This subsection describes the procedure at the source Datareplicator.

1. Terminate the source HiRDB normally.
2. Use the hdestate command to make sure that the read and write locations are the same in the extraction information queue file.
3. Use the hdestop command to terminate Datareplicator.
4. Change the duplexing definition file.
5. Prepare the physical file of the file to be duplexed.^#
6. Delete the physical file that will no longer be used.
7. Delete any file that has the same name as the logical file of the file to be duplexed.
8. Use the hdestart -i command to initialize Datareplicator.
9. Make a backup of the duplexing control file created under $HDEPATH at each node.
10. If a system switchover configuration is used, copy the duplexing control file under $HDEPATH at the target.
11. Start the source HiRDB.
12. Execute the hdeprep command.
13. Use the hdestart command to start Datareplicator.

#

Preparing the physical file means allocating a partition to use a character special file, and then creating its symbolic link.

This subsection describes the procedure at the target Datareplicator.

The basic difference from the procedure for the source Datareplicator is that there is no need to terminate HiRDB.

1. Use the hdsstate command to make sure that the read and write locations are the same in the import information queue file.
2. Use the hdsstop command to terminate Datareplicator.
3. Change the duplexing definition file.
4. Prepare the physical file of the file to be duplexed.^#
5. Delete the physical file that will no longer be used.
6. Delete any file that has the same name as the logical file of the file to be duplexed.
7. Use the hdsstart -i command to execute initial startup of Datareplicator.
8. Make a backup of the duplexing control file created under $HDSPATH.
9. If a system switchover configuration is used, copy the duplexing control file under $HDEPATH at the target.

#

Preparing the physical file means allocating a partition to use a character special file, and then creating its symbolic link.

When you perform initialization for the first time after creating the import information queue file and the import status file as character special files, use the hdsstart -i -f command to perform an initial start of the target Datareplicator. There is no need to specify the -f option thereafter.