NEC recently reviewed various OpenStack backup solutions at the OpenStack Summit in Sydney as part of an effort to identify the most efficient way to protect cloud environments. In this comprehensive overview, they discuss why we need to protect OpenStack clouds, what to look for when you’re evaluating a solution, and some of the options they encountered during their own research. Watch the 45- minute talk below, or scroll down to see our summary of their talk.
Why You Need an OpenStack Backup Solution
OpenStack was once viewed as a compute-only cloud platform, limited to supporting ephemeral Virtual Machines (VMs). As the technology evolved and companies invested more resources into building a top-notch infrastructure, other business units began to move their workloads to this new environment, widening the variety of applications in use.
Today, as more and more enterprises adopt OpenStack for production purposes, IT workloads are mixed — part stateless and part stateful applications — making data protection a critical component again. Any company storing resources in their OpenStack volumes should give serious thought to backing up that data, not only for disaster recovery but for day-to-day operational recovery as well. With a solid data protection solution, you can easily replicate and restore individual workloads, reuse environments for test/dev purposes, and selectively restore files that disappear or become corrupted.
Why OpenStack Needs Its Own Solution
There are many data protection solutions available for clouds in general, but they’re usually vendor-specific (like Microsoft and VMware) or use a legacy approach to data protection that is ill-fitted to cloud computing. For OpenStack cloud administrators, this poses a problem: how can I find a backup and recovery solution that is as scalable as the cloud I’ve built?
The answer is to find a native OpenStack backup solution that is built on the core constructs of your cloud environment:
- Limitless Scalability: OpenStack has no limit to the number of nodes you can have, but most traditional backup solutions have limitations on the storage and systems they can handle.
- Multi-Tenancy: Most traditional backup solutions manage user access through Active Directory integrations or similar, but OpenStack’s multi-tenant design allows you to create groups of users as tenants residing on your cloud. You need a solution that is different for each tenant, and that allows those tenant to back up the workloads and VMs in their environment.
- Context Awareness: It will take your team too long to piece together a full image from individual user loads that legacy solutions back up: you need network settings, configuration files, services, and audit requirements/trails.
- Distributed Architecture: Most applications are distributed clustered applications and aren’t as simple to back up as single node applications.
- Cloud-Specific Resources: Anything specific to your cloud environment needs to be captured somewhere in a way that can be easily recovered, including configuration logs, cloud databases, and other persistent data.
Given this, traditional data protections simply don’t fulfill these requirements. OpenStack admins need to find a way to make sure that their cloud backups are not scattered in bits and pieces, requiring error-prone manual processes to maintain and recover in the event of backup failure.
Key Considerations for an OpenStack Backup Solution
When evaluating OpenStack backup solutions, you must define the internal requirements for RTO, RPO, and more.
- How often do you want to backup your data?
- How many backups do you want to keep, and for how long?
- Should the backups be kept offsite?
- How often should they be tested?
- What is the reliability of the targets (Ceph, Swift)?
- What is your company’s expected RPO & RTO?
- What are the expectations of your internal SLAs?
WHAT to Back Up:
- Log files (Noah, Neutron, Cinder)
- Security groups
- Storage volumes
- Data & metadata
- Backup type: full backup, differential backup, incremental backup, file-level backup
- Backup control: policy-driven, automatic
- Restore type: one-click restore, selective restore, file level recovery
- Guest OS agents
- Unlimited data transfer
- Consolidated GUI
TrilioVault checks the box for all of these core requirements. It’s the only OpenStack native data protection solution that delivers the full breadth of required functionality for all of your data recovery use cases: operational, DR, one-click backup, and more.
- Multi-tenant, self-service, policy-based solution
- Scalable, agentless
- Non disruptive
- Point-in-time backup and recovery of entire workloads
- Selective restore of VMs to networks, different availability zones, regions, and different clouds
- File-level backup
- Backup of NFS, Ceph, Swift, S3
There are some open source options when it comes to backing up OpenStack, but these solutions fail to capture entire workloads incrementally forever, and typically don’t offer a one-click restore option in the event of a disaster.