The replication path for database information includes the IBM i remote journal function, the MIMIX database reader process, and one or more database apply processes.
The IBM i remote journal function transfers journal entries from the source system to the target system. A brief overview of this function is described in Remote journaling.
The database reader process (DBRDR) process reads journal entries from the target journal receiver of a remote journal configuration and places those journal entries that match replication criteria for the data group into a log space. This occurs regardless of which configuration method is used to identify objects for replication.
All journal entries deposited into the source journal will be transmitted to the target system. The database reader process performs the filtering that is identified in the data group definition parameters and file and tracking entry options.
The database apply process applies the changes stored in the target log space to appropriate database or replicated object on the target system. The configuration of your MIMIX environment determines how many apply processes run in parallel for a data group and whether one of the apply processes is allowed to use a multithreaded job for greater efficiency in processing record-level transactions. Transactions that are not part of a commit cycle are immediately applied to the target system. For transactions that are part of a commit cycle, processing varies depending on how the data group is configured. Default configuration values in new data groups allow MIMIX to immediately apply transactions that are part of a commit cycle. Data groups can be configured so that MIMIX processes transactions that are part of a commit cycle but does not apply those transactions until the open commit cycle completes. Delayed commit mode processing is not possible when multithreaded database apply processing is configured.
Data group entries identify the journaled objects to be replicated, any mapping between source and target file names, and identify additional information used by database replication processes.
Depending on how the data group is configured, you may be able to specify file entry options for a data group or in individual data group entries to specify a particular apply session to use for processing on the target system or to control the type of lock that database apply uses when processing members on the backup system.
A status code in the data group file entry also stores the status of the file or member in the MIMIX process. If a replication problem is detected, MIMIX puts the member in hold error (*HLDERR) status so that no further transactions are applied. Files can also be put on hold (*HLD) manually.
Putting a file on hold causes MIMIX to retain all journal entries for the file in log spaces on the target system.
When configuring a data group, several parameters on the Create Data Group Definition command affect how the database apply process (DBAPY) performs its operations on the target node. These topics provide additional information about the most significant configuration choices:
Message cleanup: Each replication process removes unnecessary messages from its job log on a daily basis.