System, method and program to preserve a cache of a virtual machine

• US 8,060,683 B2
• Filed: 12/17/2004
• Issued: 11/15/2011
• Est. Priority Date: 12/17/2004
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Tkatschow, Xenia, Shultz, Steven
• Primary Examiner(s): Bataille; Pierre-Michel
• Assistant Examiner(s): Verderamo, III; Ralph A
• Application Number:
  US11/015,167
• Publication Number:
  US 20060136667A1
• Time in Patent Office: 2,524 Days
• Field of Search: 711/6, 711/118, 718/1, 718/104
• US Class Current: 711/6
• CPC Class Codes: G06F 12/0842 for multiprocessing or multitasking
• CPC Class Codes: G06F 12/0866 for peripheral storage systems, e.g. disk cache
• CPC Class Codes: G06F 9/5077 Logical partitioning of resources; Management or configuration of virtualized resources specific details on emulation or internal functioning of virtual machines G06F9/455

Abstract

A system, computer program product and method for managing a cache of a virtual machine. A cache is defined in memory, and a virtual machine is assigned to the cache. An identity of the cache is recorded in storage. The virtual machine terminates, and the cache and contents of the cache are preserved despite the termination of the virtual machine, such that if the virtual machine subsequently resumes operating, the virtual machine can access the cache and its contents. There is also a system, method and computer program product for managing a cache of an LPAR. A cache is defined in memory, and assigned to an LPAR. A record is made of an identity of the cache in storage. The LPAR terminates, and the cache and contents of the cache are preserved despite the termination of the LPAR, such that if the LPAR subsequently resumes operating, the LPAR can access the cache and its contents.

Claims

1. A method for managing virtual machines, said method comprising the steps of:

a first virtual machine being operational and requesting to attach to a cache, and in response, attaching said first virtual machine to said cache, and recording that said first virtual machine is assigned to said cache;

a second virtual machine being operational and requesting to attach to said cache, and in response, attaching said second virtual machine to said cache, and recording that said second virtual machine is assigned to said cache;

subsequently, said first virtual machine ceasing to operate due to failure of said first virtual machine while said second virtual machine is operational, and in response, determining that a virtual machine is operational and attached to said cache, and in response, preserving said cache; and

subsequently, resuming operation of said first virtual machine, and in response, determining from said record that said first virtual machine is assigned to said cache and reattaching said first virtual machine to said cache.

2. A method as set forth in claim 1 further comprising the prior steps of a hypervisor defining resources for said first and second virtual machines from a same real computer;

and wherein the step of attaching said first virtual machine to said cache is performed by said hypervisor, the step of attaching said second virtual machine to said cache is performed by said hypervisor, and the step of reattaching said first virtual machine to said cache is performed by said hypervisor.

3. A method as set forth in claim 1 further comprising the steps of:

said second virtual machine making a request to a hypervisor of said first and second virtual machines to preserve said cache even if there is no operational virtual machine which is attached to said cache, and in response, said hypervisor making a record that said cache should be preserved even if there is no operational virtual machine which is attached to said cache, and subsequently, said first and second virtual machines terminating and there are no operational virtual machines attached to said cache, and in response, preserving said cache based on said record that said cache should be preserved even if there is no operational virtual machine attached to said cache, and subsequently, said second virtual machine resuming operation, and in response, reattaching said second virtual machine to said cache.

4. A method as set forth in claim 1 wherein:

a hypervisor defined said second virtual machine to preserve said cache in the event said first virtual machine ceases to operate; and

a program within said second virtual machine repeatedly queries said hypervisor for identities of caches used by said first virtual machine and other virtual machines in said real computer, and attaches to said caches used by said first virtual machine and other virtual machines in said real computer for a sole purpose of preserving said caches used by said first virtual machine and other virtual machines in said real computer in the event said first virtual machine and other virtual machines in said real computer terminate.

5. A method as set forth in claim 1 wherein said first virtual machine terminated without normal shut down of said first virtual machine.

6. A computer system for managing a cache, said system comprising:

a processor;

a hypervisor executing on said processor to define first and second virtual machines, said first and second virtual machines sharing said processor;

means within said first virtual machine for requesting attachment to a cache, and means within said hypervisor, responsive to said request by said first virtual machine, for attaching said first virtual machine to said cache and recording that said first virtual machine is assigned to said cache;

means within said second virtual machine for requesting attachment to said cache, and means within said hypervisor, responsive to said request by said second virtual machine, for attaching said second virtual machine to said cache and recording that said second virtual machine is assigned to said cache;

means within said hypervisor, responsive to subsequent cessation of operation of said first virtual machine while said second virtual machine is operational, for determining that at least one virtual machine is operational and attached to said cache, and in response, preserving said cache; and

means within said hypervisor, responsive to subsequent resumption of operation of said first virtual machine, for determining from said record that said first virtual machine is assigned to said cache and reattaching said first virtual machine to said cache.

7. A system as set forth in claim 6 further comprising:

means within said second virtual machine for making a request to said hypervisor to preserve said cache even if there is no operational virtual machine which is attached to said cache;

means within said hypervisor, responsive to the request to preserve said cache, for making a record that said cache should be preserved even if there is no operational virtual machine which is attached to said cache;

means, within said hypervisor, responsive to subsequent termination of said first and second virtual machines and any other virtual machines attached to said cache, for preserving said cache; and

means, within said hypervisor, responsive to subsequent resumption of said second virtual machine, for reattaching said second virtual machine to said cache.

8. A system as set forth in claim 6 wherein said hypervisor defined said second virtual machine to preserve said cache in the event said first virtual machine ceases to operate, and further comprising a program within said second virtual machine to repeatedly query said hypervisor for identification of caches used by said first virtual machine and other virtual machines in said real computer, and attach said second virtual machine to said caches used by said first virtual machine and other virtual machines in said real computer for a sole purpose of preserving said caches used by said first virtual machine and other virtual machines in said real computer in the event said first virtual machine and other virtual machines in said real computer terminate.

9. A system as set forth in claim 6 wherein said first virtual machine terminated without normal shut down of said first virtual machine.

10. A method for managing virtual machines, said method comprising the steps of:

a first virtual machine being operational and making a request to a hypervisor to attach to a first cache and preserve said first cache in case said first virtual machine terminates, and in response, said hypervisor defining a first table which correlates first virtual addresses to said first cache and recording that said first table should be preserved upon termination of said first virtual machine, said first virtual machine using said first virtual addresses to access said first cache, and subsequently, said first virtual machine terminating due to failure of said first virtual machine and said first cache being preserved, and subsequently, said first virtual machine resuming operation, and in response, reattaching said first virtual machine to said first cache; and

a second virtual machine being operational and making a request to said hypervisor to attach to a second cache but not preserve said second cache in case said second virtual machine terminates, and in response, said hypervisor defining a second table which correlates second virtual addresses to said second cache and recording that said second table should not be preserved upon termination of said second virtual machine, said second virtual machine using said second virtual addresses to access said second cache, and subsequently, said second virtual machine terminating, and in response, said second table not being preserved, and subsequently, said second virtual machine resuming operation and attaching to a third cache, said second cache not being available to said second virtual machine upon the resumption of operation of said second virtual machine; and

wherein

said hypervisor defines computer resources for said first and second virtual machines in a same real computer.