JPS62204360A - Inter-system data base sharing system - Google Patents

Abstract

PURPOSE:To attain the quick degeneration of access by inhibiting the access to be given to the corresponding data resources in case the access holding information on a file device having a fault, etc., exists. CONSTITUTION:A system control file 5 is provided to store the access holding information on the faults of file devices 61-63 provided within the shared data resources or the access holding information that had an access after the file device is separated. The access holding information stored in the file 5 is set to the file control table formed in the data processors 1-4 respectively. Then an access is given to the data resources shared by processors 1-4. Here the access holding information stored in the file managing table is checked and an access is inhibited to the file device having a fault, etc., in case the access holding information on the faulty file device exists. Thus it is possible to attain the quick degeneration of access to the data resources having a fault, etc.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a computer system in which a plurality of data processing devices access shared data resources. The present invention relates to an inter-system database sharing method suitable for degenerating data resources such as separation of data resources.

[Prior art]

In large database systems where multiple data processing units need to access shared data resources,
Concurrent access must be managed to maintain data integrity.

Therefore, when a plurality of data processing apparatuses access the same shared data resource at the same time, exclusive control is performed. Specifically, when multiple jobs attempt to access one data set on a direct access storage device at the same time, when one job is accessing it, access from other zips is prohibited or queued. By doing this, one data set can be used by multiple jobs.

Incidentally, a related computer system of this type is described in Japanese Patent Application Laid-Open No. 57-64850. This may be inconsistent in a multiprocessing, multiprogramming general-purpose computing device that controls access to shared data resources due to failure of a communications link, operating system, subsystem, or lock controller. The present invention provides a device that prohibits access to data resources that are not available.

[Problem that the invention seeks to solve]

However, in the conventional exclusive control method in which multiple data processing devices access a shared data resource, if a physical failure occurs in one file device of the shared data resource, the system continues to access the failed file device. When accessing the site, many failures occur.If there are many failures,
No consideration is given to the fact that the system cannot operate normally due to a rapid increase in the utilization rate of the console device for processing reports to the operator and a rapid increase in the CPU utilization rate due to failure processing.

In addition, for hardware maintenance of the file device, the file device can be disconnected from the system.
If the service is stopped and the separated file device is accessed incorrectly, a failure will occur. No consideration is given to the fact that the occurrence of such a failure also impedes the normal operation of the system.

An object of the present invention is to prevent a sudden increase in the usage rate of console devices by preventing recurrence of the same failure in a system in which a plurality of data processing devices access shared data resources;
It also prevents sudden increases in CPU utilization, ensures normal system operation, and protects file devices or disks that have failed.
- It is an object of the present invention to provide a highly reliable inter-system database sharing method by promptly degenerating access to a file device that has been disconnected for software maintenance or the like and whose service has been stopped.

[Means for solving problems]

The above object is a computer system having a plurality of data processing devices, a communication device that connects the plurality of data processing devices, and a shared data resource consisting of a plurality of file devices through which the plurality of data processing devices share data. , a system control file that stores access pending information related to a failure that has occurred in one or more file devices in a shared data resource, or access pending information that has been accessed after one or more file devices have been disconnected due to hardware maintenance, etc. established,
The access pending information stored in the system control file is set in a file management table created within each data processing device, and from then on, when accessing data resources shared by each data processing device, access in the file management table is set. Check the pending information, and if there is access pending information regarding the file device with the problem, access the file device with the problem! This is achieved by providing an intersystem database sharing method that stops

[Effect] The inter-system database sharing method of the present invention is as follows.

It works as follows.

Information accessed by each data processing device when a failure occurs in a file device in a database shared by multiple data processing devices, or when the file device is disconnected due to a service stop for hardware maintenance, etc. (access pending information) is stored in a system control file that is shared by multiple data processing units.

Next, based on the access pending information stored in the shared system control file, access is maintained to the file management table that manages the access pending information created inside each data processing device.

After setting the lightning information, each data processing device checks the file management table when accessing either of the two file devices thereafter, and prohibits access to the faulty file device.

Additionally, if a failure occurs in one file device while accessing data resources shared by one data processing device,
The access pending information is stored in the shared system control file, and at the same time, the access pending information is set in the file management table within its own data processing device. Furthermore, access suspension information regarding the occurrence of a failure is communicated to other data processing apparatuses via the communication device, and similar access suspension information is held in the file management tables in the other data processing apparatuses. When another data processing device accesses a shared data resource, it checks the file management table in this data processing device, and even if this data processing device is not accessing the failed file device, Access to a file device in which a failure has occurred is prohibited based on access pending information of another data processing device that has access experience.

