Advanced Replication Guide : Using advanced replication with applications : Designing and implementing Intelligent Transactions : Creating a synchronization error log table for an application
  
Creating a synchronization error log table for an application
It is highly recommended that a synchronization error log table be created as part of the database schema of a decentralized system. The log table allows application developers to store information about the application-level errors that may occur during synchronization.
If the error requires manual resolution, the error log serves as a source of information required to correct the error. For instance, if an update conflict occurs, a log can contain information about which row was conflicting and what was done to the conflicted row by the transaction. Below is an example of an error log table:
CREATE TABLE ERRLOG
  (ID CHAR(20) NOT NULL,
  LOG_TITLE CHAR(80) NOT NULL,
  LOG_DESCRIPTION VARCHAR NOT NULL,
  UPDATETIME DATETIME NOT NULL,
  PRIMARY KEY (ID));
The ID is a generated, globally unique key of the log entry.
The LOG_TITLE and LOG_DESCRIPTION columns contain the information about the error, such as an update conflict.
The UPDATETIME column contains the timestamp of the last update operation of the error log row.
See also
Designing and implementing Intelligent Transactions