Techniques for Virtual Machine Management

• US 20130275808A1
• Filed: 12/19/2011
• Published: 10/17/2013
• Est. Priority Date: 12/21/2010
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: McNeeney, Adam James, Mulcajy, James
• Granted Patent: US 9,690,662 B2
• Time in Patent Office: Days
• US Class Current: 714/20
• CPC Class Codes: G06F 11/1407 Checkpointing the instruction stream
• CPC Class Codes: G06F 11/1438 Restarting or rejuvenating
• CPC Class Codes: G06F 11/1482 by means of middleware or OS functionality
• CPC Class Codes: G06F 11/2023 Failover techniques
• CPC Class Codes: G06F 11/203 using migration
• CPC Class Codes: G06F 11/2094 Redundant storage or storage space G06F11/2056 takes precedence
• CPC Class Codes: G06F 11/2097 maintaining the standby controller/processing unit updated initialisation or re-synchronisation thereof G06F11/1658 and subgroups
• CPC Class Codes: G06F 2009/4557 Distribution of virtual machine instances; Migration and load balancing
• CPC Class Codes: G06F 2201/815 Virtual middleware or OS functionality using virtual machines to implement generic software techniques for error detection or fault masking G06F11/1484
• CPC Class Codes: G06F 2201/82 Solving problems relating to consistency ensuring consistency in mirrored systems G06F11/2064
• CPC Class Codes: G06F 9/45558 Hypervisor-specific management and integration aspects
• CPC Class Codes: H04L 41/0663 Performing the actions predefined by failover planning, e.g. switching to standby network elements
• CPC Class Codes: H04L 41/40 using virtualisation of network functions or resources, e.g. SDN or NFV entities

Abstract

A technique for operating a group of virtual machines (VMs) includes utilizing a checkpoint procedure to maintain secondary VMs to assume tasks of primary VMs within a cluster in the event of failover. On failover of a first one of the primary VMs, a first one of the secondary VMs assumes the tasks from the checkpoint immediately preceding a failover event. Each of the primary VMs is connected to receive data from remaining ones of the primary VMs via an internal bus and process the data on receipt. Checkpoints for the primary VMs are synchronized. For each of the primary VMs, release to the external bus of data generated on the basis of received internal bus data is prevented until a subsequent checkpoint has occurred. On failover of one of the primary VMs, all of the primary VMs are directed to initiate failover to an associated one of the secondary VMs.

Claims

1. A method of operating a group of virtual machines (VMs) that includes two or more primary VMs and two or more secondary VMs with the primary VMs within a cluster being linked to an external bus and with each of the primary VMs being operable to generate data and transmit the data on the external bus and to receive and process data from the external bus, the method comprising:

for each of the primary VMs, utilizing a checkpoint procedure to maintain an associated one of the secondary VMs to assume tasks of the associated one of the primary VMs in the event of failover; and

on failover of one of the primary VMs, causing the associated one of the secondary VMs to assume the tasks of the associated one of the primary VMs from the checkpoint immediately preceding a failover event, wherein each of the primary VMs is connected to receive data from remaining ones of the primary VMs via an internal bus and process the data received on the internal bus on receipt;

synchronizing the checkpoints for the primary VMs;

for each of the primary VMs, preventing release to the external bus of data generated on the basis of received internal bus data until a subsequent checkpoint has occurred; and

on failover of one of the primary VMs, causing all of the primary VMs to initiate failover to an associated one of the secondary VMs.

2. The method of claim 1, wherein for at least one of the primary VMs, on the occurrence of failover within the cluster, determining whether the at least one of the primary VMs has received internal bus data since the preceding checkpoint and, if not, preventing the at least one of the primary VMs from failing over to its respective secondary VM when other of the primary VMs failover.

3. The method of claim 1, wherein the cluster further includes at least one additional primary VM that does not have an associated secondary VM and operation of the at least one additional primary VM is stopped when the primary VMs failover.

4. The method of claim 3, wherein the additional primary VM is connected to receive data from the primary VMs via the internal bus and process the data on receipt, and wherein release to the external bus of data generated on the basis of received internal bus data for the additional primary VM is prevented until a subsequent checkpoint occurs.

5. The method of claim 1, further comprising:

redesignating the secondary VMs as new primary VMs of a new cluster following failover; and

designating new associated secondary VMs for the new primary VMs.

