Block storage for KubeVirt virtual machines on Red Hat® OpenShift® — live migration, VM snapshots, and sub-millisecond latency for VM disk I/O.
Red Hat® OpenShift® Virtualization uses KubeVirt to run virtual machines as Kubernetes workloads — consolidating container and VM infrastructure on a single platform. Storage for OpenShift Virtualization has requirements that differ from container storage: VM disks need block semantics, live migration requires storage-level support, VM snapshots carry different consistency expectations, and VM disk I/O patterns put different demands on the storage layer than container workloads. Simplyblock provides NVMe block storage purpose-built for these requirements — live migration across nodes, instant copy-on-write VM disk snapshots, and the sub-millisecond latency that VM operating systems and databases running inside VMs require. All integrated via the standard Kubernetes CSI API without manual storage administration.
The VM Storage Challenge
Virtual machines have distinct storage requirements that go beyond what most Kubernetes container storage is designed for.
KubeVirt VM live migration moves a running VM from one node to another without downtime. This requires either ReadWriteMany block volumes or storage that supports live migration natively. Most container-optimized storage — including local PVs and many CSI drivers — does not meet this requirement, making live migration impossible.
Virtual machines running full guest operating systems generate I/O patterns — boot sequences, swap activity, OS-level disk scheduling — that differ significantly from containerized applications. Storage optimized only for container workloads may not deliver acceptable performance for VMs running databases, ERP systems, or Windows applications.
Taking a consistent snapshot of a VM disk requires coordination between the hypervisor, the guest OS, and the storage layer. Storage that only supports crash-consistent PVC snapshots may not be sufficient for VM workloads that require application-consistent backup points.
When VM disk I/O and container persistent volumes share the same storage backend without isolation, a VM running a backup job or a high-I/O database can degrade latency for container workloads on the same storage pool — and vice versa.
How Simplyblock Addresses It
NVMe block storage with live migration support, VM disk snapshots, and QoS isolation — integrated with OpenShift Virtualization via standard CSI without manual storage configuration.
Simplyblock provides the storage semantics that KubeVirt VM live migration requires: ReadWriteMany block volumes and block storage live migration via the Kubernetes CDI (Containerized Data Importer) API. VMs can migrate across nodes in an OpenShift cluster without storage-layer downtime or manual volume reconfiguration. This enables node maintenance, cluster updates, and workload rebalancing without VM downtime. See the KubeVirt storage page for the full live migration architecture details.
Simplyblock's instant copy-on-write snapshots integrate with the Kubernetes CSI snapshot API used by OpenShift Virtualization for VM disk operations. VM disk snapshots complete in milliseconds regardless of disk size — making frequent snapshot schedules practical without the I/O overhead that thick-copy snapshots impose. Cloning a VM disk creates a writable copy-on-write clone that is immediately usable. See the data protection page for snapshot scheduling and recovery workflows.
Simplyblock delivers sub-millisecond P99 latency for VM disk I/O — important for guest operating system performance, database workloads running inside VMs, and Windows application workloads that are sensitive to storage latency. Per-volume IOPS and throughput limits prevent individual VMs from degrading storage performance for other workloads. See the storage QoS page for how per-PVC isolation applies to mixed VM and container workloads.
What Teams Gain
VM storage that enables live migration, supports consistent snapshots, and delivers the latency VM workloads require — without a separate storage system for VMs and containers.
Teams migrating workloads from VMware to OpenShift Virtualization can preserve live migration as an operational capability — without depending on a SAN or changing the storage architecture to support it.
Copy-on-write snapshots complete in milliseconds and do not require scheduled maintenance windows. VM backup schedules that previously required off-hours windows can run continuously without impacting VM performance.
Sub-millisecond NVMe latency gives guest operating systems, databases running inside VMs, and Windows application workloads the storage performance they need — not the throttled performance that container-optimized storage delivers to VM disk I/O.
Simplyblock serves both container persistent volumes and KubeVirt VM disks from the same storage pool with per-workload isolation. Teams eliminate the operational burden of managing separate storage systems for virtual machines and containerized applications.
Simplyblock's hyperconverged topology runs on the same nodes as OpenShift workloads and VMs. Dedicated storage nodes are optional. Teams growing their OpenShift Virtualization cluster add nodes that serve as both compute and storage.
Teams running OpenShift Virtualization on bare metal or on-premises get the same live migration support and VM snapshot capabilities as cloud deployments — without cloud storage dependencies. See OpenShift bare metal storage for the full architecture.
Run the numbers
Enter your usable capacity to see the hardware and compute split for VM workloads, with erasure coding at 1.5× raw instead of 3× for replication. No sign-up, and the result is printable.
Yes. Simplyblock provides ReadWriteMany block volumes and supports the CDI (Containerized Data Importer) live migration API used by KubeVirt. VMs can migrate across nodes without storage-layer downtime. See the KubeVirt storage page for the complete migration architecture.
Simplyblock implements the Kubernetes CSI VolumeSnapshot API. OpenShift Virtualization uses this API for VM disk snapshot and restore operations. Simplyblock snapshots are copy-on-write and complete in milliseconds regardless of disk size — making frequent snapshot schedules practical without performance impact.
Yes. Simplyblock provides standard block storage volumes consumed by VM disks via the Kubernetes CSI driver. The guest operating system — whether Linux or Windows — sees a standard NVMe block device. There is no guest-OS-specific configuration required.
Yes. Simplyblock serves container persistent volumes and KubeVirt VM disks from the same storage pool. Per-volume QoS limits prevent individual workloads from monopolizing storage I/O, enabling safe colocation of VMs and containers on shared storage.
Yes. Simplyblock runs on bare metal OpenShift clusters without cloud provider dependencies. VM live migration, snapshots, and NVMe performance are all supported on bare metal. See the OpenShift bare metal storage page for the bare metal-specific deployment architecture.
Both provide storage for OpenShift Virtualization. Simplyblock delivers lower latency (NVMe-native vs. Ceph file system overhead), lighter resource consumption on the same nodes, and no ODF licensing cost stacked on top of the OpenShift subscription. See the OpenShift ODF alternative page for a direct comparison.
Cost is driven mainly by two things: how much raw capacity the protection scheme needs, and how much compute the storage latency frees up. A 4+2 erasure coding scheme needs 1.5x raw for the usable capacity you deploy, versus 3x for three-way replication, which is roughly half the NVMe drives for the same fault tolerance. Because simplyblock runs on the compute nodes, there is also no separate storage tier to buy. Enter your usable capacity in the TCO calculator to see the three-year hardware and compute split for your own estate.
Ask your AI assistant to compare storage options for OpenShift Virtualization KubeVirt workloads — ODF/Ceph, local PVs, external SAN, and simplyblock — on live migration support, snapshot capability, and VM disk latency.
Free, no obligation
Book a 30-minute call with a simplyblock storage architect. We will help you cut storage cost and hit your IOPS and latency targets on Kubernetes, with practical recommendations you can act on right away.
Book my session