7.3. Creating a Database Cluster
 | Note |
|---|---|
The Clusters you create can run different versions of the backend. |
To create a new Database Cluster:
Select File => New or click the New Toolbar button.
The system displays the New Cluster dialog.
Choose a service name (such as rhdb-test or rhdb-production) for the cluster. All Clusters will receive names such as rhdb-nickname, as the name rhdb is reserved for the service that is automatically created by the RPM installation.
Either type or browse for the path for the location of the new cluster's main data area. Note that, if necessary, the pgsql and data path components are automatically added so that the resulting path is of the form */pgsql*/data.
Select the Database Engine that will be used to initialize and access the cluster. If the desired engine is not shown on the list of available engines, you can use the associated Browse button and then Add Engine from the dialog that appears.
Select a port and indicate if TCP/IP connections are to be accepted. You can always change this setting later (see Section 7.11 Editing Database Cluster Run Time Options).
Choose whether the cluster should be activated at boot.
Click the Advanced tab to select advanced options such as the cluster's encoding and the options --debug and --noclean.
See the entry for initdb in the PostgreSQL Server Applications section of the PostgreSQL 7.3 Reference Manual.
Confirm the request for cluster creation by selecting OK.
If the engine selected is of a compatible catalog version, you are offered the option to skip the initialization and reuse all of the data on the specified data area. In this case you should leave the options Encoding, Locale, Collate, Ctype, Debug, and No Clean in the Advanced tab in their default settings as they are used only at initialization. For security reasons, the port number and enabling of TCP/IP settings will be set to what is specified in the dialog, regardless of what values are stored in the run-time options. For the remaining options, if the values set in the dialog are different from the default, the dialog values are used and replace the ones stored in the data area. To change options to the default value, use Edit Cluster afterwards.
If the operation succeeds:
The init script and associated configuration files for the new cluster are installed on the system.
The main data area (PGDATA) is created and initialized.
If TCP/IP was specified, the option tcpip_socket in postgresql.conf is set to true.
If Activate on boot was selected, the service for this cluster is configured to be started automatically the next time the system boots.
/sbin/service and chkconfig are fully functional on the new service.