A Connection Broker cluster is a group of Connection Brokers that share the same PostgreSQL or Microsoft SQL Server® database. Clustering Connection Brokers allows you to support more connections per second, enhance availability, and improve disaster recovery capabilities.
The simplest way to create a cluster is as follows:
For more on the benefits of using clusters and how to effectively manage clusters, see "Using Clusters to Maximize Availability" in the Leostream Scalability Guide.
1. Install a standalone Connection Broker. By default, the Connection Broker uses an internal database.
2. Once that Connection Broker is configured, you can begin building your cluster. First, you need the address and credentials for a PostgreSQL or Microsoft SQL Server 2012 or 2014 database server. Connect all the Connection Brokers in you cluster to the same database.
3. Next, switch the first Connection Broker to an external database. Go to the Connection Broker > System > Maintenance page, select one of the options to switch to an external database, and click Next.
4. In the Database form, switch this Connection Broker to the new external database. When switching the database, note the database name and Site ID for this Connection Broker. See "Switching to an External Database" in the Leostream Scalability Guide for complete instructions.
5. To add Connection Brokers to the cluster, install individual Connection Broker virtual appliances on different virtualization hosts. These Connection Brokers can be located in any data center, as long as the Connection Broker can communicate with your external database.
6. For each additional Connection Broker, go to the Connection Broker > System > Maintenance page.
7. Switch these Connection Brokers to the same database. Each Connection Broker must be given a unique Site ID. Because the Connection Broker database already exists, information stored in the additional Connection Broker internal databases is not copied over. The additional Connection Brokers are simply attached to the existing SQL Server database.