High Availability deployment model
The High Availability (HA) deployment model provides failure tolerance for the Control Room servers, services, and databases.
- Single/one node: This deployment model is not considered high availability with only one Control Room node configured in the data center.
- Three nodes: This deployment model is considered full high availability with three Control Room nodes configured in the data center cluster. This is the minimum required deployment model because it supports one node failure at a time.
The following shows the Automation Anywhere and data center components.
In this example, the Control Room servers and Microsoft SQL Servers have HA redundancy. The asterisk (*) on the PostgreSQL server and Subversion Version Control (SVN) server, indicate that those servers do not require HA redundancy. However, see Graceful Degradation for bot support if either of these fail.
- Multiple users have access the Control Room cluster through their web browsers. The web browsers communicate to the Control Room cluster through the load balancer.
- Multiple Bot Runners communicate to the Control Room cluster through the load balancer.
- The server message block (SMB) file share and the Microsoft SQL Server store data from the Control Room cluster.
- Microsoft SQL Server uses redundancy through active-standby replication syncing to the clustered Microsoft SQL Server. The primary server actively replicates data to the secondary (or standby) server.
- Maintains availability when server failures occur within a single data center.
- Does not provide protection against data center outage.
- Use Cases
- Small to medium-size businesses that do not require multi-site disaster recovery.