Best practices for Disaster Recovery

This topic provides basic guidelines that you should consider when designing a disaster recovery strategy for TrueSight Capacity Optimization servers. For TrueSight Presentation Server and for Remedy Single Sign-on, refer to the respective product documentation. 

General Requirements:

• Select the appropriate commercial tool that meets your organizational requirements to implement the disaster recovery strategy.
• The host names that are stored in the configuration properties of TrueSight Capacity Optimization is used to communicate internally or with the database. So, ensure that the host names in the disaster recovery installation must be same as the production installation. 
  To understand the communication between the servers, see Communication-ports-and-protocols.

• The file systems of TrueSight Capacity Optimization servers and the content of the database instances in the active nodes and the recovery machine must be in sync.
  Depending on the size of the data to synchronize and the network and hardware performance, you can determine the frequency. However, it would be ideal that the content is synchronized at least once a day. 

Here are some considerations, based on the server functionality:

Application Server Requirements

The deployment folder (default folder is /opt/bmc/BCO) of the production server must be in sync with the deployment folder of the recovery server.

• On an application server machine, if the File System is on the Application Server as local folder (shared with other Application Server), you need to synchronize this folder.
• If the repository is on a remote folder (such as on a Network-Attached Storage), you need to create a remove folder in the same way and keep the production and recovery installations in sync.

ETL Engine Requirements

The requirements are the same as for the Application Server.  All the routes to the data sources (example, route to the vCenters) need to be opened.

Database Requirements

To create a database for the disaster recovery site:

• The database must be an exact copy of the production database
• As some configuration properties are stored in the database, the database must be in synchronization with the production database  
  When there is a node switch, a misalignment between the production and disaster recovery database can cause a data gap.