Synchronization of physical disks

• US 10,162,873 B2
• Filed: 12/21/2012
• Issued: 12/25/2018
• Est. Priority Date: 12/21/2012
• Status: Active Grant

• Current Assignee: Red Hat, Inc. (International Business Machines Corporation)
• Original Assignee: Red Hat, Inc. (International Business Machines Corporation)
• Inventors: Bonzini, Paolo
• Primary Examiner(s): Obisesan, Augustine K.
• Application Number:
  US13/724,326
• Publication Number:
  US 20140181015A1
• Time in Patent Office: 2,195 Days
• CPC Class Codes: G06F 16/273 Asynchronous replication or reconciliation

Abstract

In a process for migrating a virtual machine'"'"'s storage from a source disk to a destination disk, during a steady state (i.e., wherein the contents of the virtual machine stored on the source disk and the destination disk are equal), a virtual machine monitor receives a set of write requests from a guest operating system (“guest”) of the virtual machine, provides confirmation of the completion of the set of writes to the source disk, and asynchronously replicates the set of write requests to the destination disk. Upon receipt of a flush request from the guest, the virtual machine monitor confirms completion of the flushing of the destination disk following replication of the write requests to the destination disk. Upon receipt of a switch request from a virtual machine manager, the virtual machine monitor switches the virtual machine to the destination disk and issues subsequent write requests to the destination disk.

Claims

1. A method comprising:

identifying, by a processing device executing a virtual machine monitor, a steady state wherein a first count of writes associated with a virtual machine stored on a source physical disk and a destination physical disk are equal;

receiving, by the virtual machine monitor, a set of write requests comprising data from a guest operating system of the virtual machine;

providing, by the virtual machine monitor to the guest operating system, confirmation of a completion of a set of writes to the source physical disk;

following providing to the guest operating system the confirmation of the completion of the set of writes to the source physical disk, asynchronously replicating, by the virtual machine monitor, the set of write requests to the destination physical disk;

receiving, by the virtual machine monitor, a flush request from the guest operating system of the virtual machine, wherein the flush request comprises an instruction to the destination physical disk to write the data pending in cache of the destination physical disk to persistent storage of the destination physical disk;

determining a second count of writes completed on the source physical disk at a time of a previous flush request is less than or equal to the first count of writes identified at the steady state;

flushing, by the virtual machine monitor, the destination physical disk in response to the determining;

providing, by the virtual machine monitor, to the guest operating system, confirmation of completion of the flushing of the destination physical disk following replication of the set of write requests to the destination physical disk and writing of the data to persistent storage of the destination physical disk;

receiving, by the virtual machine monitor, a switch request from the virtual machine manager; and

switching, by the virtual machine monitor, the virtual machine to the destination physical disk and issuing subsequent write requests to the destination physical disk.

2. The method of claim 1, further comprising reporting to the virtual machine manager, by the virtual machine monitor, entry into the steady state.

3. The method of claim 1, wherein identifying the steady state further comprises:

maintaining a dirty bitmap to record a disk sector that have been written to on the source disk but not the destination disk; and

emptying the dirty bitmap.

4. The method of claim 1, further comprising operating in a synchronization phase prior to the steady state, wherein the synchronization phase comprises migrating the contents of the source disk to the destination disk.

5. The method of claim 4, further comprising entering the synchronization phase in response to at least one of receiving a request from the virtual machine manager or detecting an error in the replication of the set of write requests to the destination disk.

6. The method of claim 4, wherein the source disk is divided into a base image and an overlay image, and wherein only contents of the overlay image are migrated to the destination disk during the synchronization phase.

7. The method of claim 1, further comprising delaying processing of a write request to the source disk for a set period of time upon determining the contents stored on the source disk and the destination disk are equal.

