Scenarios and Best Practices › Scalability Best Practices › Scalability Recommendations › CA Server Automation vCenter Management Recommendations › Performance Considerations during Initial Discovery

Performance Considerations during Initial Discovery

Performing the initial discovery and database load of the vCenter environment you want to manage can be a time consuming process. During this process, the following take place:

1. The vCenter AIM parses the entire vCenter environment and publishes the results for the managers.
2. The CA Server Automation manager retrieves the published data from the vCenter AIM and creates an internal cache for processing.
3. The internal cache is synchronized with the current CA Server Automation manager database contents, with discoveries performed for Computer Systems not currently within the database.

During initial management of the vCenter server, the database has no Computer Systems in the database, so all of these objects are discovered and created. Consider the estimated time to complete initial discovery. Based upon baseline testing, the average throughput is: from eight to nine Computer Systems per minute.

Examples:

The following examples provide the sizes of environments and corresponding estimated times to complete initial discovery:

• 2,500 Computer Systems - approximately five hours
• 8,000 Computer Systems - approximately 15 hours

During this initial population, CPU usage may be high for extended periods.

Note: The initial discovery process takes the vast majority of this time. However, the initial discovery is done once for the lifetime of the product. The vCenter AIM and CA Server Automation internal caching processes, which are done whenever these processes are restarted, are considerably quicker. For example, 2500 Computer Systems are typically published through vCenter AIM and fully cached by the CA Server Automation manager in approximately 5 minutes.