New features

Each release offers new functionalities and improvements. Read this information for a high-level summary of the new features introduced in the Fusion Data Foundation 4.21 release.

CNSA integration with Fusion Data Foundation

IBM Fusion 2.13 introduces integration between Fusion Data Foundation and Container Native Storage Access (CNSA) by using the remote mount feature. This feature enables Fusion Data Foundation to deploy and manage the required IBM Storage Scale or Global Data Platform resources during installation. It also allows administrators to configure and attach external IBM Storage Scale clusters directly from the Red Hat® OpenShift® Container Platform 4.21 user interface.

The integration simplifies the installation, upgrade, and lifecycle management of remote mount functionality across IBM Fusion deployments, supports seamless upgrades and migrations from earlier releases, and ensures continuity for existing workloads while enabling a unified, UI‑driven storage management experience.

Erasure coding support in Fusion Data Foundation

IBM Fusion 2.13 adds erasure coding support to Fusion Data Foundation through the Storage efficiency configuration option. This feature improves usable storage capacity while maintaining strong data protection, providing an alternative to replica-3 storage configurations.

With this capability, administrators can select the Storage efficiency option during storage configuration to apply erasure coding. Erasure coding splits data into data and parity chunks that are distributed across storage nodes, delivering fault tolerance with significantly higher usable capacity compared to performance-optimized configurations. Use this option for workloads such as archival storage, backups, and other capacity-driven use cases.

Erasure coding requires a minimum number of storage nodes, based on the selected data and parity layout. The minimum supported configuration requires six storage nodes. For example, a 4+2 configuration requires at least six nodes, while an 8+2 configuration requires at least ten nodes.

Important:
  • Erasure coding is supported starting with Red Hat OpenShift Container Platform 4.21 and later and Fusion Data Foundation 4.21 and later.
  • This feature is supported only for Block storage (RBD).
  • Clusters upgraded from earlier releases do not support erasure coding.

Disaster Recovery

Multi-volume consistency for disaster recovery

Multi-volume consistency group support is now fully supported again for both, RBD and CephFS volumes. Consistency groups are used by default for all newly protected applications. Already protected non-CG applications on an upgraded cluster applications should be unprotected and protected again to migrate to consistency groups.

Support for CephFS volumes using non‑default StorageClasses

Fusion Data Foundation DR solution supports data replication for CephFS volumes that use multiple or non‑default StorageClasses, such as replica‑2 file storage. This enhancement enables disaster recovery for workloads backed by customized CephFS configurations, including cases where different pools or secrets require separate StorageClass, VolumeSnapshotClass, and VolumeReplicationClass resources.

Enhanced disaster recovery capabilities in the RHACM KubeVirt UI

Disaster recovery integration for virtual machines

The Red Hat Advanced Cluster Management for Kubernetes (RHACM) KubeVirt UI now integrates Fusion Data Foundation disaster recovery, allowing users to view VM protection status and initiate DR protection directly from the VM view. This capability is available for both GitOps‑managed and discovered VMs, with disabled actions providing guidance when prerequisites are not met.

Improved visibility for failover and relocate operations

The RHACM UI now surfaces detailed, step‑by‑step progression for disaster recovery failover and relocate workflows using progression data from Ramen. Enhanced modals show each operation phase, status indicators, and any associated errors, improving transparency, troubleshooting speed, and overall observability of DR processes.