In the Manual offline migration using an iSCSI storage pool recipe, we created an iSCSI storage pool and used it while performing manual offline migration. With storage pools, we can delegate the operation of a shared storage to libvirt rather than manually having to log in/log out iSCSI targets, for example. This is especially useful when we perform live migrations with the
virsh command, as we are going to see in the next recipe. Even though the use of storage pools is not required, it simplifies and centralizes the management of backend volumes.
In this recipe, we are going to use the GlusterFS network filesystem to demonstrate an alternative way of manually migrating a KVM instance, this time not using storage pools.
GlusterFS has the following two components:
- Server component: This runs the
GlusterFSdaemon and exports local block devices named bricks as volumes that can be mounted by the client component
- Client component: This connects...