Design your cluster

Identify your needs to determine how to design your cluster.

Because there are a variety of ways to use clustering depending on what you are hoping to achieve, it is important to spend some time identifying your needs to determine how to design your cluster.

• Design your network for clusters
  Before you configure your networks for clustering, you need to plan carefully and do some initial pre-cluster configuration involving TCP/IP.
• Multiple-release clusters
  If creating a cluster that will include nodes at multiple cluster versions, then certain steps are required when you create the cluster.
• Identify servers to include in a cluster
  In order to identify the servers that you want to include in a cluster, you need to decide which servers are capable of providing adequate backup for the data and applications you need to run your business.
• Comparison between primary-backup and peer models
  Primary-backup CRGs and peer CRGs provide resiliency for resources within a cluster; however, it is important to understand their differences and uses.
• Identify applications to include in a cluster
  Not every application will give you the availability benefits of clustering.
• Plan for data resilience
  Data resilience is achieved when data is always available to an user or application. You can achieve data resilience through the use of logical replication or switchable independent disk pools.