Couchbase Server is largely self-managing. There are few configuration variables, and most of these do not need to be modified by most users. You also do not need to perform regular maintenance to your cluster or data. Couchbase automatically manages this as part of the standard operation.
The result is that management of your cluster is only required when human intervention of some kind is required, for example:
As you store information in your cluster and the load on your cluster increases or requires more data storage, you will need to increase the size of your cluster. The converse may also be true. For example, if you have deployed your application within a cloud service such as Amazon EC2, you will want to shrink the size of your cluster when the RAM or disk I/O requirements recede to lower your running costs.
The act of expanding and shrinking your cluster is called rebalancing and can be performed on a live, running cluster.
If a note within your cluster fails for some reason (hardware or network failure are the most common), then Couchbase Server can switch to one of the replicas of the data. Failover can be manual or automatic, and there are different considerations for the two methods, which are both described later in this chapter.
Although Couchbase Server automatically distributes your data, and creates 1-3 replicas of your bucket data, it is good practice ...