Help Q Replication and Event Publishing

Detailed structure of version columns for peer-to-peer replication

Tables that participate in peer-to-peer replication require two version columns, a timestamp column and a small integer column, which are maintained by triggers. These two columns allow the Q Capture and Q Apply programs to perform version-based conflict checking that is required for peer-to-peer replication.

The columns also allow the two programs to resolve conflicts so that the tables within a Q subscription group maintain convergence. The values in these columns reflect which version of a row is most current.

These extra columns and triggers are created when you use the ASNCLP command-line program or Replication Center to create Q subscriptions for peer-to-peer replication. When you create a Q subscription for peer-to-peer replication, you must subscribe to both of these columns.

Important: Make sure that the version columns are not part of a primary key, unique constraint, or unique index to avoid having them selected by the replication administration tools as keys for replication.

Table 1 provides a brief description of the extra columns in user tables that are required for peer-to-peer replication.

Table 1. Extra columns required for peer-to-peer replication
Column name Description
IBMQREPVERNODE Data type: SMALLINT; Nullable: No, with default

A number that identifies the database or subsystem that contains the table within a peer-to-peer group. Default: 0

IBMQREPVERTIME Data type: TIMESTAMP; Nullable: No, with default

A timestamp that records when a change occurs in the table. Default: 0001-01-01-00.00.00



Send your feedback | Information roadmap | The Q+SQL Replication Forum

Update icon Last updated: 2013-10-25