JPH036633A - File access system among plural host computers - Google Patents

Abstract

PURPOSE:To shorten the time required for an access given to a file of a shared storage device from a host computer except a control host computer by making an access to the file of the shared storage memory with no intervention of the control host computer. CONSTITUTION:When the host computer 4 performs the reading/writing processes to a file, a copy of a file control block of the file to be processed is produced on a shared storage 6. Consequently, the computer 4 can perform the process by referring to and updating the copy of the file control block of the storage 6 with no intervention of a host computer 3 serving as a control host computer. Thus it is possible to shorten the time required an access given to the file of the storage device 6 from a computer 4 except the computer 3.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a file access method, and particularly to a file access method between multiple host computers.

[Conventional technology]

Conventionally, files on an auxiliary storage device (hereinafter referred to as a shared storage device) that can be accessed from multiple host computers are stored in one
Centralized management (data management) with one host computer
In a file access method between multiple host computers in a computer system that uses data management, when a file on a co-located storage device is accessed from a host computer that is not managing data, That access was done through a host combinator (called a host combinator).

[Problem to be solved by the invention]

In the conventional file access method between multiple host computers described above, when a file on a shared storage device is accessed from a host computer that does not manage data (a host computer other than the management host computer), the file is accessed via the management host computer. Since the access is being performed by a host computer other than the management host computer, the access time for files on the shared storage device from host computers other than the management host computer becomes longer, the load is concentrated on the management host computer, and the processing efficiency of the computer system becomes worse. In view of the above-mentioned points, it is an object of the present invention to shorten the access time for files on a shared storage device from host computers other than the management host computer, and reduce the load on one host computer (management host computer). An object of the present invention is to provide a file access method between multiple host computers that can avoid concentration of files.

[Means to solve the problem]

The file access method between multiple host computers of the present invention is a computer system in which one management host computer centrally manages files on a shared memory mW that can be accessed from multiple host computers. A unique memory device that stores file control blocks created for control, etc., and is accessible from each host computer, and when files on a shared storage device are accessed from a host computer other than the management host computer. File control of the file based on the file control block of the file on the specific memory device 1 A shared memory device that stores a copy of the wholesale block, and a file on the shared storage device that cannot be accessed from a host computer other than the management host computer. When the file is opened, a file control block for the file is created on the specific memory device and a copy of the file tldJ jB profile of the file that can be directly referenced/updated from the host computer other than the management host computer is created. and a centralized data management mechanism in a management host computer created on the shared memory device based on file control blocks.

[Effect]

In the file access method between multiple host combiners of the present invention, a unique memory device stores a file control block created for file control by a management host computer, and can be accessed from each host computer. When a shared memory device accesses a file on a coexisting storage device from a host computer other than the management host combiator, what is the file system for the file on the private memory device? The central data management mechanism in the management host computer stores a copy of the file control file of the file based on the 'n block, and the central data management mechanism in the management host computer performs the following operations when the file on the co-located storage device is accessed from a host computer other than the management host computer. When a file is opened, a file control block for the file is created on the specific memory device and can be referenced directly from the host computer other than the management host computer.
A copy of the file control block of the file that can be updated is created on the shared memory device based on the file control block of the file.

〔Example〕

Next, please refer to the drawings regarding the present invention! ! This will be explained in ffL.

FIG. 1 is a block diagram showing the configuration of an embodiment of a file access method between multiple host computers according to the present invention. The file access method between multiple postcomputers of this embodiment centralizes data management of the user program 1 running on the host computer 3, the user program 2 running on the host computer 4, and the files on the shared memory 17. A host computer 3 (management host computer) that has a centralized data management mechanism 30 for data management, a host computer 4 that does not perform data management, and a file control block created for file control etc. by the host computer 3 are recorded. t9 specific memory device 5 (host computer 3
exists in conjunction with. A memory device 6 (which can be implemented on both the main storage device and the auxiliary storage device) that can be accessed from both the host computer 3 and the host computer 4 ) and a shared storage device 7 which is an auxiliary storage device that can be accessed from both the host computer 3 and the host computer 4 and has a file control information area 70 (an area for storing control information of existing files). It is made up of.

FIG. 2 is a diagram showing the configuration of a file control block created in the private memory device 5 (naturally, the configuration of a copy of the file control block created in the shared memory device 6 is also the same). A file control block with control information includes information such as a copy existence flag (a flag indicating whether or not a copy of the file control block exists), used block status, block size, and processing being performed on the file. has been done.

Next, the operation of the file access method between the plurality of host computers of this embodiment configured as described above will be explained.

First, the operation of the file access method between multiple host computers of this embodiment when the user program 1 is started on the host computer 3 will be described.

The user program 1 started on the host computer 3 is executed by the centralized data management mechanism 3 in the host combiator 3.
0, an open request (file open request) is made to file Fl (file used by user program 1).

