Microsoft AZ-802: Implement Windows Server High Availability

Topic snapshot

Sample questions

These questions are original IT Mastery practice items aligned to this topic area. They are designed for self-assessment and are not official exam questions.

Question 1

Topic: Implement Windows Server High Availability

You administer a four-node Windows Server failover cluster that uses Storage Spaces Direct. Node S2D-N3 is offline after a system board failure. The vendor confirms that the node will not be repaired, and its local disks will be discarded. Clustered roles are running on the other nodes, and no storage repair jobs are pending. A replacement server, S2D-N5, has supported hardware and a matching OS/patch level. Which action plan should you perform next?

Options:

• A. Clear the cluster configuration on S2D-N3 only.

• B. Run storage repair on S2D-N3 and resume the node.

• C. Add S2D-N5 before removing S2D-N3 from the cluster.

• D. Evict S2D-N3, validate S2D-N5, then add it to the cluster.

Best answer: D

Explanation: For a failed cluster node that will not return, the correct recovery path is replacement, not repair. Because S2D-N3 is permanently unavailable and its disks are being discarded, remove the failed node from the cluster membership first, typically from a healthy node by using Failover Cluster Manager or Remove-ClusterNode -Force. After the failed member is evicted, run cluster validation that includes the replacement server and then add the validated server to the cluster. In a Storage Spaces Direct cluster, this keeps cluster membership accurate before storage is rebalanced and the replacement contributes capacity.

• Repairing the failed node does not fit because the hardware will not be repaired and the disks will be discarded.
• Adding the replacement first leaves the permanently failed node as an active cluster member and does not complete the replacement workflow.
• Clearing local cluster configuration is for cleanup on an accessible node and does not remove the failed node from the running cluster.

Question 2

Topic: Implement Windows Server High Availability

An organization is replacing a small SAN with two on-premises Windows Server 2022 Datacenter servers that are certified for Storage Spaces Direct. A SQL Server failover cluster instance will run on the cluster and must use WSFC-managed shared storage while tolerating the failure of one node or one drive. Which storage design is the best fit?

Options:

• A. Use Storage Replica between standalone data volumes

• B. Enable Storage Spaces Direct and create mirrored cluster volumes

• C. Replicate local volumes by using DFS Replication

• D. Place the database files on an SMB share from one node

Best answer: B

Explanation: Storage Spaces Direct is the best fit when a failover cluster needs resilient shared storage but no external SAN is available. It pools the local drives in the cluster nodes and creates resilient virtual disks that can be exposed as Cluster Shared Volumes or cluster disks. In this scenario, the servers are certified for Storage Spaces Direct, run Windows Server Datacenter, and must tolerate a node or drive failure, so mirrored Storage Spaces Direct volumes align with the workload and resiliency requirements. A quorum witness may still be needed for the two-node cluster, but it does not replace shared storage. Replication or a normal file share does not provide the same WSFC-managed shared-disk semantics for a failover cluster instance.

• DFS Replication keeps folders synchronized but does not create cluster disks or provide coordinated shared storage.
• Single-node SMB share introduces a dependency on one node and is not a resilient cluster disk design.
• Storage Replica is for volume replication and disaster recovery patterns, not for creating shared storage from local drives in this scenario.

Question 3

Topic: Implement Windows Server High Availability

A company is deploying a new three-node Windows Server failover cluster for Hyper-V and Storage Spaces Direct. Each node has the same physical NIC layout. The operations team wants repeatable host-network configuration for management, compute, and storage traffic, and wants the cluster to detect drift from the intended configuration. Which design is the best fit?

Options:

• A. Use Cluster-Aware Updating to configure NICs

• B. Create SET switches manually on each node

• C. Use Network ATC with cluster network intents

• D. Apply Group Policy firewall rules to cluster nodes

Best answer: C

Explanation: Network ATC is designed for consistent, intent-based host-network deployment on Windows Server clusters. For a Hyper-V and Storage Spaces Direct cluster with matching adapters, you can define network intents such as management, compute, and storage, and Network ATC applies the required host networking configuration across the cluster nodes. This reduces manual variation in virtual switch, QoS, and storage-network settings, and it supports ongoing validation so drift is visible. Manual switch creation can work technically, but it does not provide the same cluster-wide intent model or drift handling. The key design fit is using Network ATC when consistency and repeatability of cluster host networking are explicit requirements.

