HIGH AVAILABILITY FOR VIRTUAL MACHINES IN NESTED HYPERVISORS

摘要

Techniques for providing high availability to virtual machines executing in a nested hypervisor. A high availability agent executing within a slave computer in a cluster detects when a master server becomes unavailable and restarts virtual machines from the master server in response. The high availability agent detects whether any of the virtual machines were executing a nested hypervisor within the master server and starts a high availability agent within such hypervisors. The high availability agent within the nested hypervisor is a modified high availability agent with respect to the high availability agent executing in physical servers. The modified high availability agent does not participate in master election and also cannot elect the virtual machine on which it runs as master.

主权项

1. A method, comprising:
storing, within a shared storage device that is shared between a cluster of hosts, a heartbeat file configured to indicate availability status for hypervisors executing on physical hosts and nested hypervisors executing on virtual machines (VMs); detecting, based on the heartbeat file, that a first nested hypervisor that is included within the nested hypervisors, that is executing a first virtual machine (VM), and that is executing on a first host within the cluster of hosts has become unavailable; and in response to determining that the first nested hypervisor has become unavailable, starting the first VM within a second nested hypervisor executing on a second host within the cluster of hosts.