About
A rollback journal consists of records of the actions of transactions, primarily before they are committed.
Its name comes from the fact that its primary function is to roll back (undo) changes from a transaction but it's the central point of all other important data functionalities such as commit atomicity, (See features for a full list)
Features
A Rollback journal of records is used to:
- Implement atomic commit to provide transaction atomicity
- fault tolerance If a failure prevents modified data from being permanently written to the original data files, then the changes can be obtained from the rollback journal (not from the transaction log ?), so work is never lost.
- recover the database. During database recovery, rollback records are used to undo any uncommitted changes applied from the transactions log to the original data files.
- Provide read consistency through multiple sessions. To manage the multiversion read consistency, the database must create a read-consistent set of data when a table is simultaneously queried and updated.
- Analyze data as of an earlier point in time. See Oracle Flashback Query
- Recover from logical corruptions
Implementation
There is two well known implementations:
Management
Reduce
You can reduce the rollback data generation by introducing multiple commits in a loop while loading data, but the overall elapsed time will increase.