How Backup/Restore Works

Backup/Restore Processing

Appliance backup uses a system utility such as rsync to copy appliance data as a complete snapshot (the image of your data at the time of backup). The utility determines the difference between the data on the appliance and the previous backup data on the remote backup device before the data transfer.

• First backup - all data is copied
• Subsequent backups - only the difference since the previous backup is copied. This saves appliance resources and network bandwidth. The snapshot on the backup system is updated to include the differences copied in the latest incremental backup.

Data collection continues during backup processing.

The backup system creates a directory named for the IP address of the backed-up appliance and stores its backup data there. This allows multiple appliances to be backed up to a backup system large enough to hold their cumulative data, as long as each appliance has a unique IP address.

After each backup completes, both data sets (the original appliance and the backup server data set) are identical. When restoring an appliance, the backed up information is copied back to the appliance. Processing does not continue during a restore process. For the duration of restore operation, all engines that access BFQ or MySQL (except engine_backup and engine_tomcat) are stopped. All log processing is resumed after the restore is completed.

Failure Situations & Workarounds

If the appliance crashes during a backup, the data in previous successful backups is protected. Each backup, including incremental backups, is written to a new location so no overwriting of existing backup data occurs.