High-availability replication systems
You can combine HCL® OneDB® Enterprise Replication and high-availability clusters to create a high-availability replication system.
- A primary database server, which receives updates, and can participate in Enterprise Replication.
- Secondary servers, which mirror the primary server and are perpetually applying logical-log records from the primary server, and cannot participate in Enterprise Replication.
A minimal high-availability cluster consists of a primary server and a HDR secondary server that are tightly coupled. Transactions on the primary server are not committed until the log records containing the transactions are sent to the HDR secondary server.
High-availability clusters can also contain shared-disk (SD) secondary servers and remote standalone (RS) secondary servers. A SD secondary server does not maintain a copy of the physical database on its own disk space; it shares disks with the primary server. An RS secondary servers maintains a copy of the physical database on its own disk space.
If the primary server in a high-availability cluster becomes unavailable, one of the secondary servers takes over the role of the primary server. In a high-availability replication system, if the primary server fails, a secondary database is promoted to primary server, and Enterprise Replication can continue with the new primary server.
You can configure Connection Managers to direct client requests to replication servers, and to control which secondary server takes over if the primary server becomes unavailable.
A high-availability replication system is effective when you use a hierarchical or a forest of trees topology.