Replication

Replication enables you to maintain automation data between source and target servers, even across an air gap. To replicate a cluster of objects from one server to another, you can export them from a source server and then import to a target server.

Replication overview

During replication, the data is exported to and imported from a single file. If the servers are not physically connected, you need to manually upload a flat file that contains source server data to the target server.

You import data to the target server according to the following replication import strategies:

Strategy Description
Object import strategy This strategy defines how to match objects and resolve conflicts between objects during import. For details, see Object import strategies.
Plugins import strategy This strategy specifies how to import plugins and processes with plugin steps. The strategy defines the conditions for upgrading plugins during replication import. For details, see Plugins import strategies.
Properties merge strategy This strategy specifies how to merge conflicts between property sheets and property sheet definitions. The strategy was previously called Merge Strategy, and its functionality is the same. For details, see Replication merge strategies.

When you replicate data, take into account:

  • You can export a set of applications. All the linked objects, such as components, environments, and pipelines, are exported automatically with the applications.
  • You can export a set of environments. If you omit an environment that is a part of a pipeline, the pipeline is also omitted from the export.
  • Exported files include component versions of the associated components.
  • You can use snapshots instead of component versions.
  • Export and import works with a single file containing all the data.
  • Process execution history is not exported.
  • Component mapping works only if the target server agents have the same names as those on the source server.
  • During import, the creation date for all objects is set to the date of the current system. The creation date from the source server is ignored.

Back to top

Next steps: