After-image interdependency between source and target databases

Source and target databases are interdependent in two ways. First, the target database requires after-image blocks and transaction log records (notes) generated by the source database in order to remain synchronized with the source database. All source database update operations generate transaction log records that are written into after-image extents.

The second and subtler dependency is the fact that the state of the source database after-image extents depends on normal replication activity between the source and target databases. As long as normal replication occurs between the source and its target databases, filled source after-image extents can be archived and emptied. However, if normal replication is not occurring, source database after-image extents continue to fill but remain in a LOCKED state until they are replicated in their entirety. Source database activity occurring without replication actively being performed relies upon the available space in the source after-image extents being greater than that of non-replicated databases.

In addition to the increased capacity of source after-image extents, there is another factor involved when determining the size of after-image extents for a source database. When a database is enabled as a source database, the database generates additional transaction log records that are written into the recovery log (BI) and the after-image log. You must consider the overhead generated by these additional transaction log records as you develop and implement successful after-image sizing and management schemes for OpenEdge Replication.