Replication Extract supports the use of DB2 compression. If any of the primary tables reside in a DB2 compressed tablespace, the DB2 log can contain compressed data. To expand this data, the DB2 IFI API requires access to the DB2 compression dictionary.
It is very important for Replication Extract that the DB2 IFI API is able to access the appropriate DB2-built compression dictionary. DB2 stores this dictionary in the compressed tablespace. Whenever possible, the API return expanded data. If the dictionary cannot be retrieved from the compressed tablespace, Replication Extract attempts to retrieve the dictionary from an appropriate full image copy.
The DB2 LOAD, REORG, RECOVER TOCOPY, RECOVER TORBA utilities can change the dictionary stored in the compressed tablespace. When you use any of these utilities, make a full image copy before the tablespace is made available for updates. This ensures Replication Extract can always access the appropriate compression dictionary.
These utilities may also stop, start, lock, or unlock a tablespace. Occasionally, a log record is written before a matching image copy is registered with DB2. When this occurs, Replication Extract waits until the tablespace or image copy is available before processing continues.
