Archivelog | Noarchivelog   «Prev  Next»
Lesson 3Noarchivelog mode
ObjectiveDiscuss noarchivelog mode.

Discuss noarchivelog mode in Oracle

Remember that the database by default starts in noarchivelog mode. If we operate a database in noarchivelog mode we must perform either a logical backup or a physical offline mode backup. (Again, logical backups will be covered in the third part of this series, in the discussions on the Export/Import utilities.)

Noarchivelog mode

Several things happen when you operate a database in noarchivelog mode. As users access the database, information is written from the redo log buffer area to redo log files in a cyclical fashion. Information is first written to one redo log file, a log switch occurs and then information starts being written to the next redo log file. If there are only two redo log files, the next log switch will start to write data over the first redo log file. Once we start writing over the first redo log file, any information contained in this log file is lost.
The following Slide Show illustrates redo log file processing with noarchivelog mode enabled.
1) archive-redo-log-buffe 1
2) archive-redo-log-buffe 2
3) archive-redo-log-buffe 3

Backup and recovery options

With noarchivelog mode enabled you only have one option to backup and recover your database. You must do a cold backup. This means that the database is shutdown and you do an operating system backup of all the data files including the redo and control files. Your restore will be an operating system restore of these same files. You will lose any data entered into the database since this backup.
There is a very limited alternative. Let us say you have a read only database, one with limited transactions or containing many large disk drives. You can set up groups of multiplexed redo logs files, such that you never overwrite any redo log file. This could apply to maybe 1% of all the Oracle databases.
The next lesson is about media recovery with noarchivelog mode.