The cluster must contain at least half of its nodes to function. That is, when the cluster is formed and becomes stable, only half of the nodes must remain functional for the cluster to continue operating. Note, however, that the cluster can survive more than half the nodes failing as long as the cluster has stabilized between failures. For example, a four-node cluster that loses one node would continue to function with the remaining three nodes.
A tie-break situation can occur if exactly half the nodes in a cluster fail at the same time, or if the cluster is divided so that exactly half the nodes in the cluster cannot communicate with the other half. For example, in a cluster of four nodes, if any two nodes fail at the same time, or any two cannot communicate with the other two, a tie-break exists and must be resolved.
The cluster automatically chooses three managed disks to be quorum disks and assigns them quorum indices of 0, 1, and 2. One of these disks is used to settle a tie-break condition.
If a tie-break occurs, the first half of the cluster to access the quorum disk after the split has occurred locks the disk and continues to operate. The other side stops. This action prevents both sides from becoming inconsistent with each other.
You can change the assignment of quorum disks at any time by typing the svctask setquorum command.