Implementing and Managing Windows Server High Availability
Core concepts
High availability centers on failover clustering. Know quorum models (node majority, disk witness, file share witness, cloud witness), dynamic quorum, and how to choose a witness. Understand cluster networks, live migration, CSV (Cluster Shared Volumes), and Hyper-V failover clustering. For file services, know Scale-Out File Server, DFS Namespaces and Replication, and Storage Replica for synchronous or asynchronous replication. Network Load Balancing (NLB) remains relevant for stateless services. Understand cluster-aware updating and rolling upgrades.
Traps to watch
Quorum is the most tested area. With an even number of nodes, a witness is required; with odd nodes, node majority can suffice. Cloud witness requires internet access and an Azure storage account. Dynamic quorum adjusts votes automatically but does not replace a witness in all cases. CSV is not the same as a shared disk; it allows simultaneous access. Live migration requires shared storage or storage migration, and constrained delegation or Kerberos configuration. NLB does not provide application-level health checks.
Worked example
A four-node cluster spans two sites (two nodes each). Rule of thumb: use a file share witness in a third site or a cloud witness, because an even node count plus a site failure leaves two nodes and no majority without a witness. Verify the witness is reachable by all nodes and not hosted on a cluster node.
Exam approach
Count votes, consider failure scenarios, then pick the witness. For file services, match scale and consistency needs to SOFS, DFS, or Storage Replica.