Disaster recovery that brings back the whole OpenStack application
Orchestrated, storage-integrated failover across two OpenStack clouds. Trilio recreates your instances, volumes and networks on the second site in one operation, with an RPO you choose, from minutes to near-zero.
- Planned, unplanned and test failover
- No cloud-admin credentials
Try it: run a failover and watch the group come back on Site B in boot order.
- Minutes to near-zeroRPO, chosen per protection group
- One operationto stand up VMs, volumes and networks
- Non-disruptivetest failover that proves the plan
- Self-serviceunder scoped, least-privilege roles
Replicating disks recovers bytes, not a running application
OpenStack has no built-in answer for cross-site disaster recovery. A protected workload is a set of instances, boot and data volumes, flavors, networks and port mappings spread across Nova, Cinder, Neutron and Keystone.
With storage replication alone
Manually rebuilds the application at the remote site during an outage.
- Recreate every instance and flavor
- Re-attach every boot and data volume
- Remap networks and ports manually
- Runbooks that drift and are rarely tested
With Trilio Site Recovery
The whole recovery is orchestrated, not just the storage.
- Instances recreated automatically, in dependency order
- Replicated storage promoted and volumes attached
- Cross-site metadata sync and resource mapping
- Test failovers on a schedule, without touching production
How Trilio Site Recovery works
Four steps take a workload from protected to running on the second site, and back again.
Group
Put the VMs and volumes that fail over together into a crash-consistent protection group.
Replicate
Trilio configures array-level replication between sites and continuously syncs the OpenStack metadata each group needs.
Fail over
Planned or unplanned, Trilio promotes storage on Site B and recreates instances, volumes and network attachments in order.
Fail back
Replication reverses and workloads return to the original site, onto their original VMs and volumes.
Every step runs through scoped OpenStack roles, so recovery is a self-service operation for DR operators, not a shared-admin fire drill.
See Trilio Site Recovery in action
Watch a protection group fail over to a second OpenStack site and fail back again, start to finish.
- Protection groups
- Test failover
- Failover
- Failback
Want to see it on a topology like yours?
Request a live demoChoose the RPO each workload actually needs
Replication mode is set per protection group, so a tier-0 database and a batch tier can share one DR platform at the cost each warrants.
Asynchronous
RPO of minutesScheduled snapshot replication at a configurable interval. The sensible default for most workloads.
Available todayContinuous
Near-zero RPONear-continuous replication with promote-on-failover, for demanding applications that can’t tolerate a scheduled window.
Available todaySynchronous
True zero RPOActive-active stretched storage. The capability is GA on the array; the Trilio integration is planned, not a supported mode today.
On the roadmapWhere teams use Site Recovery
Site and data-center outage
Bring production back on the second site in a single operation when the primary is lost.
Disaster recovery solutionsRansomware and corruption
Recover a protection group to a known-good recovery point instead of a compromised live state.
Ransomware protectionPlanned maintenance and migration
Move workloads between sites for upgrades or data-center moves, then fail back cleanly.
VMware to OpenStack migrationAudited DR testing
Run non-disruptive test failovers on a schedule to satisfy compliance and prove the plan works.
See a test failoverWorkload mobility
Relocate applications between OpenStack regions without rebuilding them by hand.
Application mobility
Why teams choose Trilio Site Recovery
Recover in minutes, not hours
One orchestrated operation stands up VMs, volumes and networks on the second site. No manual runbook.
Prove it before you need it
Test failover validates recovery end-to-end without touching production or pausing replication.
Fail back cleanly
Reverse replication returns each group to its original instances and volumes, not orphaned copies.
No shared admin
DR operators run failover and failback under scoped, least-privilege OpenStack roles, never cloud-admin.
Storage-integrated & consistent
Array replication delivers crash-consistent protection with minimal host and network overhead.
Open and extensible
A vendor-neutral storage layer keeps your DR free of single-vendor lock-in.
Fits the OpenStack you already run
Trilio Site Recovery runs as two containerized services, an API service and an orchestration engine. It authenticates through Keystone and is tested without privileged credentials.
Already protecting OpenStack with backups? See Trilio for OpenStack for point-in-time backup and recovery, or read the documentation.
- Distributions
- Kolla-AnsibleCanonical Charmed OpenStackRed Hat OpenStack on OpenShift (RHOSO)Mirantis MOSK, in progress
- Operator tools
- Native
openstack drCLI and the Trilio DR dashboard in Horizon - Storage
- Everpure FlashArrayIBM FlashSystem (Storage Virtualize), coming soon
- Access model
- Scoped, non-admin project roles through Keystone
- Topology
- Two sites, each an OpenStack cluster with its own storage array
See a full failover on a topology like yours
In a 30-minute demo, a solution architect walks through protection groups, a non-disruptive test failover and a clean failback.