The Longhorn team is excited to announce the release of Longhorn v1.11.0. This release represents a major step forward in stability, observability, and scheduling intelligence, while marking an important milestone for the V2 Data Engine, which officially enters the Technical Preview stage.

V2 Data Engine: Entering Technical Preview

With v1.11.0, the V2 Data Engine officially graduates to Technical Preview, reflecting significant improvements in stability, performance, and operational readiness.

Note: Live upgrade is not supported yet for the V2 Data Engine. V2 volumes must be detached before upgrading the engine.

High-Performance Frontend with ublk

Longhorn supports configuring UBLK performance parameters globally, per volume, or via StorageClass to improve I/O performance.

Deprecation of V2 Backing Image

The Backing Image feature for the V2 Data Engine is deprecated in this release and will be removed in v1.12.0. Users are encouraged to adopt Containerized Data Importer (CDI) for volume population, aligning Longhorn with the broader Kubernetes and KubeVirt ecosystem.

Faster Recovery: Smarter Replica Rebuilding

Parallel Rebuilding for V1 Data Engine

Replica rebuilding in the V1 Data Engine is now significantly faster. Instead of rebuilding from a single healthy replica, Longhorn can stream data from multiple replicas in parallel, dramatically reducing rebuild time for fragmented or large volumes.

This improvement shortens exposure windows during failures and helps clusters recover redundancy more quickly.

Smarter Scheduling and Better Resource Balance

Balance-Aware Replica Scheduling

Longhorn introduces a balance-aware disk selection algorithm that considers disk usage across nodes and disks when scheduling replicas.
This reduces uneven capacity distribution, avoids hotspots, and improves long-term cluster health—especially in large or heterogeneous environments.

Topology-Aware Provisioning

With support for Kubernetes StorageClass.allowedTopologies, administrators can now constrain volume provisioning to specific zones, regions, or nodes, enabling better alignment with failure domains and infrastructure layouts.

Disk Health Monitoring and Observability

Node Disk Health Monitoring

Longhorn now actively monitors disk health using S.M.A.R.T. data, allowing early detection of failing disks and enabling proactive maintenance before volumes are impacted.

Administrators can also disable disk health monitoring if required for specific environments.

Expanded Kubernetes Feature Support

ReadWriteOncePod (RWOP)

Longhorn now fully supports the Kubernetes ReadWriteOncePod (RWOP) access mode, enabling stricter single-pod attachment guarantees and improving safety for stateful workloads that rely on exclusive access semantics.

Flexible Networking for Share Manager

Users can now configure an additional network interface for the Share Manager, supporting more complex network segmentation and security requirements for RWX volumes.

Get in touch

As always, these improvements are the direct result of community feedback and collaboration. If you’re new to Longhorn or open source, remember: every contribution counts, and your voice matters. Join the conversation on the CNCF #longhorn Slack channel or GitHub discussions and let us know how you’re using Longhorn v1.11.0!