The topology defines the relative location of a data store and
its messaging engine, as well as the relationship with other data stores.
You have several options for the relative location of a data store and
its messaging engine. The following principles apply:
- The default embedded Apache Derby data store always runs in the same server
process as the messaging engine.
- For a messaging engine that can fail over within a
WebSphere® Application Server cluster,
you must ensure that, for any server that the policy might choose to run, the
messaging engine can gain access to the data store. The embedded Apache Derby
data store is not supported for a messaging engine running in a cluster. Use
either Apache Derby in Network Attach mode, or an alternative relational database
management system.
To support high availability (HA), you can configure the data store as
follows:
- Configure the data store to use a shared, network-mounted, file system.
Ensure that you protect against the file system being a single point of failure.
- Configure the data store to use a remote database server. Ensure that
you protect against a failure of the database server.
- Configure the data store to use a shared disk in an HA cluster. The HA
cluster performs failover of the data store and, optionally, can coordinate
failover of the messaging engine.