TUNING
To minimize delays that occur when users, servers, or API programs open and close databases on a server, each server maintains a database cache. When a database closes and there are no users or processes using the database, Domino® puts the database in the cache so it can close it quickly. The database remains in the cache until it's opened again or for about 15 to 20 minutes, whichever comes first. Databases in the cache can be opened quickly.
The database cache is available to the first process that starts on a machine and to any processes spawned from it. If you run the Domino Administrator and the Domino server on the same machine -- a configuration that is not recommended -- start the server before you start the Domino Administrator. If you start the Domino Administrator first, it owns the cache and prevents the Domino server from using it effectively.
By default, the number of databases that the cache can store simultaneously is the greater of these values:
The actual number of databases allowed in the cache is 1.5 times the maximum allowed. This buffer increases the chance that when a user opens a database from the cache, Domino can return the database to the cache when the user closes it.
Databases are dropped from the cache by an "ager" thread that performs necessary writes, de-allocates memory, and completes other tasks to close databases. This process happens over a period of 15 to 20 minutes. Ideally, databases are dropped from the cache in time to allow new databases to be added without exceeding the maximum databases allowed in the cache. However, if the maximum is exceeded, one of the following occurs:
Related reference Monitoring the database cache