OpenStack is an open-source software platform for cloud computing that enables organizations to manage private and public cloud environments. It provides Infrastructure as a Service (IaaS), allowing users to provision and manage compute, storage, and networking resources in a flexible and scalable manner.
Understanding OpenStack Concepts
The following OpenStack services and architecture concepts are essential for backup and restore operations:
Concept / Service | Description |
Keystone (Identity Service) | Handles user identity, authentication, projects, and permissions. Validates user authorization to perform API operations. |
Nova (Compute service) | Provisions and manages virtual machines (referred to as Nova instances or servers). |
Cinder (Block Storage Service) | Manages persistent block storage volumes and block-level snapshots. |
Glance (Image Repository Service) | Stores base operating system images (for example, Ubuntu and Windows) used as boot templates. |
Neutron (Network Service) | Manages virtual networks, routers, floating IP addresses, and security groups. |
Project | A logical grouping of resources. Every virtual machine, volume, and network belongs to a specific project. |
Region | A Region is a top-level geographic or administrative deployment of OpenStack services. Each Region has its own independent API endpoints for Nova, Cinder, Neutron, Keystone, etc. Regions are completely independent of each other. |
Availability Zone (AZ) | Physical fault domain subdivisions (such as separate racks or power sources) within a region. A region contains one or more AZs. |
❗ Important
Currently, Canonical OpenStack is supported
Backing up and restoring OpenStack VMs
Druva performs agentless, image-level backups of OpenStack virtual machines (VMs) with the help of a dedicated virtual machine known as the backup proxy.
Some of the important concepts are:
A Project can contain virtual machines across multiple regions.
Discovery of virtual machines is performed at the project level, meaning VMs will be discovered from all the regions.
To back up or restore VMs in a specific region, a backup proxy must be deployed in that region.
You can restore full virtual machines to the same or alternate projects or regions.
Other capabilities
Agentless protection: Protects VMs directly at the storage layer using native Cinder snapshots and therefore no software installation is required inside guest operating systems.
Direct integration: Leverages native OpenStack APIs to manage backup and recovery workflows seamlessly.
Efficiency: Deduplicates and compresses backup data locally on the proxy before securely transferring it over HTTPS to the Druva Cloud.
Flexible execution: Supports both scheduled, automated backups and on-demand manual execution.
High Availability support: Protects virtual machines spread across multiple Availability Zones in a Region, ensuring continuous backup operations during local infrastructure outages.
OpenStack Architecture
The above infographic illustrates the architecture for direct-to-cloud backups.
To facilitate virtual machine backups and restores, a Druva backup proxy must be deployed within the target OpenStack project.
At least one backup proxy is required in every region where the project hosts virtual machines.
The proxy discovers virtual machines within the project and processes backup and restore operations.
The proxy performs source-side deduplication so that only unique data blocks are transmitted across the network.
Once the transfer is complete, the backup data is securely committed to the Druva Cloud.