System to pass through resource information

• US 5,446,875 A
• Filed: 01/14/1994
• Issued: 08/29/1995
• Est. Priority Date: 08/17/1989
• Status: Expired due to Term

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Shimizu, Kunio, Ogisu, Yukihisa
• Primary Examiner(s): Baker, Stephen M.
• Application Number:
  US08/181,368
• Time in Patent Office: 592 Days
• Field of Search: 395/575, 364/285.2
• US Class Current: 714/1
• CPC Class Codes: G06F 11/1482 by means of middleware or OS functionality
• CPC Class Codes: G06F 11/1492 by run-time replication performed by the application software
• CPC Class Codes: G06F 11/2097 maintaining the standby controller/processing unit updated initialisation or re-synchronisation thereof G06F11/1658 and subgroups

Abstract

A system for passing through resource information in a computer system including an actual process, a back-up process, and a shared memory connected to the actual process and the back-up process. The actual process includes information which is stored in the shared memory and passed through to the back-up process by passing the contents of the shared memory to the back-up process. The back-up process performs an operation as a new actual process after being passed the contents of the shared memory when the actual process is not operable. The shared memory includes two sets of areas which respectively include an original information area storing information which is transmitted from the actual process to the back-up process, and a difference area storing information related to a change when the actual process changes the original information area. The back-up process reads out information stored in a difference area of one of the two sets of areas and passes through the information stored therein related to the change. The actual process stores the changed information that is transmitted to the back-up process in the original information area of the back-up process and stores the information related to the change in the difference area of the back-up process. The original information stored in the original information area of an other of the two sets of areas is copied and the two sets of areas are switched when at least one of a preset time interval is reached and a preset volume stored in the difference area is reached. In an alternative embodiment, the information passed from the actual process to the back-up process is divided into two groups according to the frequency of change of the resource information. The resource information in a first group is passed through in a batch to the back-up process when the actual process crashes. The resource information in a second group is passed through to the back-up process intermittently after preset time intervals or after volumes of changed information reach a preset threshold.

Claims

1. A system to pass through resource information in a computer system having an actual process, a back-up process, and a shared memory to which said actual and back-up processes are connected, said actual process having information to be stored in said shared memory and to be passed through to said back-up process, said back-up process being passed the contents of said shared memory and performing an operation as a new actual process after being passed said contents of said shared memory when said actual process is not operable, said system comprising:

means for providing two sets of areas in said shared memory, each being composed of an original information area that stores information that is transmitted from said actual process to said back-up process and a difference area that memorizes the information related to a change when said actual process changes said original information area;

said back-up process reading out information related to said resource information stored in the difference area of one of said two sets of areas and passing through the information related to said change;

and said actual process storing the changed information that is to be transmitted to said back-up process in an original information area of said back-up process and the information related to said change in a difference area of said back-up process;

means for copying the original information memorized in the original information area of said other set by at least one of the criteria of a preset time interval being reached and a preset volume stored in said difference area being reached; and

means for switching two sets of areas each having said actual process and said back-up process so that said back-up process updates data.

2. The system to pass through resource information according to claim 1, wherein

said actual process notifies said back-up process of said switching when a used memory of the two sets is switched,

and said back-up process passes through information related to said change by reading out said resource information stored in the difference area of the area of the other set switched.

3. The system to pass through resource information according to claim 2, wherein

the information stored in said difference area related to said change refers relates to changed data in said original information area.

4. The system to pass through resource information according to claim 3, wherein

said area of said actual process switches at a crash time, said back-up process carries over the final resource information from said set formerly used by said actual process, and becomes the actual process.

