5V0-3221 Sample Questions & Answers

The requirement to extend a Layer 2 network segment specifically points to the need for an L2 VPN. This service, configured on the tenant's Edge Gateway (backed by NSX), allows for the stretching of a broadcast domain between the on-premises environment and the Cloud Director VDC. This facilitates VM migration without changing IP addresses. A standard IPsec VPN provides L3 connectivity but does not extend the L2 segment.

The correct method for sharing catalogs between different VMware Cloud Director instances is the publish and subscribe model. The source catalog (New York) is published, creating a shareable URL. The destination instance (London) then creates a subscribed catalog pointing to that URL. This model periodically synchronizes the catalog content, storing a local copy in the London instance. This ensures that when London tenants deploy VMs from the catalog, the data is served locally, avoiding performance issues caused by the high-latency WAN link.

The recommended procedure for a failed cell upgrade (when the primary has succeeded) is to not roll back the entire environment. The primary cell and database have already been modified. The correct action is to revert only the failed cell to its snapshot, review its specific upgrade logs (e.g., in /opt/vmware/vcloud-director/logs/) to diagnose the failure, resolve the underlying issue, and then try upgrading that cell again. Reverting the entire cluster is a last resort and not the immediate next step.

To achieve this, the administrator must make both underlying storage policies (one for each datastore) available to the Provider VDC. Then, by setting the policy corresponding to 'Hybrid-DS' as the default, any VM provisioned without an explicit storage policy selection will be placed on that datastore. The 'All-Flash-DS' policy is still available to be selected by tenants for specific VMs, fulfilling the premium offering requirement.

The valid allocation models for an Organization VDC are Pay-As-You-Go, Allocation Pool, Reservation Pool, and Flex. 'On-Demand' and 'Burstable' are not standard VCD allocation model names. Pay-As-You-Go allocates resources on demand, Reservation Pool guarantees resources, and Flex provides a hybrid approach.

The ability for tenants to manage the Distributed Firewall (DFW) for east-west traffic is enabled through Data Center Groups. A provider administrator must create a Data Center Group, add one or more Organization VDCs to it, and then explicitly activate the Distributed Firewall service on that group. Once activated, tenants belonging to the member VDCs can define and manage DFW rules via the Tenant Portal.

The configuration of the shared NFS mount point, which serves as the transfer server storage for VMware Cloud Director, is a core appliance setting. This is managed through the VMware Appliance Management Interface (VAMI), which is accessible via a web browser on port 5480 of any cell appliance. This is where the administrator specifies the NFS server's IP or hostname and the exported path.

In the Pay-As-You-Go allocation model, CPU and Memory resources are only consumed from the underlying Provider VDC when a VM is powered on. For powered-off VMs, there is no reservation or consumption of these compute resources. This model provides the highest consolidation ratio as it relies on overcommitment, but offers no performance guarantees. Storage is the only resource consumed by powered-off VMs in this model.

This case study describes a classic use case for the VMware Cloud Director catalog publish and subscribe feature. Publishing the catalog in Frankfurt makes its contents available via a secure, encrypted URL. Subscribing to this catalog in Dublin creates a local copy of all vApp templates. This process can be scheduled to run automatically, ensuring the DR site always has the latest versions. Because the templates are stored locally in Dublin, they can be deployed quickly during a DR event without relying on the WAN link. This meets the requirements for automation, local storage for performance, and encrypted transport.

The 'Bridged' fence mode connects a vApp network directly to an Organization VDC network, effectively acting as a Layer 2 bridge. VMs in the vApp network are on the same subnet as the Org VDC network and can communicate without routing. This is distinct from 'Direct' which connects to an External Network, 'Isolated' which has no external connectivity, and 'NAT-Routed' which places the vApp behind its own routing and NAT service.