Database data recovery

  • Method : Establish redundant data and use the redundant data to recover to a certain moment b, and then use the registration log file to resubmit a series of operations during this period.
  •  The most commonly used techniques for establishing redundant data are: data dump ( ie backup ) and log file registration.

           Data dumps are divided into dynamic dumps ( high efficiency, achieving concurrency, but error-prone, and the operation of each transaction during the dump must be recorded in conjunction with the log file ) and static dumps ( not easy to make mistakes, depending on whether other transactions are allowed to run during the dump ) Low efficiency ); It is also divided into massive dumps              and incremental dumps according to the amount of dumps . (So there are four methods for permuting and combining the available data dumps )

  • Recovery strategy: The corresponding strategy should be selected according to the type of failure. That is, transaction failure-undo the operation of the transaction; system failure-undo is not completed, redo has been completed; media failure-reinstall the database, redo has been completed

Guess you like

Origin www.cnblogs.com/jhwlxx/p/12689622.html