When you choose the shipped default configuration options, MIMIX uses data group object entries to identify whether to process transactions for library-based objects. Collectively, the object entries identify which library-based objects can be replicated by a particular data group.
When you configure a data group to use journal-centric configuration for object types of *ALL or *FILE, *LIB, *DTAARA, or *DTAQ, data group object entries are not needed to identify what to replicate. For the object types specified in the data group for journal-centric configuration, MIMIX keeps track of which objects are journaled to the user journal associated with the data group. However, for the specified journal-centric object types, you may still need data group object entries to identify exceptions to exclude from replication. Data group object entries are still required for object types that cannot be journaled to a user journal, for identifying objects with differences to ignore during auditing, and for objects that you do not want to replicate from the user journal for performance reasons.
Each data group object entry identifies one or more library-based objects. An object entry can specify either a specific or a generic name for the library and object. In addition, each object entry also identifies the object types and extended object attributes (for *FILE and *DEVD objects) to be selected, defines a configured object auditing level for the identified objects, and indicates whether the identified objects are to be included in or excluded from replication.
For most supported object types which can be identified by data group object entries, only the system journal replication path is available. For a list of object types, see Supported object types for system journal replication. This list includes information about what can be specified for the extended attributes of *FILE objects.
A limited number of object types which use the system journal replication path have unique configuration requirements. These are described in Identifying spooled files for replication, Replicating user profiles and associated message queues and Replicating jobs associated with job queue objects.
For detailed procedures, see Configuring data group entries.
Replication options for object types journaled to a user journal - For objects of type *FILE, *DTAARA, and *DTAQ, MIMIX supports multiple replication methods. For these object types, additional configuration data is evaluated when determining what replication path to use for the identified objects.
For *FILE objects, the extended attribute and other configuration data are considered when MIMIX determines what replication path to use for identified objects.
-
For logical and physical files, MIMIX supports several methods of replication. Each method varies in its efficiency, in its supported extended attributes, and in additional configuration requirements. See Identifying logical and physical files for replication for additional details.
-
For other extended attribute types, MIMIX supports only system journal replication. Only data group object entries are required to identify these files for replication.
For *DTAARA and *DTAQ objects, MIMIX supports replication using either system journal or user journal replication processes. Additional information, including configuration requirements, is described in Identifying data areas and data queues for replication.
Replication options for object types journaled to a system journal - These additional options are available for *FILE, *DTAARA, and *DTAQ objects that are configured for system journal replication:
-
For *FILE, *DTAARA, and *DTAQ objects configured for replication through the system journal, the Omit content (OMTDTA) parameter provides the ability to omit a subset of data-changing operations from replication. For more information, see Omitting object content from system journal replication.
-
For *FILE objects configured for replication through the system journal, MIMIX caches extended file attribute information for a fixed quantity of *FILE objects. For more information, see Caching extended attributes of *FILE objects