System for recovery from a virtual machine monitor failure with a continuous guest dispatched to a nonguest mode

• US 5,437,033 A
• Filed: 11/04/1991
• Issued: 07/25/1995
• Est. Priority Date: 11/16/1990
• Status: Expired due to Fees

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Yamamoto, Tadashi, Inoue, Taro, Tanaka, Shunji, Umeno, Hidenori, Ohtsuki, Toru
• Primary Examiner(s): Black, Thomas G.
• Assistant Examiner(s): AMSBURY, WAYNE P
• Application Number:
  US07/787,074
• Time in Patent Office: 1,359 Days
• Field of Search: 395/275, 395/400, 395/575, 395/700
• US Class Current: 714/10
• CPC Class Codes: G06F 11/1438 Restarting or rejuvenating
• CPC Class Codes: G06F 11/1484 involving virtual machines
• CPC Class Codes: G06F 2201/815 Virtual
• CPC Class Codes: G06F 9/45533 Hypervisors; Virtual machine monitors

Abstract

A system and method for continuous operation of a virtual machine system having operation modes including a guest mode in which virtual machines are operated and a nonguest mode in which a virtual machine monitor for controlling the virtual machines is operated. The continuous guest is a virtual machine which does not stop executing operation at the occurrence of a failure due to program error of the virtual machine monitor. A main storage is provided with two areas. One of the two areas is a continuous guest area having the same host absolute address in the nonguest mode as a guest absolute address in the guest mode, the area is used by the continuous guest which is a virtual machine which continues to operate on transition of the operation mode from the guest mode to the nonguest mode. The other is an area in which a program module for dispatching the continuous guest in response to the transition of the operation mode from the guest mode to the nonguest mode. The continuous guest is allocated to the guest area on transition of the operation mode from the guest mode to the nonguest mode. Occurrence of a failure due to a program error of the virtual machine monitor is detected and by starting the program module for dispatching the continuous guest in response to the detection of the occurrence of the failure, the operation of the continuous guest does not stop executing.

Claims

1. A system for recovery from a virtual machine monitor failure of a virtual machine system having operation modes including a guest mode in which virtual machines are operated and a nonguest mode in which a virtual machine monitor for controlling said virtual machines is operated, the system comprising:

a main storage which is connected with said virtual machine system including,

a continuous guest area having the same host absolute address as a guest absolute address, and being occupied by a continuous guest virtual machine which continues to operate even if a failure due to a program error of said virtual machine monitor has occurred, and

an area in which is stored a program module for dispatching said continuous guest virtual machine in the nonguest mode in response to detection of said failure;

means for allocating said continuous guest virtual machine to said guest area;

means for detecting the occurrence of said failure due to said program error of said virtual machine monitor; and

means for starting said program module for dispatching said continuous guest virtual machine in response to the detection of the occurrence of said failure by said detecting means.

2. The system for recovery from a virtual machine monitor failure of a virtual machine system according to claim 1, wherein said virtual machine monitor includes means for dedicating to said continuous guest virtual machine an instruction processor of said virtual machine system corresponding to the instruction processor which is used by said continuous guest virtual machine.

3. The system for recovery from a virtual machine monitor failure of a virtual machine system according to claim 1, wherein said virtual machine monitor includes means for dedicating to said continuous guest virtual machine an input/output interruption subclass of said virtual machine system corresponding to the input/output interruption subclass used by said continuous guest virtual machine.

4. The system for recovery from a virtual machine monitor failure of a virtual machine system according to claim 1, wherein said main storage further includes a hardware system area (HSA) and said virtual machine system includes means for saving in said hardware system area (HSA), information which is necessary for continuous operation of said continuous guest virtual machine in response to occurrence of the interception/host interruption.

5. A system for recovery from a hypervisor failure of a logical partition system having operation modes including a guest mode in which logical partitions are operated and a nonguest mode in which a hypervisor for controlling said logical partitions is operated, the system comprising:

a main storage which is connected with said logical partition system including,

a continuous guest area having the same host absolute address as a guest absolute address, and being occupied by a continuous guest which is a logical partition which continues to operate even if a failure due to a program error of said hypervisor has occurred, and

an area in which is stored a program module for dispatching said continuous guest in the nonguest mode in response to detection of said failure;

means for allocating said continuous guest to said guest area;

means for detecting the occurrence of said failure due to said program error of said hypervisor; and

means for starting said program module for dispatching said continuous guest in response to the detection of the occurrence of said failure by said detecting means.

