Logging and recovery
In IT, logging and recovery refers to a collection of measures and techniques to prevent data loss in database systems (DBS) after an error (e.g. after a system crash or a failure of hard drives) and to maintain the accuracy ( consistency ) of the database to guarantee. Possible types of error are: transaction errors (e.g. abortion of a current transaction due to blocking by another transaction), system errors (e.g. computer failure), device or external storage errors (e.g. hard drive failure due to a head crash) or the so-called Disaster (e.g. the complete loss of the data center after an earthquake).
The transaction log writes all changes to the database during operation and the recovery provides the last current (prevention of data loss) and error-free ( consistent ) status after an error (e.g. after the reboot if a system crash was responsible for the error ) the database.
Logging and recovery are therefore responsible for maintaining consistency, durability and atomicity in a database, three of the ACID properties.
The system components required for logging and recovery are temporary log files (for backing up the log buffer of a DBS or DBVS ), the archive log (for backing up the DB buffer of a DBS or DBVS), and the archive copy of the database (for Backup of the database of a DBS).
Transaction log
A transaction log ( English transaction log, transaction journal, database log, binary log ) is a history of actions carried out in the area of databases . The aim is to retain ACID properties even after a hardware failure. Physically, the transaction log is a log file . With the help of the transaction log, a recovery of the database can be carried out after an error in order to avoid data loss or errors in the database.
See also
literature
- Database systems - concepts and techniques of implementation , Theo Härder and Erhard Rahm, Berlin New York Tokyo, Springer, 2nd edition, 2001