What are the redo log files used for in the oracle database?


Article ID: 160289


Updated On:


Data Loss Prevention Enforce


In the same directory as the Oracle database files, or *.dbfs, there are three "redo*.log" files.


Although they have the .log extension, the redo.log files are most assuredly a part of the Oracle Database and are important to a cold database backup.  These, along with all other files in the directory, are needed to restore the database should something catastrophic happen to the database.   


Redo logs contain transaction information in case the database is interrupted prior to writing data changes from memory to the database.  There are three files (in our setup) that work in a “round-robin” fashion; when one fills or reaches a timing threshold,  Oracle switches to the next.  At any given time only one redo log is “active”.  Typically, when Oracle writes the associated data changes to disk, the statements in redo are no longer necessary.  If a database crash occurs, on recovery, the redo logs are applied, “rolling forward” any unsaved changes they may contain.  Rather than having actual data, the redo logs contain Data Manipulation Language statements, or DML.  In SQLPlus, this would be any statement that changes data in some way: INSERT, UPDATES, DELETES.  Redo logs are used to generate archive logs which are used in more advanced backup and recovery methods.  Symantec DLP does not provide support for any backup-recovery method apart from cold database backup and recovery.


Even though we don’t take advantage of archive logging, the redo logs are still important to any backup-recovery method, including cold backup.