Use the multiple masters configuration in this section to plan the installation and configuration of the Sun Cluster HA for Sun Java System HADB. The Sun Cluster HA for Sun Java System HADB data service might support additional configurations. However, you must contact your Enterprise Services representative for information about additional configurations.
HADB configurations are defined by data nodes, or a collection of processes. Each node is a dedicated area of main memory with one or more secondary storage devices. These storage devices are not shared storage. Each HADB data node must have exclusive access to an area of main memory and several areas of disk space. HADB data nodes are active or spare.
The recommended minimum requirement for a cluster running the Sun Java System HADB and the Sun Java System Application Server is four active data nodes plus two spare nodes. See Creating a Sun Java System HADB Database for an example of how to create a Sun Java System HADB with four active and two spare nodes by using the hadbm comand-line utility. For high availability, you configure the HADB data redundancy unit (DRU) to use the Sun Cluster interconnect. See Sun Java System Application Server 7 Administrator's Guide in the Sun Java System Application Server 7 Update 1 Collection (Solaris Edition) for detailed information about DRUs. The HADB JDBC driver in the client application handles highly-available access to the database.
The following figure illustrates the recommended minimum configuration for the Sun Java System HADB and the Sun Java System Application Server.