Release Notes


|13.2 New Archive Logging Behavior

|Prior to FixPak 4, DB2 only checked for archive completion when a new log |file was needed. Now DB2 checks for archive completion whenever the |first active log changes. As a result, information is recorded to disk |earlier and more often.

|The benefit of this change is that if the system crashes, the information |stored on disk (related to which log files are successfully archived) is more |accurate and DB2 does not have to reissue the archive request for log files |that are already archived.

|There is no change to what DB2 does after detecting the successful archive |of a particular log file.

|DB2 now detects the completion of log archives earlier and will rename them |earlier. Inactive truncated log files are deleted. As a result, |the number of log files remaining in the active log path can be less than the |LOGPRIMARY database configuration value. In this case, DB2 will create |new log files when needed.

|Before this change, restarting the database reduced the number of logs to |equal the value of LOGPRIMARY. Now, when you restart a database, DB2 |first examines the database log directory. If the number of empty logs |is fewer than the number of primary logs, DB2 will allocate new logs to make |up the difference. If more empty logs are available than there are |primary logs in the database directory, DB2 will allow the database to be |restarted with all the available empty logs in the database directory.

|After database shutdown, any secondary log files in existence will remain |in the active log path at restart time. To clear out the active log |path, the DB2 ARCHIVE LOG command may be used.


[ Top of Page | Previous Page | Next Page | Table of Contents | Index ]