Migrating KVM instances is the process of sending the state of the guest virtual machine's memory, CPU, and virtualized devices attached to it, to a different server. Migrating KVM instances is a somewhat complicated process, depending on what backend storage the VM is using (that is, directory, image file, iSCSI volume, shared storage, or storage pools), the network infrastructure, and the number of block devices attached to the guest. There are following the two types of migrations as far as libvirt is concerned:
- Offline migration involves downtime for the instance. It works by first suspending the guest VM, then copying an image of the guest memory to the destination hypervisor. The KVM machine is then resumed on the target host. If the filesystem of the VM is not on a shared storage, then it needs to be moved to the target server as well.
- Live migration...