• Manual switches are more error-prone because each node must be configured and maintained consistently outside an intent model.
• Firewall Group Policy can standardize security rules but does not build the cluster host networking stack.
• Cluster-Aware Updating coordinates patching and reboots; it is not a NIC or virtual switch configuration framework.

Question 4

Topic: Implement Windows Server High Availability

A company runs a four-node Windows Server failover cluster for a critical workload. Two nodes are in Datacenter A and two nodes are in Datacenter B. The cluster currently uses node majority with no witness. Either datacenter can reach Azure Storage if the other datacenter fails. You must modify quorum so the workload can remain online after the simultaneous loss of either datacenter. Which design is the best fit?

Options:

• A. Configure node majority and rely on dynamic quorum.

• B. Configure a file share witness in Datacenter A.

• C. Remove quorum votes from the Datacenter B nodes.

• D. Configure an Azure Cloud Witness for quorum.

Best answer: D

Explanation: For an even-numbered multi-site failover cluster, an external witness is commonly used to avoid a split vote and preserve quorum after a site failure. With two nodes in each datacenter, losing either site leaves two node votes. Adding an Azure Cloud Witness creates an additional vote outside both datacenters, so the surviving site can form majority as long as it can reach Azure Storage. This matches the stated failure pattern without placing the witness in a datacenter that might be lost.

• Local witness placement fails because a file share witness in Datacenter A is unavailable if Datacenter A is lost.
• Dynamic quorum alone does not guarantee quorum after a simultaneous two-node site failure from a four-node cluster.
• Removing B votes makes the design asymmetric and prevents Datacenter B from maintaining availability if Datacenter A fails.

Question 5

Topic: Implement Windows Server High Availability

You are preparing to enable Storage Spaces Direct on a four-node Windows Server failover cluster. Cluster validation passes network and system checks, but the storage validation report shows this warning:

Storage Spaces Direct inventory
Node count: 4
Physical disks discovered: 24
Disk bus type reported: RAID
Eligible disks for S2D: 0
Result: No disks with supported bus types were found

What is the most likely root cause?

Options:

• A. The storage controller is presenting RAID virtual disks.

• B. Cluster Shared Volumes have not been created yet.

• C. The nodes are missing a storage replication partnership.

• D. The cluster does not have a quorum witness configured.

Best answer: A

Explanation: Storage Spaces Direct pools local disks across cluster nodes, so each node must expose its local drives directly to Windows through supported device types such as SAS, SATA, or NVMe. The key clue is that all discovered disks report the bus type as RAID and the validation report finds zero eligible disks. That usually means a hardware RAID controller is abstracting the physical disks as virtual disks instead of exposing them in HBA or pass-through mode. Quorum, CSV creation, and replication are separate cluster or storage features and do not explain why S2D inventory cannot find any eligible local disks. The next implementation step is to correct the controller/disk presentation and rerun cluster validation before enabling S2D.

• Quorum witness affects cluster voting resilience, not whether local disks are eligible for S2D pooling.
• CSV timing is backwards because S2D creates the pool and volumes before CSVs are used for clustered workloads.
• Replication partnership applies to technologies such as Storage Replica, not the initial S2D disk eligibility check.

Question 6

Topic: Implement Windows Server High Availability

You are performing a rolling operating system upgrade on a four-node Windows Server failover cluster that uses Storage Spaces Direct. After upgrading and restarting the first drained node, virtual machines remain online, but Windows Admin Center shows the following health state:

Node1: Up, upgraded OS
Storage jobs: Repair - Running
Virtual disks: Warning
CSV state: Online (redirected access observed)

To avoid avoidable downtime, what is the best next diagnostic action before upgrading another node?

