How Data Initially Replicates

The database server uses storage-space and logical-log backups (both those backed up to tape and those on disk) to perform an initial replication of the data on the primary database server to the secondary database server.

To replicate data

Install user-defined types, user-defined routines, and DataBlade modules on both database servers, and then register them on the primary database server only.
To make the bulk of the data managed by the two database servers the same, create a level-0 backup of all the storage spaces on the primary database server, and restore all the storage spaces from that backup on the secondary database server in the data-replication pair.
The secondary database server that you restored from a storage-space backup in the previous step then reads all the logical-log records generated since that backup from the primary database server.

The database server reads the logical-log records first from any backed-up logical-log files that are no longer on disk and then from any logical-log files on disk.

For detailed instructions on performing the preceding steps, refer to Starting HDR for the First Time. The IBM Informix Backup and Restore Guide discusses how to start replication using ON–Bar.

You must perform the initial HDR with a storage-space backup. You cannot use data-migration utilities such as onload and onunload to replicate data because the physical page layout of tables on each database server must be identical in order for HDR to work.

When HDR is working, the primary database server is online and accepts updates and queries just as a standard database server does. The secondary database server is in logical-recovery mode and cannot accept SQL statements that result in writes to disk (except for sorting and temporary tables).