This scenario involves a VCF multi-AZ design where AZs must operate independently (no shared dependencies) and achieve a 99.9% availability SLA (allowing ~8.76 hours of downtime annually). The design decisions must ensure resilience, fault isolation, and recovery capabilities across AZs.
Requirement Analysis:
Independent AZ operation:Each AZ must function standalone, with no single point of failure or dependency across AZs.
99.9% availability:The design must minimize downtime through redundancy, replication, and recovery mechanisms.
Option Analysis:
A. Configure array-based replication between the selected AZ(s) for the management domain:Array-based replication (e.g., vSphere Replication or SAN replication) for the management domain (vCenter, NSX Manager, SDDC Manager) ensures that critical management VMs are duplicated across AZs. If one AZ fails, the other can take over with minimal downtime, supporting independent operation and high availability. The VCF 5.2 Design Guide recommends replication for multi-AZ deployments to meet SLAs, as it provides a recovery point objective (RPO) near zero. This option enhances availability and is correct.
B. Make sure all configuration backups are replicated between the selected AZ(s):Replicating configuration backups (e.g., SDDC Manager backups, NSX configurations) ensures that each AZ has access to recovery data. If an AZ’s management components fail, the other AZ can restore operations independently using its local backup copy. This supports the independence requirement and reduces downtime (contributing to 99.9% SLA) by enabling quick recovery. The VCF Administration Guide emphasizes backup replication for multi-AZ resilience, making this option correct.
C. Make sure the recovery VLAN for the infrastructure management components has access to both AZ(s):A recovery VLAN spanning both AZs implies a shared network dependency. If this VLAN fails (e.g., due to a network outage), both AZs could be impacted, violating the independence requirement. Multi-AZ designs in VCF favor isolated networks per AZ to avoid cross-AZ single points of failure. The VCF Design Guide advises against shared VLANs for critical components in independent AZ setups. This option undermines the requirements and is incorrect.
D. Choose two distant AZ(s) and consider each AZ the DR for the other:Distant AZs (e.g., separate data centers) with mutual DR (disaster recovery) roles enhance geographic fault tolerance. However, “operate independently” in VCF typically means each AZ can run workloads standalone, not that one is a passive DR site. Distant AZs introduce latency, complicating synchronous replication needed for 99.9% availability, and may rely on shared management, conflicting with independence. The VCF Multi-AZ Guide focuses on active-active AZs, not DR-centric designs, making this less suitable.
E. Choose two close proximity AZ(s) and configure a stretched management workload domain:A stretched management domain (e.g., using vSAN stretched cluster) spans AZs with synchronous replication, ensuring high availability. However, this creates a dependency: both AZs share the same vCenter and management stack, so a failure (e.g., vCenter outage) could affect both, violating independence. The VCF 5.2 Design Guide notes stretched clusters are for single logical domains, not independent AZs. This option contradicts the requirement and is incorrect.
F. Configure a non-routable separate recovery VLAN for the infrastructure management components within each AZ:A non-routable, AZ-specific recovery VLAN isolates management recovery traffic (e.g., for vMotion, backups) within each AZ. This ensures that each AZ’s management components operate independently, with no cross-AZ network reliance. If one AZ’s network fails, the other remains unaffected, supporting the SLA through fault isolation. The VCF Multi-AZ Design Guide recommends separate, isolated networks per AZ for resilience, making this option correct.
Conclusion:The three design decisions areConfigure array-based replication between the selected AZ(s) for the management domain (A),Make sure all configuration backups are replicated between the selected AZ(s) (B), andConfigure a non-routable separate recovery VLAN for the infrastructure management components within each AZ (F). These ensure independent operation and meet the 99.9% SLA through replication and isolation.
References:
VMware Cloud Foundation 5.2 Design Guide (Section: Multi-AZ Design)
VMware Cloud Foundation 5.2 Administration Guide (Section: Backup and Recovery)
VMware Cloud Foundation Multi-AZ Deployment Guide (Section: Networking)
VMware vSphere 8.0 Update 3 Documentation (Section: vSAN Stretched Clusters)
Submit