Why your KubeVirt VMs can’t move between clusters — and how EVPN fixes it
Your team has moved VMs to Kubernetes with KubeVirt. The workloads are running. The team is gaining confidence. Then someone The post Why your KubeVirt VMs can’t move between clusters — and how EVPN fixes it appeared first on The New Stack .
KubeVirt allows teams to run VMs on Kubernetes, but moving them between clusters poses challenges. Disaster recovery, cluster upgrades, and capacity rebalancing are reasons for cross-cluster live migrations, yet the network team raises concerns. To maintain IP and MAC addresses on the destination cluster, a stretched Layer 2 domain is required.
This traditionally involves new VLANs, switch configurations, and possibly new hardware, leading to change windows, tickets, and weeks of work. Instead, EVPN (Extensible Virtual Routing Family) and VXLAN address this issue. Two Custom Resource Definitions (CRDs) in the OpenPERouter manage EVPN configuration: Underlay CR for BGP peering between clusters and top-of-rack switches, and L2VNI CR for Layer 2 overlay networks using VXLAN segments.
A second L2VNI CR enables IP routing between subnets and connects clusters to external networks. The migration network, a distinct L2VNI, isolates migration traffic from application traffic. The target VM is provisioned on the destination cluster with its MAC address and IP configuration intact. A migration sender on the source cluster initiates the transfer using the receiver's synchronization addresses.
Once the migration completes, traffic flows to the VM in the new cluster without disruptions. The IP Address Management (IPAM) strategy ensures no conflicts across clusters. OpenPERouter simplifies the process, removing the need for network team involvement and minimizing downtime.
Written by urgent.news from The New Stack's reporting — not their text. Machine-written — it may contain errors, so check the original before relying on it.
