At the recent Red Hat Summit, the sentiment toward OpenShift-Virtualization was overwhelmingly positive, with one massive “but”: Live Migration. For many, moving stateful workloads between nodes feels like a high-wire act. When the storage layer can’t keep up with the I/O demands of a migrating VM, latency spikes, the “stun” period lasts too long, and migrations fail. The common refrain? “I love OpenShift-Virt, but getting Live Migration to work is a nightmare.”
Our response is simple: You’re just using the wrong storage.
The Torture Test: 1.2 Million IOPS
We didn’t want to show you a “lab-clean” migration. We wanted to show you what happens when the infrastructure is screaming.
The Setup:
- 3 Nodes running OpenShift-Virt.
- 14 VMs distributed across the cluster.
- The Load: Each VM is pushing ~75K IOPS, totaling 1.2 Million IOPS—all hitting just 3 Lightbits storage nodes.
Beyond Just Moving Data
In the video below, we demonstrate that even under this massive synthetic load, storage management remains instantaneous. Using the Lightbits CSI integration, we perform a series of multi-tenant operations:
- Instant Volume Creation: No waiting for provisioning.
- Zero-Copy Snapshots & Clones: Creating clones directly from volumes and snapshots without impacting the 1.2M IOPS baseline.
- Cleanup: Deleting snapshots and volumes on the fly.
The “Aha” Moment: Live Migration under Load
The climax of the demo is the Live Migration. Watch the storage metrics closely: as OpenShift-Virt triggers the move, Lightbits seamlessly handles the NVMe-over-TCP multi-pathing.
You will see the storage connection expand to the target node while the source node is still active. Once the memory state handoff is complete, the old connection is pruned. The latency? Stable. The VM? Unaware it just jumped hosts.
Watch the Demo
Stop fighting your storage and start scaling your virtualization. See how Lightbits turns the “Live Migration nightmare” into a non-event. Watch the movie:
