OpenStack Backups Summary
In OpenStack, backups are crucial for ensuring data integrity and recovering from failures. All Glance snapshots and images are periodically backed up into a tape archive to ensure long-term protection and disaster recovery.
Backups typically involve two key components:
Glance Snapshots (repository of images)
- Glance snapshots capture the state of a running or stopped instance's root disk.
- They create an image of the instance’s disk, which can be used to launch new instances or restore the current state.
- Snapshots are stored in Glance and can be viewed with
openstack image list
. - Keep in mind: If the instance is booted from a volume, the snapshot will be created on the volume only — not in the Glance image repository. However, you will see a snapshot listed in the images repository with a size of 0, which consists only of metadata about block device mappings.
- When launching a new instance from the snapshot, OpenStack will recreate the original Cinder volumes if they were part of the snapshot.
- If you want to store the root disk on a volume in the image repository, you need to create a volume from the snapshot and then upload the volume to an image.
Cinder Volume Backups
- Cinder backups create a full copy of a volume that can be restored or cloned.
- The backup service stores the volume data independently, allowing recovery even if the original volume is lost. Backup service at PSNC is stored on independet S3 backend.
Best Practices:
- Automate backups using OpenStack CLI or Heat and Mistral templates.
- Schedule regular backups to minimize data loss.
- Store backups in a separate storage backend for added resilience.
- Monitor backup status and test recovery procedures regularly.