Clusters are commonly used to increase availability by making node failures more transparent. However, maintaining this high level of availability increases the complexity of changes (even minor hardware changes). For example, changes involving the cluster topology can be difficult or restricted because of failure fencing concerns. Although the SC2.2 software enables a smooth transition between different service demands, efficient planning and processes are essential to minimize difficulties.
This section discusses methods for making hardware and topology changes to existing clusters.
One meaning of the term cluster topology refers only to the number of nodes ...