In addition, when part or all of a data processing device restarts due to a system failure, or restarts after a service stop due to another business, a new data processing device may share data resources. When restarting the data processing device, the access hold information at the time of a failure that previously occurred in this data processing device and other data processing devices is read from the shared system control file, and a file management table is created in the data processing device to be restarted. , set access hold information, and thereafter refer to the file management table to prohibit access to a file device with a problem.

〔Example〕

An embodiment of the present invention will be described below with reference to FIGS.

FIG. 1 is an example of a system configuration diagram in which the inter-system database sharing method of the present invention is implemented. According to this example, the database system is composed of four data processing devices = data processing devices 1, 2, and 5.4.
They are interconnected through communication devices 7°, 8.9.10, and share the database group 6. Shared database group 6
, which stores shared data resources, is composed of file devices 61, 62, and 63. Also,
The data processing devices 1, 2, 5.4 are the service system control 7
i) L15 is shared.

The shared system control file 5 is used when each data processing device accesses a shared database group, and accesses a file device that has failed or has been disconnected from the system due to service suspension due to hardware maintenance. This is a place where records (access pending information) are stored. Each data processing device 1゜2.5
．． 4 is an operating system 11, 2 which is a collection of techniques and procedures for operating a data processing device and is prepared to make the processing device easy to use and utilize effectively.
1, 31. 41 and shared resource database group 6 are accessed 5 subsystems 15, 14, 25, 24, 53,
54, 45, 44, and exclusive control managers 12, 22, which perform exclusive control to enable each subsystem to access the shared resource database group 6 without contradiction.
32.42.

Now, in the configuration of the embodiment described above, the operation of the embodiment will be explained focusing on the data processing device 1. Data processing device 11C
When the user requests access to the shared database group 6 through the subsystem 1S, the subsystem 13 issues an exclusive control request to the exclusive control manager 12 when access to the shared resources of the database group 6 occurs. The exclusive control manager 12 acquires a lock that allows exclusive access to the shared resources of the database group 6. After acquiring the lock, the subsystem 15 of the data processing device 1 can exclusively access the target file device, so the subsystem 15 accesses the °C shared data resource via the operating system 11. When the access is completed, the lock is no longer required, so the subsystem 13 issues an exclusive control release request to the exclusive control manager 12 to release the lock and unlock it. Further, the exclusive control manager 12 transmits the exclusive control information to the exclusive control managers 22, 32.4 included in other data processing devices 2, 5.4 using the communication device 7.
2 to release the lock, waiting for requests from other subsystems, and erroneously assigning two or more jobs to the same shared data resource for multiple requests, resulting in an unprocessable Tedlock. Check to see if there are any.

Next, we will outline the processing when a file device of a database group that stores shared data resources is in a suspended state, that is, when it is inaccessible from a subsystem.
A pending state for a file device occurs when a failure occurs in the file device, or when the file device is disconnected from the system and service is stopped for hardware maintenance or the like for the file device. The exclusive control manager receives access pending information from subsystems accessing pending file devices, stores it in the shared system control file, and also releases access pending information to all subsystems other than the subsystem to which the access pending information was delivered. It is a system for communicating information.

Details of the process when the exclusive control manager 12 accesses the file device 61 in the shared database group 6 that is in a pending state due to a request from the subsystem 13 in the shared data processing device 1, or when the file device 61 becomes in a pending state during access. Explain.

The subsystem 13 of the data processing device 1 is the file device 6
1 is accessed and it is found to be in a pending state, or when the subsystem 13 is accessing the file device 61 and a failure occurs midway and this is detected, the file created within the subsystem 13 The access pending information stored in the shared system control file is set in the file management table that manages the access pending information stored in the shared system control file when the system is started or restarted after a failure. Based on the information, not only access suspension information (access suspension release information is also managed as a history). The sent access suspension information is communicated to the exclusive control manager 12. The manager 12 stores the access pending information in the shared system control file 5, and also uses the communication device 7.8 to store the exclusive control lane manager 22j in the other data processing devices 2 and 5.4.
2,421C process/access reservation information. Subsystems 14 and 23 other than the accessed subsystem 15
,24,33°64.45.44 is subsystem 13
The access pending information regarding the failure that occurred while accessing the file device 61 or the failure that occurred during the access is sent to each subsystem 14.23, 24, 33, 34, 4.
5. Set in each file management table in 44. Since each subsystem accesses the shared database group 6 by referring to its respective file management table, the file device 61 in which the failure has occurred will not be accessed thereafter. Additionally, when a file device is disconnected from the system due to hardware maintenance, etc. for the file device, and the service is stopped, each subsystem sets access pending information in the file management table and returns the file device to the disconnected file device. Prohibit access to. Furthermore, when releasing the hold, the access hold release information is set in the file management table of each subsystem according to the same procedure, and the file management table is referred to when accessing the database group 6 shared by each subsystem. Therefore, if access suspension is released, access will be performed.