Options:

• A. Evict the upgraded node and re-add it later

• B. Raise the cluster functional level immediately

• C. Drain the next node because quorum is maintained

• D. Verify storage repair completion and virtual disk health

Best answer: D

Explanation: In a rolling upgrade of a failover cluster that uses Storage Spaces Direct, node-by-node maintenance must account for both cluster role availability and storage resiliency. The evidence shows the first upgraded node is back online, but storage repair is still running and virtual disks are in a warning state. Draining another node during that repair can reduce resiliency further and increase the chance of VM impact or outage. The next diagnostic step is to confirm that storage jobs have completed and that virtual disks, CSVs, and the storage pool have returned to a healthy state before proceeding. Quorum alone is not enough for a safe S2D rolling upgrade.

• Functional level timing is wrong because the cluster functional level is raised only after all nodes are upgraded and rollback is no longer needed.
• Quorum-only reasoning fails because Storage Spaces Direct health and repair state also affect safe rolling maintenance.
• Evicting the upgraded node is unnecessary and could introduce additional disruption when the issue is ongoing storage repair.

Question 7

Topic: Implement Windows Server High Availability

A four-node Windows Server failover cluster uses Storage Spaces Direct for CSV storage. Node S2D-NODE3 suffered a motherboard failure and will be replaced with a newly installed server using the same name, Windows Server version, drivers, and network configuration.

Exhibit: Current status

Get-ClusterNode
Name        State
S2D-NODE1   Up
S2D-NODE2   Up
S2D-NODE3   Down
S2D-NODE4   Up

Add-ClusterNode -Name S2D-NODE3
Error: A node with this name already exists in the cluster.

What should you do next to recover the failed node?

Options:

• A. Clear the replacement server’s data disks before joining the cluster.

• B. Evict S2D-NODE3 from the cluster, then add the replacement node.

• C. Run Repair-ClusterStorageSpacesDirect before replacing the node.

• D. Run full storage validation against the online production cluster.

Best answer: B

Explanation: The evidence points to a stale cluster node membership problem, not an immediate storage repair problem. Get-ClusterNode shows S2D-NODE3 still exists in the cluster but is down, and Add-ClusterNode fails because that node name is already present. When a failed node is being replaced, remove the failed node object from the cluster from a healthy node, then add the prepared replacement server. After the node rejoins, allow Storage Spaces Direct to resynchronize and monitor storage jobs or health as needed.

Do not start by wiping disks or running storage repair when the replacement node cannot yet join the cluster.

• Storage repair first fails because the immediate blocker is cluster membership, not a virtual disk repair operation.
• Clearing disks is risky and unsupported as a first step because it can remove useful pool metadata or destroy data.
• Full storage validation is not the next step shown by the evidence and can be disruptive on an online production cluster.

Question 8

Topic: Implement Windows Server High Availability

A Windows Server failover cluster is being prepared for consistent host-network configuration by using Network ATC. The administrator runs a cluster-wide intent for management, compute, and storage traffic.

Exhibit: Network ATC status

Intent name: ConvergedIntent
Adapter names in intent: pNIC01, pNIC02

Node   Status    Detail
HV01   Success   Intent applied
HV02   Success   Intent applied
HV03   Failed    Physical adapter pNIC01 was not found

What is the most likely root cause?

Options:

• A. The cluster quorum witness is offline

• B. HV03 has different physical NIC names

• C. Live Migration is using the wrong network

• D. CSV ownership is assigned to HV03

Best answer: B

Explanation: Network ATC uses intents to deploy and maintain consistent host-network configuration across cluster nodes. When an intent references specific physical adapter names, each target node must expose matching adapters for Network ATC to apply the configuration. The status shows that HV01 and HV02 succeeded, but HV03 failed because pNIC01 was not found. The next administrative focus is to validate and correct the physical NIC naming and symmetry on HV03 before reapplying the intent.

