Checkpoint method for a highly available computer system
AssigneeStratus Technologies, Inc.
InventorLei Cao
Life of this patent
Estimating the term…
Abstract
A HA computer system is configured to support the operation of an active and a standby virtual machine, and the active virtual machine supports one or more first computer processes and the standby virtual machines supports duplicates of the one or more first computer processes. Each active virtual machine monitors and stores a state of the first computer processes and periodically pauses to only synchronize some of the state of the first computer processes running on the active VM with state associated with the duplicate processes running on the standby VM stored in a first region of virtual memory.
Claims17
A method of performing a checkpoint operation, comprising:
- configuring a first virtual machine running on an active host computer to allocate one or more pages from a first virtual memory region to a first process, and to allocate one or more pages from a second virtual memory region to a second process, the first virtual memory region being backed by physical memory pages of a first size that is smaller than physical memory pages of a second size backing the second virtual memory region;
- and
- controlling the operation of the first virtual machine to synchronize only those physical memory pages allocated to the first process with physical memory pages allocated to a duplicate process maintained in association with a second virtual machine running on a standby host computer.
A method for partially synchronizing an operating state of a first virtual machine with a second virtual machine;
- configuring the first virtual machine running on an active host computer to identify a first and a second virtual memory region, the first virtual memory region being backed by physical memory pages of a first size that are smaller than physical memory pages of a second size backing the second virtual memory region;
- allocating one or more pages from the first virtual memory region to a first process running on the first virtual machine, and allocating one or more pages from the second memory region to a second process running on the first virtual machine, and
- controlling the operation of the first virtual machine to only synchronize those physical memory pages allocated to the first process with physical memory pages allocated to a duplicate process maintained in association with the second virtual machine running on a standby host computer.
A highly available computer system, comprising:
- an active host computer configured to support the operation and maintaining state of a first virtual machine on which is running a plurality of processes under control of a virtual machine manager;
- a standby host computer supporting the operation of a second virtual machine that maintains state associated with a plurality of processes that are duplicates of the processes running on the active host computer;
- a checkpoint engine that is configured to only synchronize a portion of the state of the first virtual machine with the state of the second virtual machine;
- wherein, the operational state associated with a first one of the plurality of the processes running on first virtual machine is maintained in a first region of virtual memory, the operational state of a second one of the processes running on first virtual machine is maintained in a second region of the virtual memory, and the first virtual memory region is backed by physical memory pages of a fist size that are smaller in size than physical pages of a second size backing the second virtual memory region;
- and
- the portion of the virtual machine state that is synchronized is one or more physical pages maintained in the first region of virtual memory.
Dependencies are inferred from the claim text.