In Oracle RAC environments that run on Linux and UNIX platforms, you can use the CLUSTER_INTERCONNECTS initialization parameter to specify an alternative interconnect to the one Oracle Clusterware is using for the private network.
Note:
The CLUSTER_INTERCONNECTS initialization parameter should not be set to highly available IP (HAIP) addresses provided by Redundant Interconnect Usage. HAIP addresses are recognized automatically.
If you set multiple values for CLUSTER_INTERCONNECTS, then Oracle Database uses all of the network interfaces that you specify for the interconnect, providing load balancing if all of the listed interconnects remain operational. You must use identical values, including the order in which the interconnects are listed, on all instances of your database when defining multiple interconnects with this parameter.
Note:
Oracle does not recommend setting the CLUSTER_INTERCONNECTS initialization parameter, which overrides the default interconnect settings at the operating system level.
Instead, the best practice is to use Redundant Interconnect Usage, available with Oracle Grid Infrastructure 11g release 2 (11.2) for Oracle RAC and Oracle Real Application Clusters One Node 11g release 2 (11.2) databases, and later. For databases that precede Oracle Database 11g release 2 (11.2), use operating system-based network bonding technologies to enable high availability (and load balancing) for network interface cards meant to be used as the cluster interconnect. If you want to use multiple database versions in one cluster, you can combine both techniques. Redundant Interconnect Usage will use the interfaces as presented on the operating system level, regardless of bonding. For more information regarding bonding technologies contact your operating system vendor.
See Also:
Oracle Clusterware Administration and Deployment Guide for more information about Redundant Interconnect Usage