At this time, the user program I estimates the number of used blocks for the file F1 in advance and determines the number of used blocks I as the number of blocks to be allocated to the shared storage device 7.

Upon receiving the open request, the centralized data management mechanism 30 allocates the area of the file Fl in the coexisting storage device 7 for the specified number of blocks, and creates a file control block of the file F1 (a file control block containing information as shown in FIG. 2). ) in the shared storage device 7 @i area 70
is created on the specific memory device 5 based on the information therein.

When creating this file control block, the centralized data management mechanism 30 recognizes that the file F1 was opened by the user program l,
Based on this recognition, the copy existence flag in the file control block is set to 0FF (information indicating that there is no copy of the file control block).

When the above-described open processing (processing based on an open request) is completed, the centralized data management mechanism 30 notifies the user program 1 that the open processing of the file F1 has been completed, and returns control to the user program 1.

User program 1 creates data in the process of processing,
Write to file F1 based on the data creation
The request (data write request) is sent to the central data management mechanism 30.
Issued to.

The centralized data management mechanism 30 that received the write request performs wriLe processing (data writing processing) to the file F1 on the coexisting storage device 7, and reflects the Wr and ILe processing to the file F1 on the specific memory device 5. file F1
Updates information such as the usage block status in the file control block.

When the write process described above ends, the centralized data management mechanism 30 notifies the user program 1 that the write process has ended, and returns control to the user program 1.

User program 1 saves r of file Fl during processing.
An ead request (data read request) is issued to the centralized data management mechanism 30.

Upon receiving the read request, the centralized data management mechanism 30 refers to the file control block of the file F1 existing on the private memory device 5 and reads the file F1 on the co-located storage device 7.
1 read processing (data reading processing) is performed.

When the above read process is completed, the centralized data management mechanism 30 notifies the user program 1 that the read process has ended, and returns control to the user program 1.

Here, the user program 1 updates the data that has been read and performs write operations on the data again.
Assume that a request is issued to the centralized data management facility 30.

In the write process performed when the user program 1 issues a write request again, if the data updated by the write process exceeds the size (block size) of the file F1, the central data management mechanism 30
is the a11ocate process as shown below - (region 7
0 Cate processing).

That is, the necessary amount of area is allocated on the shared storage device 7.

Furthermore, the file F existing on the specific memory device 5
A copy existence flag (in this case, OFF is set) in the file control block of file F1 is referred to, and a copy of the file control block of file F1 exists on the shared memory device 6 based on the OFF of the copy existence flag. The information in the file control block of the file F1 existing on the specific memory device 5 is transferred to the above al.
Update to reflect the locate process.

When the write process including the above-mentioned allocate process is completed, the central data management R side 30 notifies the user program l that the write process has ended, and returns control to the user program l.

When all the processes for file F1 in user program 1 are completed, user program l
A close request (file close request) is issued to the centralized data management mechanism 30.

Upon receiving the close request, the centralized data management mechanism 30 closes the unused block f in the file Fl on the shared storage device 7.
Performs ree processing (release processing), refers to the copy existence flag in the file control block of file F1 existing on the specific memory device 5, and turns the copy existence flag OFF.
The information in the file control block of the file F1 existing on the specific memory device 5 based on the above f re
Update to reflect e-processing.

Further, the centralized data management mechanism 30 updates the file control information area 70 on the shared storage device 7 based on the update, and then deletes the file control block of the file F1 existing on the private memory device 5. .

When the above-described close processing is completed, the centralized data management mechanism 30 notifies the user program 1 that the close processing has ended, and returns control to the user program 1.

Next, run the user program 2 on the host computer 4.
The operation of the file access method among a plurality of host computers according to this embodiment when the host computer is activated will be described.

The user program 2 started on the host computer 4 requests the host computer 3 via the host computer 4 to open the file F2 (the file used by the user program 2).
control is then transferred to the host computer 3).

At this time, the user program 2 estimates the number of used blocks for the preliminary file F2, and specifies the number of used blocks as the number of blocks to be allocated to the shared storage device 7.

The host computer 3 that received the open request
A pen request is issued to the central data management mechanism 30 to start the central data management mechanism 30.

The centralized data management mechanism 30 allocates the area of the file F2 for the specified number of blocks in the shared storage device 7, and transfers the file control block of the file F2 (the file control block having information as shown in FIG. 2) to the shared storage device 7. It is created on the specific memory device 5 based on the information in the file control information area 70 in 7.

When creating this file control block, the centralized data management mechanism 30 recognizes that the file F2 was opened by the user program 2 (open request from the host computer 4 side), and based on this recognition, , sets the copy existence flag in the file control block to ON (information indicating that a copy of the current file II? wholesale block exists), and copies the file control block of file F2 on the shared memory device 6. create.

When the above-mentioned open processing is completed, the centralized data management mechanism 30 notifies the user program 2 running on the host computer 4 via the host computer 3 that the open processing of the file F2 has been completed. Control is transferred to computer 4, and control is returned to user program 2).