5. A system to pass-through resource information in a computer system having an actual process, a back-up process, and a non-volatile shared memory to which said actual and back-up processes are connected, said actual process having information to be stored in said shared memory and to be passed through to said back-up process, said back-up process being passed the contents of said shared memory and performing an operation as a new actual process after being passed said contents of said non-volatile shared memory when said actual process is not operable, said system comprising:

means for enabling an area of the non-volatile shared memory to which the resource information is set, to comprise a dual structure with a current version and a new version;

means for setting original information areas storing the information that is transmitted from the actual process to the back-up process and difference areas storing information related to a change in the original information area in a current version and a new version, respectively;

means for the actual process writing into the current version;

means for the back-up process reading from the new version;

means for enabling the actual process to copy the content of the original information area of the current version to the original information area of the new version and to switch a current version to a new version and a new version to a current version after elapse of a preset time period or after a preset volume threshold stored in said difference area is reached; and

means for enabling the actual process to notify the back-up process of the switching between the current version and the new version.

6. A method to pass through resource information in a computer system having an actual process, a back-up process, and a shared memory to which said actual and back-up processes are connected, said actual process having information to be stored in said shared memory and to be passed through to said back-up process, said back-up process being passed the contents of said shared memory and performing an operation as a new actual process after being passed said contents of said shared memory when said actual process is not operable, said method comprising the steps of:

classifying said information to be passed-through into first information changing relatively frequently during a normal operation and second information changing relatively frequently during a normal operation; and

collectively passing said first information to said back-up process at the crash time of said actual process, and periodically passing said second information to said back-up process at least at predetermined time intervals or when the amount of differences in said second information has reached a predetermined value.

7. The system to pass through resource information according to claim 6, wherein each of said first information and second information include an amount of information which can be read in less time than a time interval defined by the predetermined time interval.

8. The system to pass through resource information according to claim 7, wherein

said resource information has a flag showing the use of a resource,

and said batch pass-through affects resource information when said flag is ON.

9. The system to pass through resource information according to claim 7, wherein

said intermittent pass-through affects resource information with the flag turned ON to show usage in an intermittent pass-through entry.

10. The system to pass through resource information according to claim 9, wherein

at the completion of said intermittent pass-through, the flag showing the usage set in said pass-through entry is turned OFF.

11. A System to pass through resource information in a computer system, having an actual process and a back-up process, wherein the back-up process continues handling tasks instead of the actual process when the actual process is not operable to pass through the information from the actual process to the back-up process, said system comprising:

means for dividing the resource information passed through from the actual process to the back-up process into two groups by the change frequencies of resource information;

means for passing the resource information in a first group through in a batch to the back-up process at a crash time of the actual process; and

means for passing the resource information in a second group through intermittently to the back-up process after preset time intervals or after the volumes of changed information reach a preset threshold.

12. A method for passing-through resource information used in a computer system with an actual process and a back-up process that passes through resource information necessary to continue handling tasks stored in a non-volatile shared memory connected to said actual process and said back-up process, from said actual process to said back-up process at least when said actual process is not operable, wherein two sets of areas are provided in said non-volatile share memory, each being composed of an original information area for storing information transmitted from said actual process to said back-up process and a difference area for storing the information related to a change when said actual process changes said original information area, said method comprising the steps of:

reading out information related to said resource information stored in the difference area of one of said two sets, and passing through the information related to said change, said reading out being performed by said back-up process;

storing by said actual process the changed information to be transmitted to said back-up process in the original information area of said back-up process and the information related to said change in the difference area of said back-up process;

copying the original information stored in the original information area of said other set by at least one of the criteria of a preset time interval being reached or a preset volume stored in said difference area being reached; and

switching two sets of areas between said actual process and said back-up process.

13. A computer system comprising:

a first processor module to perform an actual process;

a second processor module to perform a back-up process corresponding to the actual process, based on a failure of the actual process; and

a non-volatile shared memory coupled to the first and second processor modules, to store first and second information associated with the actual process, the first and second information being written into the non-volatile shared memory by the first processor module, the first information changing relatively frequently and the second information changing relatively infrequently, the second processor module reading at least some of the second information from the non-volatile shared memory during a normal operation of the actual process, and the second processor module reading the first information and any portion of the second information not previously read by the second processor module, based on the failure of the actual process.