You can define keyed replication for a data group when you are initially configuring MIMIX or you can change the configuration later. To use keyed replication for all database replication defined for a data group, the following requirements must be met:
-
Before you change a data group definition to support keyed replication, do the following:
-
Verify that the data group specifies a numeric value for the Number of DB apply sessions (NBRDBAPY) parameter.
Note: If the NBRDBAPY parameter specifies a threaded value (*THDLOW, *THDMED, or *THDHIGH), the resource group with which the data group is associated must be converted to use single-threading using Checklist: Converting to single-threading. After the conversion is complete, you can continue with these instructions. -
Verify that the files defined to the data group are journaled correctly. Do not continue until this is verified.
-
If the files are not currently journaled correctly, you need to end journaling for the file entries defined to the data group. Use topic “Ending journaling for physical files.
-
-
In the data group definition used for replication you must specify the following:
-
Data group type of *ALL or *DB.
-
DB journal entry processing must have Before images as *SEND for source send configurations. When using remote journaling, all journal entries are sent.
-
Verify that you have the value you need specified for the Journal image element of the File and tracking ent. options. *BOTH is recommended.
-
File and tracking ent. options must specify *KEYED for the Replication type element.
-
-
The files identified by the data group file entries for the data group must be eligible for keyed replication. See topic “Verifying Key Attributes” in the Assure MIMIX Operations book.
-
If you have modified file entry options on individual data group file entries, you need to ensure that the values used are compatible with keyed replication.
-
Start journaling for the file entries using Starting journaling for physical files.