A Platform Engineer is configuring Kubernetes Admin Credentials for a break-glass scenario. The requirement is to enable the built-in admin user for a specific TKG cluster prod-cluster, bypassing vCenter SSO in case of an SSO outage.
Which sequence of commands/actions correctly retrieves this kubeconfig? (Choose 2.)

• A.SSH into the Supervisor Control Plane VM and copy /etc/kubernetes/admin.conf to the local machine.
• B.Inspect the TanzuKubernetesCluster resource status fields for the adminKubeconfig data.
• C.Run kubectl get secret prod-cluster-kubeconfig -n -o jsonpath='{.data.value}' | base64 -d > admin.kubeconfig.
• D.Use the kubectl vsphere login command with the --tanzu-kubernetes-cluster-name and --kubeconfig flags, referencing the SSO administrator credentials.
• E.The system does not generate a static admin kubeconfig for TKG clusters; all access must be via SSO.

Comment: *

Name: *

Email: *

Verification: *

A Cloud Architect is designing a disaster recovery plan for a mission-critical Zonal Supervisor deployment. The scenario involves a catastrophic failure of the Supervisor Cluster itself (e.g., corruption of the etcd database across all zones) during a failed upgrade, requiring a full restore.
Environment:
* VKS workloads are backed up using Velero.
* The Supervisor configuration (Namespaces, Policies) is backed up using the vCenter File-Based Backup.
What is the correct sequence of steps to restore service? (Select all that apply.)

• A.After the Supervisor is restored, use Velero to restore the TKG workloads (TanzuKubernetesClusters) into the restored Namespaces.
• B.Manually recreate all vSphere Namespaces and re-assign permissions before restoring workloads.
• C.Since the Supervisor state (etcd) is corrupted, the administrator must restore the Supervisor Cluster using the specific "Restore Supervisor" workflow (often involving the restore-supervisor.sh script or UI equivalent if available in the specific VCF version) which utilizes the backup data to reconstruct the Control Plane VMs.
• D.It is unnecessary to restore the Supervisor; simply redeploying a new Supervisor and pointing Velero to the object store will automatically recover the cluster infrastructure.
• E.Restore the vCenter Server from its file-based backup to recover the Supervisor's management context.

Comment: *

Name: *

Email: *

Verification: *

A VI Administrator needs to expand the available storage options for a specific vSphere Namespace data-science. The team requires a new storage tier for temporary scratch data that does not need replication (RAID-0).
The administrator has created a new VM Storage Policy named scratch-policy in vCenter that defines these non-replicated rules.
What is the correct procedure to make this new policy available to the DevOps team's kubectl context within that namespace? (Select all that apply.)

• A.Log in to the vSphere Client, navigate to Workload Management, select the Namespace, go to the Storage card, and click Edit Storage Policies to add scratch-policy.
• B.Instruct the DevOps team to create a StorageClass YAML with the provisioner:
  csi.vsphere.vmware.com and parameters.storagepolicyname: scratch-policy.
• C.No action is needed; all vCenter storage policies are automatically available to all namespaces as StorageClasses.
• D.Once added in the vSphere Client, verify the availability by running kubectl get storageclass within the data-science namespace context; a new class named scratch-policy should appear automatically.
• E.Log in to the Supervisor Control Plane via SSH and edit the /etc/kubernetes/manifests/kube- controller-manager.yaml file to include the new policy.

Comment: *

Name: *

Email: *

Verification: *

A Platform Engineer is designing an auto-scaling strategy for a cluster hosting a machine learning workload. The workload creates transient "Job" pods that require significant GPU resources.
Requirements:
1. The cluster usually runs with 2 worker nodes.
2. When a ML job starts, it might request 10+ pods, each needing a full GPU.
3. The cluster must scale up to satisfy these requests and scale down to 0 or 2 when idle to save costs.
4. There is a limited number of GPU-capable hosts in the vSphere cluster.
Which design considerations are critical for the correct functioning of the Cluster Autoscaler in this scenario? (Choose 2.)

• A.The TanzuKubernetesCluster YAML must define a separate worker node pool specifically using a GPU-enabled VM Class (e.g., gpu-large).
• B.The workload Pods must use podAntiAffinity to ensure they are scheduled on the Control Plane nodes.
• C.The min-size for the GPU node pool should be set to 0 (Zero) to allow scaling from zero, assuming the VKS version supports scale-from-zero behavior.
• D.The vSphere Namespace must have a resource quota that allows for the maximum potential CPU/Memory/GPU usage of the scaled-out cluster (e.g., sufficient for 12+ nodes).
• E.The Cluster Autoscaler must be configured to ignore GPU resource requests because vSphere handles GPU assignment via DRS.

Comment: *

Name: *

Email: *

Verification: *

A VI Administrator is managing the disk space consumed by the TKR Content Library. Several old versions of Kubernetes (e.g., v1.18, v1.19) are no longer needed.
The administrator deletes the specific OVA files for these versions from the Content Library using the vSphere Client.
However, when developers run kubectl get tkr, the deleted versions are still listed (though with COMPATIBLE: False or Ready: False).
What additional step is required to fully remove these versions from the Kubernetes view? (Choose 2.)

• A.The Supervisor automatically reconciles the library state; the TKR objects will eventually be garbage collected, but the sync interval might be long.
• B.Delete the VirtualMachineImage objects corresponding to the OVAs.
• C.The administrator must manually delete the TanzuKubernetesRelease Custom Resources using kubectl delete tkr to remove the metadata immediately.
• D.Remove the Content Library reference from the Namespace and re-add it.
• E.Restart the wcp-content-library-service on the Supervisor.

Comment: *

Name: *

Email: *

Verification: *