Schema Backups
Schema backups contain all the data required to reload a schema without accessing its data sources. They contain the data of the initial load as well as the data of all the subsequent activities (e.g., incremental loads, load partitions, etc...).
Backups differ from snapshots :
- they are self contained (the schema definition is included in the backup),
- they load much faster and as such can be used for hot restart of an icCube server in a fault tolerant system.
Create a Backup
Backups are created while schemas are being loaded. To activate the backup you have to define a property in the schema definition as following:
Then each time the schema is being loaded, a backup will be created. Alternatively, you can request a backup for a specific load using the LoadSchema REST API using the forceBackup parameter.
In case the backup cannot be created for whatever reason, this should not fail the load of the schema itself.
Restoring a Backup
Backups can be restored via the icCube Web interface:
as well as via the RestoreBackup REST API.
Configuration (icCube.xml)
The configuration file contains several properties about the backup:
- <backupDirectory>: the directory where the backups are saved
- <backupConfiguration>: contains several properties controlling how the backups are generated (e.g., compression).
Next chapter: Schema Snapshots is explaining what schema snapshots are.