This section contains the procedure to configure and create the initial HADB database in a Sun Cluster environment. Consider the following restrictions before creating your database.
You must specify an even number of Sun Cluster hosts by using the --hosts option during database creation.
You must must specify the Sun Cluster hosts by using the Sun Cluster private interconnect hostnames. An example of a private interconnect hostname is clusternode1-priv.
Sun Java System Application Server EE (HADB) mirror nodes must be located on different Sun Cluster nodes.
Do not use the --inetd option.
All history files, data and log devices, and the database configuration files must reside on local file systems.
If more than one Sun Cluster node is being stopped, the entire database will be shut down. The resource group should only have one node stopped at any time, or all nodes should be stopped with the scswitch -F -g command.
The HADB resource group must not be used with the Resource Group Offload feature.
The HADB resource group must not be used with HA Storage Plus.
Use the example in the following procedure to create, start, and verify the database.
Create the database. This command automatically starts the database.
# hadbm create \ -H clusternode1-priv,clusternode2-priv,clusternode3-priv,clusternode4-priv, \ clusternode5-priv,clusternode6-priv --devicesize=2048 \ -a 4 --set ManagementProtocol=rsh --dbpassword=secret12 \ -s 2 hadb
For details, see the Sun Java System Application Server documentation.
If you are using the recommended SSH setup, you do not need to specify the ManagementProtocol property.
Verify that the database is running.
# hadbm status hadb --nodes
Stop the database.
# hadbm stop hadb
Create session store and JDBC connection pool. For details, see the Sun Java System Application Server documentation.