vRealize Automation requires the appliance database for system operation. You can manage the appliance database through the vRealize Automation Appliance Virtual Appliance Management Interface.

Note:

This information applies only to deployments that use an embedded appliance database. It does not apply to deployments that use an external Postgres database.

You can configure the database as a single node or with multiple nodes to facilitate high availability through failover. The vRealize Automation installer includes a database node on each vRealize Automation appliance installation. So if you install three instances of a vRealize Automation appliance, you have three database nodes. Automatic failover is implemented on applicable deployments. The appliance database requires no maintenance unless a machine configuration changes or, if you use a clustered configuration, you promote a different node for the master.

Note:

The database clustered configuration is set up automatically when you join a virtual appliance to the cluster using the Join cluster operation. The database cluster is not directly dependent upon the virtual appliance cluster. For instance, a virtual machine joined to a cluster can operate normally even if the embedded appliance database is not started or has failed.

For high availability, vRealize Automation uses the PostgresSQL master-replica model to support data replication. This means that all of the database nodes work in a cluster with one leading node, known as the master, and several replicating nodes, known as replicas. The master node handles all database requests and the replica nodes stream and replay transactions from the master locally.

A clustered configuration contains one master node and one or more replica nodes. The master node is the vRealize Automation appliance node with the master database that supports system functionality. Replica nodes contain copies of the database that can be pulled into service if the master node fails.

Several high availability appliance database options exist. Selecting the replication mode is the most important database configuration option. The replication mode determines how your vRealize Automation deployment maintains data integrity and, for high availability configurations, how it fails over if the master or primary node fail. There are two available replication modes: synchronous and asynchronous.

Both replication modes support database failover, though each has advantages and disadvantages. To support high availability database failover, asynchronous mode requires two nodes, whereas synchronous mode requires three nodes. Synchronous mode also invokes automatic failover.

Replication Mode

Advantages

Disadvantages

Synchronous

  • Minimizes chance of data loss.

  • Invokes automatic faiiover.

  • Might affect system performance.

  • Requires three nodes.

Asynchronous

  • Requires only two nodes.

  • Affects system performance less than synchronous mode.

Not as robust as synchronous mode in preventing data loss.

vRealize Automation supports both modes, but operates in asynchronous mode by default and provides high availability only if there are at least two appliance database nodes. The Database tab on the Virtual Appliance Management Interface enables you to switch synchronization modes and to add database nodes as needed.

When operating in synchronous mode, vRealize Automation invokes automatic failover.

If you begin with one node in a non-high-availability configuration, you can add nodes later as required to enhance high availability. If you have the appropriate hardware and require maximum protection against data loss, consider configuring your deployment to operate in synchronous mode.

Appliance Database Failover

In a high availability configuration, the master constantly streams transactions to the replica servers. If the master fails, the active and working replica is ready to proceed with read-only requests. When the new master is promoted, either manually or automatically, all of the upcoming requests are moved to it.