Failover Clustering - Down Node - Current vote

Andy Summers 0 Reputation points
2024-06-11T11:53:46.1833333+00:00

3 node 2022 cluster with a Disk Witness, Dynamic Quorum

2 nodes show Current Vote 1 and one node shows Current Vote 0

If I Pause & Drain and then shutdown one of the nodes that has a Current Vote, should the Current Vote not move to to the node that is Up?

All seems to be working, and validation passes, but I expected the Vote to move from the Downed node?

Windows Server
Windows Server
A family of Microsoft server operating systems that support enterprise-level management, data storage, applications, and communications.
12,521 questions
Windows Server Clustering
Windows Server Clustering
Windows Server: A family of Microsoft server operating systems that support enterprise-level management, data storage, applications, and communications.Clustering: The grouping of multiple servers in a way that allows them to appear to be a single unit to client computers on a network. Clustering is a means of increasing network capacity, providing live backup in case one of the servers fails, and improving data security.
974 questions
0 comments

2 answers

Sort by: Most helpful
Most helpful Newest Oldest
  1. Alex Bykovskyi 1,841 Reputation points
    2024-06-11T14:22:11.06+00:00

    Hey,

    By default Dynamic Quorum won't assign vote to the Disk Witness if cluster has odd number of nodes. If your cluster has odd number of votes, Witness doesn't vote.

    https://video2.skills-academy.com/en-us/azure-stack/hci/concepts/quorum

    Check if all of your nodes are healthy in the cluster. Vote doesn't move, it is zeroed down when the node is down and Witness gets a vote, because cluster has an even amount of node left.
    Nice example: https://www.mssqltips.com/sqlservertip/4434/dynamic-quorum-for-windows-server-2012-r2-cluster-to-support-sql-server-alwayson/

    In addition, if you want to make your Disk Witness redundant, you can use StarWind VSAN to create replicated disk, which can be used for Disk Witness. https://www.starwindsoftware.com/starwind-virtual-san

    Cheers,

    Alex Bykovskyi

    StarWind Software

    Note: Posts are provided “AS IS” without warranty of any kind, either expressed or implied, including but not limited to the implied warranties of merchantability and/or fitness for a particular purpose.

    0 comments
  2. Ian Xue (Shanghai Wicresoft Co., Ltd.) 33,301 Reputation points Microsoft Vendor
    2024-06-12T02:21:00.49+00:00

    Hi Andy,

    In a Windows Server Failover Cluster with a Disk Witness and Dynamic Quorum, the quorum vote distribution and behavior can be dynamic. Typically, each node in a WSFC has one vote, and the Disk Witness also has one vote. The Dynamic Quorum feature, introduced in Windows Server 2012, allows the cluster to adjust the quorum vote count automatically to keep the cluster running even when multiple nodes fail.

    When you pause and drain a node, it's expected that the node's vote is removed to maintain the quorum. However, the vote does not "move" to another node. Instead, the total number of votes in the cluster is reduced. If the cluster is configured correctly with Dynamic Quorum, it will automatically adjust the quorum requirement so that the cluster can continue to run with the remaining nodes.

    The behavior you're observing, where the vote does not move to the up node, is consistent with how Dynamic Quorum is designed to work. The cluster dynamically adjusts the number of votes required for a quorum based on the number of available nodes, rather than moving votes between nodes.

    Here're the relevant documents for your reference:

    Configure and manage the quorum in a failover cluster | Microsoft Learn

    Set-ClusterQuorum (FailoverClusters) | Microsoft Learn

    Best Regards,

    Ian Xue

    If the Answer is helpful, please click "Accept Answer" and upvote it.

    0 comments