Backing up and restoring management database

The product database is stored on the management node (the one with the admin panel) and backed up automatically. It is also replicated to the nodes included in the HA configuration, if high availability is enabled for the management node.

This management database contains the cluster configuration, service management and monitoring parameters, and metadata of compute objects, such as projects, domains, users, virtual machines, images, flavors, volumes, networks, and other. The database does not contain user data located on storage disks (for example, volumes, volume snapshots, and compute images), as well as data stored on local root disks (for example, logs, metrics data, and service internal data).

Backups of the management database are created automatically via a daily cron job that starts at 3:00 a.m. If the management node is not highly available, restoring such a backup recovers the node in case of failure or database corruption. Backup files are stored in the /mnt/vstorage/webcp/backup/ directory. The retention policy for management node backups is the following:

  • All backups created within the last day are kept.
  • From backups created within the last 7 days, the newest for each day is kept.
  • From backups created within the last 7-14 days, the oldest one is kept.
  • From backups created within the last 14-45 days, the oldest one for each week is kept.
  • Backups older than 45 days are deleted.