Two-node cluster hangs when one node fails, no quorum.
Two nodes always need a witness (file share, disk or cloud) to decide who wins in a split-brain. No witness = the whole cluster goes down on a single failure to prevent data loss.
Try this first
- 1Open Failover Cluster Manager, click the cluster, More Actions, Configure Cluster Quorum Settings.
- 2Pick 'Select the quorum witness'. For Server 2019/2022/2025 a cloud witness (Azure Storage account) is robust and off-site.
- 3Alternative: file-share witness on another server (not a cluster node), or disk witness on shared storage.
- 4For cloud witness: create an Azure Storage account, copy the access key, paste into the cluster quorum config. Availability beats a file share on another server.
- 5Test: power off one node, confirm the other takes over (Get-ClusterNode shows 'Up' on the survivor). Reset after the real test.
When to bring us in
For stretched clusters (two datacenters) a cloud witness in a third region is almost required. A file-share witness on either DC leads to split-brain if that DC fails.
See also
- One DC or two DCs for an SMB office?Two is almost always the right answer; one DC is a single point of failure for logon, DNS and GPOs.
- Should I split FSMO roles across two DCs?For a small domain all on one DC is fine; with two DCs splitting is tidier but not required.
- How do I know my AD replication is healthy?Replication errors creep in silently; they only surface when logins or GPOs misbehave.
None of the above fits?
Describe your situation below. We pass your input plus the steps you already saw to our AI and return tailored next-step advice. If it's too risky to DIY, we'll say so.
Or skip the DIY entirely
Our Managed IT clients do not look these things up. One point of contact, a fixed monthly price, resolved within working hours.