6. A method of recovery from a virtual machine monitor failure of a virtual machine system, having operation modes including a guest mode in which virtual machines are operated and a nonguest mode in which a virtual machine monitor for controlling said virtual machines is operated, connected with a main storage including,

a continuous guest area having the same host absolute address as a guest absolute address, and being occupied by a continuous guest virtual machine which continues to operate even if a failure due to a program error of said virtual machine monitor has occurred, and

an area in which is stored a program module for dispatching said continuous guest virtual machine in the nonguest mode in response to detection of said failure, the method comprising the steps of:

allocating said continuous guest virtual machine to said guest area;

detecting the occurrence of said failure due to said program error of said virtual machine monitor; and

starting said program module for dispatching said continuous guest virtual machine in response to the detection of the occurrence of said failure by said detecting means.

7. The method of recovery from a virtual machine monitor failure of a virtual machine system according to claim 6, wherein said virtual machine monitor dedicates to said continuous guest virtual machine an instruction processor of said virtual machine system corresponding to an instruction processor which is used by said continuous guest virtual machine.

8. The method of recovery from a virtual machine monitor failure of virtual machine system according to claim 6, wherein said virtual machine monitor dedicates to said continuous guest virtual machine an input/output interruption subclass of said virtual machine system corresponding to an input/output interruption subclass used by said continuous guest virtual machine.

9. The method of recovery from a virtual machine monitor failure of a virtual machine system according to claim 6, wherein said main storage further includes a hardware system area (HSA) and said virtual machine system saves in said hardware system area (HSA) information which is necessary for continuous operation of said continuous guest virtual machine in response to occurrence of an interception/host interruption.

10. A method of recovery from a hypervisor failure of a logical partition system connected to a main storage said logical partition system, having operation modes including a guest mode in which logical partitions are operated and a nonguest mode in which a hypervisor for controlling said logical partitions is operated including,

a continuous guest area having the same host absolute address as a guest absolute address, and being occupied by a continuous guest which is a logical partition which continues to operate even if a failure due to a program error of said hypervisor has occurred, and an area in which is stored a program module for dispatching said continuous guest in the nonguest mode in response to detection of said failure, the method comprising the steps of:

allocating said continuous guest to said guest area;

detecting the occurrence of said failure due to said program error of said hypervisor; and

starting said program module for dispatching said continuous guest in response to the detection of the occurrence of said failure by said detecting means.

11. A method for recovery from a failure of a virtual machine system, said system including a real machine having operation modes including a guest mode in which virtual machines implemented by control of a control program are operated and a nonguest mode in which said control program is operated, and a main memory connected to said real machine, the method comprising the steps:

allocating a continuous guest virtual machine to an area of said main memory having the same absolute address as a 1.0 continuous guest virtual machine which executes a program to be continued for execution even if a failure has occurred in said control program;

operating said virtual machine including said continuous guest virtual machine under control of said control program;

detecting the occurrence of said failure due to a program error of said control program;

stopping operation of said continuous guest virtual machine and said control program in response to the detection of the occurrence of said failure by said detecting step; and

re-starting processings having been performed on said continuous guest virtual machine in the nonguest mode.

12. The method according to claim 11, wherein said operation step includes a step of storing information relating to said continuous guest virtual machine to a predetermined area of said main memory.

13. The method according to claim 12, wherein said predetermined area is formed of an area of a hardware system which is accessed by a given instruction.

14. The method according to claim 13, wherein said storing step is performed in response to occurrence of an interception/host interruption.

15. The method according to claim 12, wherein said restarting step includes reading information relating to said continuous guest virtual machine and stored in said predetermined area, and a step of loading the information read by said reading step to said real machine.

16. The method according to claim 13, wherein said control program dedicates to said continuous guest virtual machine an instruction processor of said virtual machine system corresponding to an instruction processor which is used by said continuous guest virtual machine.

17. The method according to claim 13, wherein said control program dedicates to said continuous guest virtual machine an input/output interruption subclass of said virtual machine system corresponding to an input/output interruption subclass used by said continuous guest virtual machine.