• log-based replication by network (case of a laptop that connects to the company network on a regular basis).
• log-based replication via portable media (case of remote sites that do not connect via a network).

What is it?

The possible actions

• a unidirectional replication:
  • from the master computer to the subscriber computers.
  • from the subscriber computers to the master computer.
• a bidirectional replication, which means from the master computer to the subscriber computer (and conversely).

How to replicate by network and by programming

1. Creating the master replica on the master computer (HCreateMasterReplica).
2. Creating the subscriber replica on the subscriber computer (HCreateSubscriberReplica). The subscriber computer must have read/write access to the disk where the master application is installed (directory of data files).
3. Using the master and subscriber applications.
4. Synchronizing (HSynchronizeReplica):
   • from the subscriber computer to the master computer (rplToMaster constant constant): the subscriber computer must have read/write rights on the disk where the master application is installed.
   • from the master computer to the subscriber computer (rplToSubscriber constant): the master computer must have read/write rights on the disk where the subscriber application is installed.
   • bidirectional (rplBidirectional constant): the master computer must have read/write rights on the disk where the subscriber application is installed and the subscriber computer must have read/write rights on the disk where the master application is installed.

Constraints and tips

• When creating the subscriber replica, the computer where the subscriber replica is installed must have read/write access onto the directory where the master replica is installed.
• To perform the log-based replication, the path of the data files to replicate can be identified by:
  • a drive letter. In this case, the two computers involved in the replication must use the same drive letter. For example, if the replication is started from the subscriber computer:
    • on the master computer, the directory of the data files is found on X,
    • on the subscriber computer, the letter X must be associated with the drive containing the data files of the master computer.
  • UNC paths
  • relative paths.
• To manage log-based replication in an application, the table of log operations (JournalOpération.fic file) must be identical for all files in replication in the same analysis.
• The files required by the replication (*.RPM and *.RPL files) should be saved in a RPL sub-directory of the application data files.
• If the log process is used for the replication only, the log files can be located in the directory of application (or in a sub-directory).
• If the log process is used to keep a backup of the files, these files should be located on a disk other than the one used for the data files (for better security).

What is it?

The possible actions

• from the master computer to the subscriber computers
• from the subscriber computers to the master computer

Constraints and tips

• To manage log-based replication in an application, the table of log operations (JournalOpération.fic file) must be identical for all the files in replication in the same analysis.
• The files required by the log-based replication (*.RPM and *.RPL files) should be saved in a RPL sub-directory of application data files.
• If the log process is used for the replication only, the log files can be located in the directory of application (or in a sub-directory).
• If the log process is used to keep a backup of the files, these files should be located on a disk other than the one used for the data files (for better security).

By programming

1. Creating:
   • the master replica on the master computer (HCreateMasterReplica).
   • the subscriber replica on the subscriber computer (HCreateSubscriberReplica). Each subscriber computer must be associated with a specific identifier. This identifier will be used during the synchronization.
     These two replicas can be created in the same directory (not recommended).
2. Installing the file of subscriber replica on the subscriber computer. This file can be installed:
   • in the replication directory specified in the analysis (\RPL subdirectory of data files by default).
   • in another directory. In this case, the subscriber application must use HChangeRplDir to specify the location where the file of subscriber replica will be sought. Indeed, when the subscriber data file is opened, the RPL file is read in order to define the range of automatic identifiers used by the data file.
3. Using the master and subscriber applications.
4. On the subscriber computer, creating the movable replica (HCreateMoveableReplica). A file with a WDZ extension is created by default. This file must be transmitted to the master computer.
5. Synchronization on the master computer (HSynchronizeReplica). To perform this synchronization, you must specify:
   • the name and path of the movable replica on the master computer
   • the identifier of the subscriber replica (specified when creating the subscriber replica).

Functions for managing log-based replication