If using an embedded SQL database storage (SQLite or SQL Server CE) to run a distributed load test, each request/response payload (session data) is stored in a local database. A load agent can replicate session data to the controller for more rich reporting.

Two types of data can be replicated:

  • Timing data: The information is necessary for waterfall charts in the test report.
  • Content data:  The request/response data used to verify server responses and debug concurrency errors.

Note:

The more data replicated during the test, the more bandwidth is required between the controller and agents.

Timing data

To enable timing data replication, set the Save sessions from agents? property to Yes.

Content data

To enable content data replication, set the Agent data scope property to the following:

  • Timing only: only timing data will be replicated.
  • Timing and errors (default): timing data and request/response content of sessions marked as errors.
  • Timing and dynamic data: timing data and request/response content of sessions marked by errors or sessions containing non-static MIME type responses. This avoids replicating images, stylesheets, and other static content from agent to controller.
  • All: timing data and request/response content for all 

Note:

If the request/response data is not replicated to the controller from the agent, it can still be retrieved on the agent after the test is over by following the steps described here.

Distributed testing with SQL server

In previous versions of StreStimulus, if using a SQL Server storage to run a distributed load test, each request/response payload (session data) is stored in the SQL Server instance. This meant that the controller and all load agents must access the SQL Server instance. This was a limitation, especially for distributed testing with provisioned AWS or Azure load agents, since these load agents don't have access to the local SQL Server. 

Starting from v5.6, StresStimulus has an option to use SQL Server on the controller and use SQLite instances on the load agents.

Set the Where to use SQL server property to:

  • Controller only: To use SQL Server for data storage on the Controller only. Each remote load agent will use local SQLite instances.

    Data replication

    Using this setting, data from the remote load agents will be replicated by the Save sessions from agents? and Agent data scope properties described above. All replicated data will be stored in the SQL Server instance.

  • Controller and all load agents (default): To use the SQL Server as a shared data storage for all test machines.

    Data replication

    If using this setting, then data from the remote load agents will not be replicated.

    SQL Server Authentication

    If using this setting, then it is recommended to use SQL Server Authentication to log in to the SQL Server instance. 


  • No labels