CloudCasa by Catalogic has added Dell PowerStore to the storage arrays CloudCasa Disaster Recovery can drive for Kubernetes failover, alongside HPE Alletra and SUSE Storage. The same feature update adds a read-only Model Context Protocol (MCP) server that lets AI assistants check backup and DR status, and a preview of Changed Block Tracking (CBT) for KubeVirt virtual machine backups, which copies only the blocks that changed since the last run.
Storage-Integrated Disaster Recovery Expands to Dell PowerStore #
PowerStore joins the HPE Alletra Storage MP B10000, HPE Alletra 9000, and SUSE Storage as replication providers for CloudCasa DR. The division of labor is the same on each platform: the array replicates persistent volume data to the recovery site, synchronously or asynchronously in PowerStore’s case, and CloudCasa handles the Kubernetes side of failover and failback for containerized applications and KubeVirt VMs. Volume data moves array to array and never passes through CloudCasa.
CloudCasa records the configuration of each protected workload in a replicated inventory volume and talks to the array through the PowerStore REST API. On failover, it brings up that inventory volume first, starts a recovery agent on the target cluster, then fails over the remaining PowerStore volume groups and rebuilds the Deployments, StatefulSets, persistent volume claims, and KubeVirt VMs against the replicated volumes. Because PowerStore fails over complete volume groups, CloudCasa advises placing workloads with different recovery requirements in separate groups.
Failback waits for operator confirmation, after which CloudCasa reverses replication on the affected volume groups so PowerStore resynchronizes only the changes made since the failover, scales down the workloads still running at the DR site, and then fails the volume groups back to the primary array and rebuilds the workloads there. CloudCasa says the approach targets recovery times measured in minutes and near-zero RPO when PowerStore replicates synchronously, with results depending on replication mode, workload size, and distance between sites.
The integration covers block (SCSI) persistent volumes provisioned through the PowerStore CSI driver; NFS and SMB volume DR is on CloudCasa’s roadmap. Supported platforms include Red Hat OpenShift and OpenShift Virtualization, SUSE Rancher/RKE2 and SUSE Virtualization, VMware vSphere Kubernetes Service, and upstream Kubernetes 1.20 and later. CloudCasa DR is licensed per worker node on an annual subscription, separately from CloudCasa’s backup product, in both the SaaS and self-hosted editions.
Model Context Protocol Server for AI Agent Integration #
The MCP server lets any AI assistant that supports remote MCP servers over HTTP query CloudCasa in plain language, with questions such as which backup jobs failed in the last 24 hours or which recovery points exist for a namespace. It calls the CloudCasa APIs on the user’s behalf, authenticates with a CloudCasa API key passed in an HTTP header, and can do only what that key’s role allows. For now it is read-only: assistants can list clusters, backup policies, jobs, recovery points, restores, alerts, and backup targets, but cannot run, create, edit, or delete anything.
“Platform teams do more of their day-to-day work using AI assistants, and they want backup status available alongside the rest of their tooling,” said Ryan Kaw, Global VP of Sales and Alliances at CloudCasa. “The MCP server makes that possible today.”
Changed Block Tracking Preview and Availability #
The CBT preview builds on the incremental backup support KubeVirt added in version 1.8, which uses QEMU and libvirt backup capabilities and works regardless of the underlying storage. CBT records which disk blocks changed since the previous backup, so CloudCasa copies only those blocks, which the company says shortens VM backup times. On Red Hat OpenShift, that KubeVirt version arrives with OpenShift 4.22, and the feature has to be enabled explicitly.
The PowerStore DR integration and the MCP server are available now, and CBT backups remain in preview. The update also adds a dedicated VMs page that lists virtual machines across all registered clusters with their running state and guest agent status, where admins can define backups and restores for each VM.