In RHEV 3.3, placing host in maintenance mode fails and VMs cannot be migrated due to resource lock errors

Issue

• Tried putting a host into maintenance to upgrade the hypervisor, but RHEV-M shows Alert Power Management test failed.

VdcBLLException: org.ovirt.engine.core.vdsbroker.vdsbroker.VDSNetworkException. 

Then when trying manually to migrate VMs off the hypervisor it complains that the VM is alreade being migrated, with below error in engine.log

18:05:47,228 WARN  [org.ovirt.engine.core.bll.InternalMigrateVmCommand] (pool-4-thread-47) [7071087c] CanDoAction of action InternalMigrateVm failed. Reasons:VAR__ACTION__MIGRATE,VAR__TYPE__VM,ACTION_TYPE_FAILED_VM_IS_BEING_MIGRATED,$VmName os-node-01