Oracle database archive log tiering
Perhaps the most important use for FabricPool is improving the efficiency of known cold data, such as database transaction logs.
Most relational databases operate in transaction log archival mode to deliver point-in-time recovery. Changes to the databases are committed by recording the changes in the transaction logs, and the transaction log is retained without being overwritten. The result can be a requirement to retain an enormous volume of archived transaction logs. Similar examples exist with many other application workflows that generate data that must be retained, but is highly unlikely to ever be accessed.
FabricPool solves these problems by delivering a single solution with integrated tiering. Files are stored and remain accessible in their usual location, but take up virtually no space on the primary array.
Policies
Use a tiering-minimum-cooling-days
policy of a few days results in retention of blocks in the recently created files (which are the files most likely to be required in the near term) on the performance tier. The data blocks from older files are then moved to the capacity tier.
The auto
enforces prompt tiering when the cooling threshold has been reached irrespective of whether the logs have been deleted or continue to exist in the primary file system. Storing all the potentially required logs in a single location in the active file system also simplifies management. There is no reason to search through snapshots to locate a file that needs to be restored.
Some applications, such as Microsoft SQL Server, truncate transaction log files during backup operations so that the logs are no longer in the active file system. Capacity might be saved by using the snapshot-only
tiering policy, but the auto
policy is not useful for log data because there should rarely cooled log data in the active file system.