MultiPools

A multipool is a metapoolthat is, it is a pool of connection pools. (See Figure 7-3.) Multipools are defined under the Services JDBC tab in the administration console. They are essentially a named, ordered list of connection pools. The scope of a multipool is limited to the local machine.

Figure 7-3. Multipools

Note that while all the connections in a single connection pool must be to the same data source, connections within a multipool may be to more than one. One important implication that follows directly is high availability. When the connection to the primary data source is perceived to be broken, the multipool can automatically fail over to a redundant data source.

Of course, this is not a coincidence . Multipools are a new development in WebLogic 6.1, designed to implement either High Availability or Load Balancing, as described below:

High Availability	If a multipool is configured for High Availability, WebLogic will service client resource requests by passing the request on to each pool in the order it is listed. If the pool handles the request, all is well. If not, the server will pass the request on to the next pool in the list, then the next , and so on until the request is fulfilled or the end of the list is reached.
Load Balancing	If a multipool is configured for Load Balancing, WebLogic will distribute client requests to each pool in the multipool in round- robin order.

To create a multipool, first configure the connection pools which will be grouped into it. (See the discussion of connection pools, above.) Next, open the JDBC tab under Services in the Administration pane. Type in the name of the connection pool, which may be anything you like. Select either High Availability or Load Balancing as the algorithm type, and click Create.

Next, click on the Pools tab. From the list of available pools, select the ones you wish to add to the connection pool. Note that the order in which you select them does matter, particularly for high-availability pools.