Method of internal data base main-spare synchronization

Abstract

The invention discloses a method for master-backup synchronization of memory databases, which comprises the following steps: initializing the database; registering tables to be synchronized in the database in the out-of-synchronization image registration area, and initially corresponding to the out-of-synchronization image area and out-of-synchronization Queue; when the main database changes, write the change record number of the change table into the out-of-sync queue of the corresponding table, and record the corresponding change in the corresponding out-of-sync image area; send the out-of-sync record to the receiving end database in packages, and at the same time Delete the corresponding records in the out-of-sync image area and out-of-sync queue; the receiver database accepts the data, updates the corresponding table records, and returns a response message; after receiving the synchronization confirmation, continue to synchronize other out-of-sync data. The present invention sets the out-of-sync image area to record the record changes, and then controls the transmission of several steps of records through the transmission control data area, so that all changed records can be recorded and real-time synchronization can be continuously and efficiently performed.

Description

technical field [0001] The invention relates to database backup technology, in particular to a method for synchronously recovering databases. Background technique [0002] The primary technique for dealing with catastrophic failures is database backups. Typically, all databases and logs are periodically backed up on inexpensive storage media. In the event of a catastrophic system failure, the database can be reloaded from the most recent backup and the system started. [0003] To avoid losing all updates since the last backup, you typically want to back up the syslog to an alternate storage medium more frequently. Since the system log is much smaller than the database, it can be backed up more frequently. This way users don't lose all transactions since the last backup. The database can be restored to the state it was in at the time of the crash by recovering all committed transactions in the log. Also, start a new log after every database backup. Therefore, when recov...

AI Technical Summary

Problems solved by technology

The technical solution described in this patent writes all the changed data into the buffer when the master and backup are synchronized. It is quite large, and the buffer is equally large, which will undoubtedly waste space; if the size of the buffer is smaller than the database, there will be a possibility of data loss

Images