For the high availability of Openstack

most of the current research is focused on the high availability of control nodes

and there is little research on high availability of virtual machines. There are some problems in the existing high availability solutions of virtual machine

for example

these schemes unable to adapt to a variety of failure scenarios

and unable to deal with failure of virtual machine granularity. In this paper

a high availability scheme of virtual machine based on Openstack is proposed. The scheme uses control component and agent component

designs different fault detection logic

and realizes high availability of virtual machine in various fault scenarios. At the same time

the concept of autonomous group is introduced to solve the problem that the agent component can not detect and report faults in the scenario of agent component failure or management network failure.