8. A non-transitory computer readable storage medium comprising instructions that, when executed by a processing device, cause the processing device:

identify, by the processing device executing a virtual machine monitor, a steady state wherein a first count of writes associated with a virtual machine stored on a source physical disk and a destination physical disk are equal;

receive, by the virtual machine monitor, a set of write requests comprising data from a guest operating system of the virtual machine;

provide, by the virtual machine monitor to the guest operating system, confirmation of a completion of a set of writes to the source physical disk;

following the confirmation of the completion of the set of writes to the source physical disk being provided to the guest operating system, asynchronously replicate, by the virtual machine monitor, the set of write requests to the destination disk;

receive, by the virtual machine monitor, a flush request from the guest operating system of the virtual machine, wherein the flush request comprises an instruction to the destination physical disk to write the data pending in cache of the destination physical disk to persistent storage of the destination physical disk;

determine a second count of writes completed on the source physical disk at a time of a previous flush request is less than or equal to the first count of writes identified at the steady state;

execute, by the virtual machine monitor, a flush of the destination physical disk in response to a determination that the second count of writes is less than or equal to the first count of writes;

provide, by the virtual machine monitor to the guest operating system, confirmation of completion of the flush of the destination physical disk following replication of the set of write requests to the destination physical disk and writing of the data to persistent storage of the destination physical disk;

receive, by the virtual machine monitor, a switch request from a virtual machine manager; and

switch, by the virtual machine monitor, the virtual machine to the destination physical disk and issuing subsequent write requests to the destination physical disk.

9. The non-transitory computer readable storage medium of claim 8, the processing device to report to the virtual machine manager, by the virtual machine monitor, entry into the steady state.

10. The non-transitory computer readable storage medium of claim 8, the processing device to:

maintain a dirty bitmap to record a disk sector written to on the source physical disk but not the destination physical disk; and

empty the dirty bitmap.

11. The non-transitory computer readable storage medium of claim 8, the processing device to operate in a synchronization phase prior to the steady state, wherein the synchronization phase comprises migrating the contents of the source physical disk to the destination physical disk.

12. The non-transitory computer readable storage medium of claim 11, the processing device to enter the synchronization phase in response to at least one of receiving a request from the virtual machine manager or detecting an error in the replication of the set of write requests to the destination physical disk.

13. The non-transitory computer readable storage medium of claim 11, wherein the source physical disk is divided into a base image and an overlay image, and wherein only contents of the overlay image are migrated to the destination physical disk during the synchronization phase.

14. The non-transitory computer readable storage medium of claim 8, the processing device to delay processing of a write request to the source disk for a set period of time upon determining the contents stored on the source physical disk and the destination physical disk are equal.

15. A system comprising:

a memory to store instructions; and

a processing device, operatively coupled to the memory, the processing device to execute a virtual machine monitor to execute the instructions to;

identify a steady state wherein a first count of writes associated with a virtual machine stored on a source physical disk and a destination physical disk are equal;

receive a set of write requests comprising data from a guest operating system of the virtual machine;

provide, to the guest operating system, confirmation of a completion of a set of writes to the source physical disk;

following the confirmation of the completion of the set of writes to the source physical disk being provided to the guest operating system, asynchronously replicate the set of write requests to the destination physical disk;

receive a flush request from the guest operating system of the virtual machine, wherein the flush request comprises an instruction to the destination physical disk to write the data pending in cache of the destination physical disk to persistent storage of the destination physical disk;

determine a second count of writes completed on the source physical disk at a time of a previous flush request is less than or equal to the first count of writes identified at the steady state;

execute, by the virtual machine monitor, a flush of the destination physical disk in response to a determination that the second count of writes is less than or equal to the first count of writes;

execute a flush of the destination physical disk in response to a determination that the second count of writes is less than or equal to the first count of writes;

provide, to the guest operating system, confirmation of completion of the flush of the destination physical disk following replication of the set of write requests to the destination physical disk and writing of the data to persistent storage of the destination physical disk;

receive a switch request from the virtual machine manager; and

switch the virtual machine to the destination physical disk and issuing subsequent write requests to the destination physical disk.

16. The system of claim 15, the processing device to report to the virtual machine manager, by the virtual machine monitor, entry into the steady state.

17. The system of claim 15, the processing device to:

maintain a dirty bitmap to record a disk sector written to on the source physical disk but not the destination physical disk; and

empty the dirty bitmap.

18. The system of claim 15, the processing device to operate in a synchronization phase prior to the steady state, wherein the synchronization phase comprises migrating the contents of the source physical disk to the destination physical disk.

19. The system of claim 18, the processing device to enter the synchronization phase in response to at least one of receiving a request from the virtual machine manager or detecting an error in the replication of the set of write requests to the destination physical disk.

20. The system of claim 18, wherein the source disk is divided into a base image and an overlay image, and wherein only contents of the overlay image are migrated to the destination physical disk during the synchronization phase.