp's own bank-records example uses precisely this design: the bucket-specific compliant rule retains an erasure-coded copy on Storage Nodes from day 0 to forever.

Therefore, D is correct. Other noncompliant buckets can still use separate ILM rules that transition data to Cloud Storage Pools after six months, but immutable bank-records objects must remain on compliant StorageGRID Storage Node placements indefinitely.

Reference topics: ILM → S3 Object Lock → Compliant ILM Rules → Storage Nodes → Forever Retention → Cloud Storage Pool Restrictions

NetApp defines a specific workflow for migrating a Linux-based StorageGRID node between hosts. The source node must first be shut down cleanly using storagegrid node stop. After the node has stopped, the administrator runs storagegrid node export on the source host. The export output identifies the /var/local storage volume and provides the corresponding import command to use on the target host.

On Host-B, the administrator runs storagegrid node import against the exported node storage. NetApp then requires the imported configuration to be checked with storagegrid node validate. After successful validation, the migrated node is started with storagegrid node start, and Grid Manager is used to confirm that the node returns to a green operational state with no alerts.

Therefore, C represents the required migration sequence. Although the option summarizes the final validation/start phase, its stop → export → import → verify ordering matches NetApp's prescribed procedure.

Option A is incorrect because Grid Manager does not automatically migrate Linux containers. Option B reverses the export/import sequence. Option D incorrectly treats migration as a new StorageGRID installation and node reconfiguration.

Reference topics: StorageGRID Maintenance → Linux Host Migration → Node Stop → Node Export → Node Import → Validate → Start → Grid Manager Verification

For unidirectional cross-grid replication, CGR must be enabled on the bucket in the grid that acts as the source. NetApp states that when CGR is enabled on only one grid, objects added to that source bucket are replicated automatically to the corresponding destination bucket on the other grid; objects written to the destination are not replicated back. Because the required direction is Grid-1 → Grid-2, CGR must therefore be enabled on Grid-1.

The administrator then selects the appropriate grid federation connection and sets Replicate delete markers = Yes. When an S3 client deletes an object without specifying a version ID, StorageGRID creates a delete marker. With this option enabled, that marker is replicated to Grid-2 so the object appears deleted on both grids, helping maintain matching bucket state. NetApp documents this as the default behavior.

Option A would configure bidirectional replication, which violates the unidirectional requirement. Option B reverses the required replication direction. Option D is incorrect because the relevant CGR configuration control is Replicate delete markers, not “Replicate metadata.”

Reference topics: Tenant Manager → Grid Federation → Cross-Grid Replication → Source and Destination Buckets → Versioning → Replicate Delete Markers

For the SG5860, which uses the SG5800 controller, StorageGRID supports both Fixed and Aggregate port-bond configurations. In Fixed mode, ports 2 and 4 are paired for the Grid Network and ports 1 and 3 for the Client Network. Active-Backup is the default network bond mode, and VLAN tagging is optional; therefore, the two-port Active-Backup configuration in A is supported.

In Aggregate port bond mode, all four 10/25-GbE ports form a single LACP (802.3ad) bond carrying both Grid and Client Network traffic. Because both networks share the aggregate, StorageGRID requires a unique VLAN tag for each network, making C the supported aggregate configuration. A four-port Active-Backup aggregate is not supported, eliminating B.

Verification note: the current NetApp SG5800 appliance documentation also permits two-port LACP bonds in Fixed mode, with VLAN tagging optional. Consequently, option D is technically supported as written, making this question ambiguous against current documentation. If restricted to the intended two-answer certification key representing the standard Fixed Active-Backup and Aggregate LACP configurations, A and C are the expected selections.

Reference topics: SG5800/SG5860 Appliance Networking → Fixed Port Bond Mode → Aggregate Port Bond Mode → Active-Backup → LACP → VLAN Tagging