Steps of the online DB backup process

Online backup enables the database to be backed up during operation. As the database comprises a series of mutually dependent files, the Data Manager may not write any data to the database during backup.

Online backup comprises the following steps:

1. The destination of the backup and the archive in question are set via a data point (this is usually set in the service panels).

2. The Data Manager then ends the DataBG and closes the database. Data in the cache is written prior to this.

3. The Data Manager generates the input file for the backup program WCCOAtoolMedia from the backup settings and starts the program.

4. Messages are written to a buffer during backup. Exceptions in this case are queries relating to historical values and _archive configs, these are answered with an error message. Connects of the managers are only allowed if they do not require any configs. In this case, the connection to the manager is closed immediately.

5. The Data Manager monitors the size of the buffer, the virtual memory and the time during backup. If the size of the buffer or the available virtual memory exceeds a configurable value or if a defined timeout has been exceeded, backup is aborted early. This will prevent the Data Manager from landing in the emergency mode due to the backup (see Data manager emergency mode, basics).

6. When backup is completed, the Data Manager opens the database again, restarts the DataBG and writes the contents of the output file of WCCOAtoolMedia into data points.

No special measures are taken for re-backup as this either affects only completed file changes or the living database, that is, the current values, last values and configuration data. The project must be restarted to restore the latter. Completed file changes must be swapped in the usual manner.