Determining how many Domino® servers to include in a cluster

Adding HCL Domino® servers to the cluster increases the cluster's ability to balance the workload so that no server becomes overloaded and performance stays high. However, if servers use too many server tasks, CPU-intensive applications, and replicas, adding servers can decrease performance because of the additional amount of cluster traffic required to keep databases synchronized on all servers.

A larger cluster is better able to absorb the workload when a cluster server fails. If you have a cluster with only two servers, for example, if one of the servers fails, the other server must absorb 100% of the failed server's workload. You could run each server at 50% capacity and absorb the workload if one goes down.

If the cluster has six servers, however, each of the remaining five servers must absorb only 20% of the failed server's workload. You could run each server at 80% capacity and absorb the workload if a server goes down.
Note: Other factors also determine how the workload of a failed server is absorbed, including replication distribution across the cluster.

Consider the following when planning hardware allocation for your cluster:

  • The more replicas you create, the more disk space you need and the more processing power you need for cluster replication.
  • The Cluster Database Directory requires approximately 2MB of disk space plus an additional 1MB for each 2,000 databases in the cluster.
  • The more servers in the cluster, the more processing power each server uses to communicate with the other cluster servers.
  • The more server tasks and CPU-intensive applications you run on a server, the more processing power you need.
  • Each server needs adequate processing power for the databases it contains and for any databases that might fail over to the server.
  • Clustered servers require more memory than nonclustered servers. The actual amount you will require depends on the level of activity on the server.

    To see if you need additional memory or processing power on your computer, check the Platform statistics.

  • When you have a large cluster or a cluster with a heavy workload, you might need to use multiple Cluster Replicators to improve the performance of cluster replication. Check the Cluster Replicator statistics to see if there is a large queue of databases waiting to be processed. If so, add Cluster Replicators one at a time until the statistics improve adequately.

    Because Cluster Replicators use system resources, the overall performance of the system could decrease as you add Cluster Replicators. Therefore, do not add more Cluster Replicators than you need.