WSFC Quorum Modes and Voting Configuration (SQL Server)
Applies To: SQL Server 2016
Both SQL ServerAlways On Availability Groups and Always On Failover Cluster Instances (FCI) take advantage of Windows Server Failover Clustering (WSFC) as a platform technology. WSFC uses a quorum-based approach to monitoring overall cluster health and maximize node-level fault tolerance. A fundamental understanding of WSFC quorum modes and node voting configuration is very important to designing, operating, and troubleshooting your Always On high availability and disaster recovery solution.
In this topic:
Each node in a WSFC cluster participates in periodic heartbeat communication to share the node's health status with the other nodes. Unresponsive nodes are considered to be in a failed state.
A quorum node set is a majority of the voting nodes and witnesses in the WSFC cluster. The overall health and status of a WSFC cluster is determined by a periodic quorum vote. The presence of a quorum means that the cluster is healthy and able to provide node-level fault tolerance.
The absence of a quorum indicates that the cluster is not healthy. Overall WSFC cluster health must be maintained in order to ensure that healthy secondary nodes are available for primary nodes to fail over to. If the quorum vote fails, the WSFC cluster will be set offline as a precautionary measure. This will also cause all SQL Server instances registered with the cluster to be stopped.
For more information, see: WSFC Disaster Recovery through Forced Quorum (SQL Server).
A quorum mode is configured at the WSFC cluster level that dictates the methodology used for quorum voting. The Failover Cluster Manager utility will recommend a quorum mode based on the number of nodes in the cluster.
The following quorum modes can be used to determine what constitutes a quorum of votes:
Node Majority. More than one-half of the voting nodes in the cluster must vote affirmatively for the cluster to be healthy.
Node and File Share Majority. Similar to Node Majority quorum mode, except that a remote file share is also configured as a voting witness, and connectivity from any node to that share is also counted as an affirmative vote. More than one-half of the possible votes must be affirmative for the cluster to be healthy.
As a best practice, the witness file share should not reside on any node in the cluster, and it should be visible to all nodes in the cluster.
Node and Disk Majority. Similar to Node Majority quorum mode, except that a shared disk cluster resource is also designated as a voting witness, and connectivity from any node to that shared disk is also counted as an affirmative vote. More than one-half of the possible votes must be affirmative for the cluster to be healthy.
Disk Only. A shared disk cluster resource is designated as a witness, and connectivity by any node to that shared disk is counted as an affirmative vote.
By default, each node in the WSFC cluster is included as a member of the cluster quorum; each node has a single vote in determining the overall cluster health, and each node will continuously attempt to establish a quorum. The quorum discussion to this point has carefully qualified the set of WSFC cluster nodes that vote on cluster health as voting nodes.
No individual node in a WSFC cluster can definitively determine that the cluster as a whole is healthy or unhealthy. At any given moment, from the perspective of each node, some of the other nodes may appear to be offline, or appear to be in the process of failover, or appear unresponsive due to a network communication failure. A key function of the quorum vote is to determine whether the apparent state of each of node in the WSFC cluster is indeed that actual state of those nodes.
For all of the quorum models except ‘Disk Only’, the effectiveness of a quorum vote depends on reliable communications between all of the voting nodes in the cluster. Network communications between nodes on the same physical subnet should be considered reliable; the quorum vote should be trusted.
However, if a node on another subnet is seen as non-responsive in a quorum vote, but it is actually online and otherwise healthy, that is most likely due to a network communications failure between subnets. Depending upon the cluster topology, quorum mode, and failover policy configuration, that network communications failure may effectively create more than one set (or subset) of voting nodes.
When more than one subset of voting nodes is able to establish a quorum on its own, that is known as a split-brain scenario. In such a scenario, the nodes in the separate quorums may behave differently, and in conflict with one another.
In order to simplify your quorum configuration and increase up-time, you may want to adjust each node’s NodeWeight setting so that the node’s vote is not counted towards the quorum.
KB2494036: A hotfix is available to let you configure a cluster node that does not have quorum votes in Windows Server 2008 and in Windows Server 2008 R2
When enabling or disabling a given WSFC node’s vote, follow these guidelines:
No vote by default. Assume that each node should not vote without explicit justification.
Include all primary replicas. Each WSFC node that hosts an availability group primary replica or is the preferred owner of an FCI should have a vote.
Include possible automatic failover owners. Each node that could host a primary replica, as the result of an automatic availability group failover or FCI failover, should have a vote. If there is only one availability group in the WSFC cluster and availability replicas are hosted only by standalone instances, this rule includes only the secondary replica that is the automatic failover target.
Exclude secondary site nodes. In general, do not give votes to WSFC nodes that reside at a secondary disaster recovery site. You do not want nodes in the secondary site to contribute to a decision to take the cluster offline when there is nothing wrong with the primary site.
Odd number of votes. If necessary, add a witness file share, a witness node, or a witness disk to the cluster and adjust the quorum mode to prevent possible ties in the quorum vote.
Re-assess vote assignments post-failover. You do not want to fail over into a cluster configuration that does not support a healthy quorum.