The user program 2 creates data in the process of processing,
Write to file F2 based on the data creation
Issue a request.

Under the control of the host computer 4, write to the file F2 on the shared storage device 7 based on this write request is executed.
The te process is performed to update information such as the status of used blocks in the copy of the file control block of the file F2 existing on the shared memory device 6.

When the write process described above is completed, control returns to the user program 2.

User program 2 saves r of file F2 during processing.
Issue an ead request.

Under the control of the host computer 4, the file F existing on the shared memory device 6 based on this read request is
The copy of the file F2 on the shared storage device 7 is referenced based on the reference.
ad processing is performed.

When the above read process is completed, control returns to the user program 2.

Here, the user program 2 updates the data that has been read and performs write operations on the data again.
shall issue a request.

In the write process performed when user program 2 issues a write request again, if the data updated by the write process exceeds the size (block size) of file F2, all
ocate processing is performed.

That is, the host computer 4 notifies the host computer 3 of an instruction to allocate the necessary amount of space on the shared storage device 7, and the host computer 4
Control then transfers to the host computer 3.

The host computer 3 activates the centralized data management mechanism 30, and the activated central data management mechanism 30 allocates a necessary amount of area on the shared storage device 7.

Further, the centralized data management facility 30 includes a specific memory device 5.
The copy existence flag (in this case, ON is set) in the file control block of the file F2 that exists above is referred to, and the file F2 is stored on the shared memory device 6 based on the ON of the copy existence flag. It is recognized that a copy of the file control block exists, and the information in the copy of the file control block of file F2 existing on the shared memory device 6 is updated to reflect the above-mentioned allocaLe processing.

y r i t e including the above-mentioned allocate process
When the processing is completed, the centralized data management mechanism 30 notifies the user program 2 running on the host computer 4 via the host computer 3 that the write processing of the file F2 has been completed (from the host computer 3 to the host computer 4). control is transferred and control is returned to user program 2).

When all the processes for file F2 in user program 2 are completed, user program 2
issue a close request.

When a close request is issued, the host computer 4
Then control is transferred to the host computer 3, and the host computer 3 starts up the centralized data management mechanism 30 based on the close request.

The centralized data management mechanism 30 performs r e processing on the unused blocks in the file F2 on the shared storage bag F7, and sets the copy existence flag in the file control block of the file F2 existing on the private memory device 5. , and based on the ON of the copy existence flag, the information in the copy of the file control wholesale block of the file F2 existing on the shared memory device 6 is updated to reflect the above-described free processing.

Furthermore, the centralized data management mechanism 30 updates the file control information area 70 on the shared storage device 7 based on the update, and then updates the file control block of the file F2 existing on the private memory device 5 and the shared memory Device 6
Delete the copy of the file control block of file F2 that exists in .

When the above-mentioned close processing is completed, the centralized data management mechanism 30 notifies the user program 2 running on the host computer 4 via the host computer 3 that the C1ose processing has ended (the host computer 3
control is then transferred to the host computer 4, and control is returned to the user program 2).

As described above, for read/write processing (access) of a file from the host computer 4, a copy of the file control block of the file to be processed is created on the shared memory device 6, so the host computer 4 Processing can be performed by directly referencing/updating the copy of the file control block on the shared memory device 6 (without going through the host computer 3 that is the management host computer).

[Effects of the Invention] As explained above, the present invention enables host computers other than the management host computer to access files on a co-located storage device without going through the management host computer. It is possible to shorten the access time for files on a shared storage device from a host computer other than a computer (the effect of speeding up access is particularly large when a large amount of data is accessed), and one host computer (
This has the effect of avoiding concentration of load on the management host computer.

[Brief explanation of drawings]

FIG. 1 is a diagram showing the configuration of an embodiment of the present invention. FIG. 2 is a diagram showing the configuration of a file control block created in the private memory device in FIG. 1. In the figure, 1, 2... User program, 3.4... Host computer, 5... Private memory device, 6... Shared memory device, 7... Shared storage device. , 30... central data management mechanism, 70... file control information area. Figure 1

Claims (1)

[Scope of Claim] In a computer system in which one management host computer centrally manages files on a shared storage device that can be accessed by a plurality of host computers, created for the purpose of controlling the files by the management host computer. a specific memory device that stores file control blocks to be stored in the shared storage device; and a specific memory device that can be accessed from each host computer and is stored in the specific memory device when a file on the shared storage device is accessed from a host computer other than the management host computer. a shared memory device that stores a copy of the file control block of the file based on the file control block of the file; and a shared memory device that stores a copy of the file control block of the file based on the file control block of the file; Create a file control block for the file on the memory device and refer to it directly from the host computer other than the management host computer.
and a centralized data management mechanism in a management host computer that creates an updateable copy of a file control block of the file on the shared memory device based on the file control block of the file. Access method.