6. An apparatus, comprising:

one or more data storage devices; and

at least one processor coupled to the one or more data storage devices, wherein the apparatus is configured to operate a group of virtual machines (VMs) that include two or more primary VMs and two or more secondary VMs with the primary VMs within a cluster being linked to an external data bus, and wherein;

each of the primary VMs is operable to generate data and transmit the data on the external bus and to receive and process data from the external bus;

for each of the primary VMs, the apparatus maintains, via a checkpoint procedure, associated ones of the secondary VMs to assume tasks of the primary VMs in the event of failover;

on detection of a failover, the apparatus initiates transfer of the tasks of the primary VMs to associated ones of the secondary VMs, with each of the secondary VMs mirroring the state of one of the primary VMs at the checkpoint immediately preceding a failover event;

the apparatus synchronizes the checkpoints for the primary VMs;

the apparatus implements an internal bus mechanism, with each of the primary VMs being connected to receive data from remaining ones of the primary VMs via the internal bus and process the data on receipt;

the one or more storage devices are utilized to buffer, for the primary VMs, external bus data generated on the basis of received internal bus data and release the data to the external bus when a subsequent checkpoint occurs; and

on failover of one of the primary VMs, all of the primary VMs are directed to initiate failover to their respective secondary VMs.

7. The apparatus of claim 6, wherein the primary and secondary VMs are linked via a hypervisor.

8. The apparatus of claim 6, wherein the primary VMs are hosted by a single processor and linked via a single hypervisor.

9. The apparatus of claim 6, wherein the primary VMs are hosted by two or more processor devices and respective hypervisors are connected for the exchange of internal bus data.

10. The apparatus of claim 6, wherein for at least one of the primary VMs, on the occurrence of failover within the cluster, the apparatus determines whether the at least one of the primary VMs has received internal bus data since the preceding checkpoint and, if not, the at least one of the primary VMs is prevented by the apparatus from failing over to an associated one of the secondary VMs when remaining ones of the primary VMs failover.

11. The apparatus of claim 6, wherein the cluster includes at least one additional primary VM whose operation is stopped when other ones of the primary VMs failover.

12. The apparatus of claim 11, wherein the additional primary VM is connected to receive data from the primary VMs via the internal bus and process the data on receipt, and wherein the additional primary VM utilizes the one or more storage devices to buffer external bus data generated on the basis of received internal bus data and release the buffered external bus data to the external bus when a subsequent checkpoint occurs.

13. The apparatus of claim 6, wherein the apparatus is a data processing system.

14. A computer program product, comprising:

a computer-readable storage device; and

code stored on the computer-readable storage device, wherein the code, when executed by a data processing system, causes the data processing system to;

operate a group of virtual machines (VMs) that includes two or more primary VMs and two or more secondary VMs, wherein each of the primary VMs within a cluster is linked to an external data bus and each of the primary VMs is configured to generate data and transmit the data on the external bus and to receive and process data from the external bus;

maintain, for each of the primary VMs, via a checkpoint procedure an associated one of the secondary VMs to assume tasks of the primary VMs in the event of failover;

on detection of a failover of at least one of the primary VMs, transfer tasks of the at least one of the primary VMs to an associated one of the secondary VMs, where each of the secondary VMs mirrors the state of one of the primary VMs at the checkpoint immediately preceding a failover event;

synchronize the checkpoints for all of the primary VMs;

connect each one of the primary VMs to receive data from remaining ones of the primary VMs via an internal bus and process the data on receipt;

buffer for each of the primary VMs external bus data generated on the basis of received internal bus data and release the external bus data to the external bus when a subsequent checkpoint has occurred; and

on failover of one of the primary VMs, direct all of the primary VMs to initiate failover to an associated one of the secondary VMs.

15. The computer program product of claim 14, wherein the primary and secondary VMs are linked via a respective hypervisor.

16. The computer program product of claim 14, wherein the primary VMs are hosted by a single processor and linked via a single hypervisor.

17. The computer program product of claim 14, wherein the primary VMs are hosted by two or more processors and respective hypervisors are connected for the exchange of internal bus data.

18. The computer program product of claim 14, wherein for at least one of the primary VMs, on the occurrence of failover within the cluster, the code causes the data processing system to determine whether the at least one of the primary VMs received internal bus data since the preceding checkpoint and, if not, the at least one of the primary VMs is prevented from failing over to an associated one of the secondary VMs when other of the primary VMs failover.

19. The computer program product of claim 14, wherein the cluster includes an additional primary VM the operation of which is stopped when other of the primary VMs failover.

20. The computer program product of claim 19, wherein the additional primary VM is connected to receive data from the primary VMs via the internal bus and process the data on receipt, buffer external bus data generated on the basis of received internal bus data for the additional primary VM, and release the external bus data to the external bus when a subsequent checkpoint occurs.