This is a host-network consistency issue, not a cluster role, quorum, or CSV placement problem.

• Quorum witness affects cluster voting and availability decisions, not whether Network ATC can find a named physical adapter.
• Live Migration network selection occurs after host networking exists and does not explain the missing adapter error.
• CSV ownership affects storage coordination and does not prevent Network ATC from locating physical NICs.

Question 9

Topic: Implement Windows Server High Availability

An administrator is creating a two-node failover cluster from Windows Server 2022 servers that are not joined to an AD DS domain. Both nodes are in the same workgroup, have matching local administrator credentials, and use a common DNS suffix. Cluster creation fails with this message:

Unable to create the cluster name account in Active Directory.
Access is denied or the object cannot be created.

What is the most likely root cause?

Options:

• A. Matching local administrator credentials are unsupported.

• B. The cluster used an AD and DNS administrative access point.

• C. The shared DNS suffix prevents workgroup cluster creation.

• D. Failover clusters require all nodes to join the same domain.

Best answer: B

Explanation: Workgroup failover clusters are supported, but they do not use AD DS computer objects for the cluster name. If cluster creation tries to use the traditional Active Directory and DNS administrative access point, it attempts to create a cluster name object in AD DS and fails because the nodes are not domain joined. For a workgroup cluster, the administrative access point should be DNS-only, with appropriate local administrative credentials and name resolution in place.

The failure message points to an AD DS object-creation attempt, not to a general lack of cluster support for workgroup nodes.

• Same-domain requirement is incorrect because modern Windows Server supports workgroup and multi-domain failover clusters.
• DNS suffix issue is not indicated; a common DNS suffix is typically needed for workgroup cluster name resolution.
• Local credentials are not the problem because matching local administrator credentials are part of the expected workgroup-cluster setup.

Question 10

Topic: Implement Windows Server High Availability

You are creating a two-node Windows Server failover cluster for a line-of-business application. The vendor requires a single shared NTFS volume that can move between nodes, and the data must survive a single disk failure in the storage array. Both servers have iSCSI connectivity to the same SAN. Which storage configuration should you use?

Options:

• A. Present a resilient iSCSI LUN to both nodes and add it as a cluster disk

• B. Store the application data in Azure Files with Azure File Sync

• C. Create identical local disks on both nodes and replicate them with DFSR

• D. Attach the SAN LUN to one node and copy data during failover

Best answer: A

Explanation: A workload that requires a shared cluster disk needs shared block storage that every cluster node can access, with ownership controlled by the Failover Clustering service. In this scenario, an iSCSI SAN LUN that is protected by the storage array and presented to both nodes satisfies both requirements: shared access for failover and resiliency against a disk failure. After presentation, the storage should be included in cluster validation and then added as available storage or assigned to the clustered role. The key distinction is that failover clustering moves disk ownership between nodes; it does not make separate local disks behave like one shared disk.

• Local disk replication fails because DFSR does not provide a single shared block device for a clustered application disk.
• Azure file synchronization solves file replication and tiering, not a vendor-required shared NTFS cluster disk.
• Single-node attachment fails because the surviving node must already have access to the shared storage during failover.

Continue with full practice

Use the Microsoft AZ-802 Practice Test page for the full IT Mastery practice bank, mixed-topic practice, timed mock exams, explanations, and web/mobile app access.

Try Microsoft AZ-802 on Web View Microsoft AZ-802 Practice Test

Related focused pages

• Free Microsoft AZ-802 Full-Length Practice Exam
• Deploy and Manage AD DS in Hybrid Environments
• Manage Windows Servers and Hybrid Workloads
• Manage Virtual Machines and Containers
• Implement and Manage Hybrid Networking
• Manage Storage and File Services
• Secure Windows Server Hybrid Infrastructure
• Implement Disaster Recovery
• Migrate Servers and Workloads
• Monitor and Troubleshoot Windows Server

Free review resource

Read the Microsoft AZ-802 Cheat Sheet for compact concept review before returning to timed practice.