Database rotation

This page describes the approaches to managing table size growth and archiving old data.

With data constantly being added to the system, and activity being logged into system tables in the database, these tables grow in size and require management. As data sets increase in size, the amount of I/O traffic associated with actions such as cleaning, deleting, and archiving can negatively effect the performance of an instance. Additionally, working with all rows in a data set, rather than a smaller working set, can create unnecessary risk.

The Database Rotation plugin preserves instance performance and averts risk associated with querying growing data sets utilizing two techniques. Both techniques are based on the concept of managing large quantities of data by separating whole sets into individual tables based on user-specified time parameters. After this task is performed, each technique handles data in a different manner:
  • Table Rotation works by rotating among a small set of tables, and deleting and reusing the old tables for new data.
  • Table Extension works by periodically starting a new table and allowing old tables to be easily archived and removed from the system.

For an additional archiving option see Archiving Data.