When using the Ingres High Availability Option, the Ingres DBMS server is installed and configured as a typical installation, provided that the database-related files are installed on the cluster file system.
At the time of installation, the installer checks if the database, checkpoint, journal, dump, work, and transaction log files are on cluster drives. For a clustered DBMS installation, the database, checkpoint, journal, dump, work, and transaction log files must be on a cluster drive. The II_SYSTEM (Ingres home directory) must be installed on a non-cluster local drive. This configuration should be done while at the Ingres DBMS Configuration screen during installation.
Note: An error message is displayed if the installer finds that the database files are not on a cluster drive.
Using a logical hostname simplifies the installation and configuration of the Ingres DBMS Server in the cluster environment. Conversely, each node can be independently configured.
To access a DBMS server in a non-cluster environment, network clients must connect to a physical server, identified by a network name and IP address. If that server fails, access to the DBMS is impossible.
Windows 2000 Clustering enables the creation of virtual servers, identified by a unique network name and IP address (different from the underlying physical server's network name and IP address). Unlike a physical server, a virtual server is not associated with a specific computer. If the node hosting the virtual server fails, clients can still access the DBMS using the same server name.
Installing the Ingres High Availability Option modifies the Ingres installation to use the virtual server name. This changes the config.dat file by replacing the server name with the virtual server name and sets II_HOSTNAME to the virtual server name. To automatically reconnect to the server after a failover, network clients must use this virtual server name when connecting to the DBMS Server.