Next, if part or all of the data processing equipment is restarted due to a system failure, or if the service is restarted after it has been stopped due to separate operations; in addition, a new data processing equipment will share the data resources. When starting as shown in Figure 2, the processing of access pending information when the file device is disconnected due to a previous failure or service stoppage due to /S-Docware maintenance is as follows. The subsystem 1S of the data processing device 1 that performs d start or restart, etc., which will be explained using a flowchart, is as follows:
Inquire about access pending information to the exclusive control manager 12,
The exclusive control manager 12 reads access suspension information and access suspension release information regarding the database group 6 stored in the shared system control file 5 (
101), the subsystem 16 sorts the access pending information, etc. in the subsystem 13, and creates a table (file management table) that serves as a history of the access pending information, etc. (102). Next, when the shared database group 6 is not accessed, the process ends as is (10B), and the step (105) for starting the access determines whether access suspension information and access suspension release information exist in the file management table. (104). If the subsystem 15 searches the file management table (104) and there is no access pending information, the subsystem 15 accesses the target file device (107).

Results of searching the file management table (104)'
.

If access pending information, etc. exists, it is further checked whether access pending information, etc. regarding the target file device exists (105).

If there is no access pending information regarding the target file device, the target file device is accessed (107). If access suspension information regarding the target file device exists in the file management table and the suspension has not been released yet, access to the target file device is prohibited (106) and the process ends (I
C! 8).

According to this embodiment, access suspension information and access suspension release information are stored in the system shared file, and before each subsystem accesses the shared database 2, the access suspension information stored in the system shared file is stored. A file management table for managing access pending information, etc. is created in the subsystem based on the information, etc., the access pending information, etc. is sected, and the shared database resource is accessed by referring to this file management table. Duplicate access can be prohibited in response to a failure, etc., and a sudden increase in the usage rate of the console device and the sudden increase in the CPU usage rate can be prevented, ensuring normal operation of the system, and maintenance of the file device or hardware where the failure has occurred. By promptly degrading access to file devices that have been disconnected and services have been stopped due to
This has the effect of improving system reliability.

In addition, if part or all of the data processing equipment has a system failure and restarts, or if the service is restarted after it has been stopped due to another business, in addition, a new data processing equipment may be required to use data resources. When the system is started or restarted after shared use, the previous access hold information and access hold release information are called from the system shared file to the subsystem, a file management table is formed, and access hold information, etc. Since the file management table is referenced when accessing shared database resources, access is prohibited due to the same failure, etc.
Since the number of operations required when files are on hold is reduced, operability and reliability are improved.

〔Effect of the invention〕

According to the present invention, access reservation information regarding the reservation state of shared data resources is accumulated in a file shared by the system, and this access reservation information is held in each data processing device, so that any data processing device can When accessing a shared data resource, the access hold information held in the data processing device is referred to, so duplicate access for the same failure can be prohibited, and access to data resources with failures can be degraded or quickly implemented. This has the effect of improving system operability and reliability.

[Brief explanation of drawings]

FIG. 1 is a configuration diagram of a database system according to an embodiment of the present invention, and FIG. 2 is a flowchart of processing when starting or restarting the system of a data processing device. Explanation of symbols 1.2, 5.4... Data processing device, 5... Shared system control file, 6... Shared database group, 61.62.63... File device, 7.8, 9 .1
0...Communication device, 11.21.31.41...Operating system, 12.22,52.42...Exclusive control manager,
15. 14, 23, 24, 53. 54, 45. 44 River subsystem.

Claims (2)

[Claims] (1) A computer comprising a plurality of data processing devices, a communication device that interconnects the plurality of data processing devices, and a shared data resource consisting of a plurality of file devices that share data with the plurality of data processing devices. In the system, a shared system control file is provided for storing access suspension information indicating that access by the data processing device to one or more file devices in the shared data resource is suspended, and the shared system control file is stored in the shared system control file. The stored access reservation information is set in a file management table that manages access reservation information formed inside each of the data processing devices, and when accessing a data resource shared by each of the data processing devices, An inter-system database sharing method characterized in that access reservation information in a management table is referenced, and when access reservation information exists, access to the corresponding data resource is prohibited. (2) In the inter-system database sharing method described in claim 1, access pending information regarding a failure that occurs while one data processing device is accessing the shared data resource is stored in the file management of the own data processing device. inter-system database sharing, characterized in that it is set in a table, communicated with all other data processing devices via a communication device, and set in each file management table in all of the other data processing devices; method.