When we have multiple share driver backends, create a shared instance,
and the list of'self.gluster_used_vols' will only be updated on the
current node. If the RPC request to delete the share instance is sent
to other nodes,'self.gluster_used_vols' will be Cannot find the
information of the glusterfs volume we want to delete, so we need to
update'self.gluster_used_vols' when deleting the instance
Change-Id: I14835f6c54376737b41cbf78c94908ea1befde15
Closes-Bug: #1894362