When you enable archiving of your truncated log file, the following process occurs:
1. An interim archive file is created in your archive directory, or in the database directory, if an archive directory is not specified. The interim archive file is named,dbname.lg.date-time.n.tmplg, where the date-time is the time the temporary archive is created.
2. Over time, the database broker writes log file entries to the interim archive in 8K chunks.
3. When the database log file is truncated, the broker writes any remaining log messages to the interim archive.
4. The interim archive is re-named to the archive file name, using the date-time when the truncation and archive criteria are met. The file is named, dbname.lg.date-time.n.
5. A new temporary archive is created the next time the broker has an 8K chunk to write to it.
6. Repeat Steps 2-5.
The temporary archive is deleted at database shutdown, or if the database log file is truncated by PROLOG.
Archive file naming convention
When you archive your log file, it is given a name using the convention, dbname.lg.date-time.n.
In this file name:
dbname — The name of the database.
date-time — The date and time when the archive file is created, using the format YYYY-MM-DD.HH-MM.
n — An identifier, only incremented when the combination of dbname and date-time does not create a unique file name.
The interim archive file uses the same convention, but adds a .tmplg extension to the file name.
dbname — The name of the database.