A database instance can fail due to hardware or logical errors.
To restore the database instance, you have to import data and log backups (see Backing Up and Restoring). If you have a lot of data to import, restoring the database instance can take a long time. The database instance is not available for use during this time.
To make the database instance available for use more quickly (high availability), you can replicate the data of the active database instance (original instance). Replication means that the active database instance is first duplicated and the duplicate then updated at regular intervals.
Replication Mechanisms
Name |
Description |
Required Database Tools and Additional Software |
The active database instance is copied once, but not updated thereafter A database copy is not a replication mechanism, but is required for setting up certain replication mechanisms, such as a standby database. |
MaxDB Database Manager |
|
You update a duplicate of the active database instance asynchronously. You copy the log backups of the original instance to the standby instance and import them into it. When the active database instance fails, you can resume operations quickly with the standby instance after importing the last redo log entries. |
MaxDB Database Manager |
|
Multiple duplicates of the database instance are automatically updated synchronously. All instances access the same log area. When a database instance fails, a standby instance is automatically activated. |
MaxDB Database Manager Cluster Software Memory management system |
|
The Synchronization Manager synchronizes the table contents of client database instances asynchronously with the table contents of a central database instance. The client computers can be offline in the meantime. Collisions are handled automatically. You have the option of filtering data. Prerequisite: the amount of changed data as a proportion of all data must be small. |
MaxDB Synchronization Manager MaxDB Database Manager |
See also:
High Availability for the MaxDB Database