JPH01163848A - Double updating preventing system for disk file - Google Patents

Abstract

PURPOSE:To improve the processing efficiency of a whole system by preventing the double updating of a block without an exclusive control from the reading of the block till its writing. CONSTITUTION:In a block 7 to be an updating object, an area to store information which smoothly indicates an updating time or the updating condition of the block 7 of an updating counter part 8, etc., is provided. Based on the information, the double updating of the block 7 from plural processes is prevented, and the exclusive control from the reading till the writing of the block 7 for the block 7 of the updating object or the file including the block is made unnecessary. Thus, the processing efficiency of the whole system is improved.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to prevention of double update of files, and particularly to prevention of double update of files on a random access storage device represented by a disk that is shared and updated simultaneously by yl number of processes. The present invention relates to a disk file double update prevention method suitable for preventing double block updates.

[Conventional technology]

In the conventional exclusive control method for shared files,
As shown in Japanese Patent No. 178455, a method was adopted in which reading from another central processing unit was made to wait between reading and updating. For this reason, processing performance deteriorates in proportion to the processing time from reading to updating.

[Problem that the invention seeks to solve]

In medium to large computer systems, multiple processes are often executed simultaneously in the central processing unit and share the same disk file. In particular, system registers such as disk files in which files are registered are always shared and updated. If multiple processes update the same memory block at the same time, it will result in a double update, so each process must take exclusive control over the target block or the file containing it before reading the target block into main memory. The block is then read, the contents of the block are updated on the main storage device, and the contents are further written to a disk file, after which the exclusive control is released. In the case of this method, the time required for exclusive control is determined depending on the processing time from block reading to writing.

Therefore, if this processing time is long, there is a problem that the exclusive control takes a long time, leading to a decrease in the processing efficiency of the entire system.

An object of the present invention is to prevent double updating of a block without exclusive control from reading to writing of the block, which would lead to a decrease in processing efficiency.

[Means for solving problems]

The above purpose is to collect information (hereinafter referred to as an update counter) that represents the uniqueness of the update state of data stored in a block on a disk, such as a block update time or a counter that is updated each time a block is updated. This is achieved by having 0. [Operation] The update counter in a block is always updated when writing is performed to update the block. Using this update counter, it can be determined whether the block to be updated has been updated by another process between reading and writing of the block being processed without exclusive control. When writing an update to the block, it is determined whether the update counter of the block is equal to the content at the time of reading, and if they are equal, the update data including the updated update counter is written to the block, and if not, it is written to the read channel again. Double updates from multiple processes can be prevented by issuing an instruction (hereinafter referred to as a channel program) to a channel connected to the disk and, if the disk is reread, processing again.

〔Example〕

Hereinafter, one embodiment of the present invention will be explained with reference to the drawings.

FIG. 1 is an equipment configuration for realizing the present invention, FIG. 2 is a flowchart showing the procedure of processing by the central processing unit of the present invention, and FIG. 3 is a channel of the present invention processed by the channel for the disk device. This shows the processing procedure of the program.

In FIG. 1, numeral 1 denotes a central processing unit, which performs input/output activation and calculations. A channel 2 controls input/output operations to the disk and transfers data to the main storage device. 3 is a main storage device, which includes a block buffer 4 for reading or writing blocks on the disk and a save area 5 for saving an update counter section. 6 is a disk, on which 7 blocks are recorded. Reference numerals 8 and 9 indicate an update counter section and a data section, respectively, which constitute a block.

Next, the operation of each part in FIG. 1 will be explained based on the flowcharts in FIGS. 2 and 5. In FIG. 2, a block to be updated recorded on the disk is read into a block buffer (201). The contents of the update counter section of the block read into the block buffer are saved to a save area (202). Update the update counter section of the block in the block buffer (for example, add 1) (
2OS). The data portion of the block in the block buffer is updated with the processing data to be updated (205). The channel program shown in FIG. 3 is issued to the channel connected to the disk (205). In the channel, position it in the block based on the procedure shown in Figure 3 (
5at). The update counter at the time of reading saved in the save area at 202 in FIG. 2 is compared with the update counter of the block on the disk, and it is determined whether there has been an update from another process since the time when the block was read at 201 (
302). If there is no update, the update counters are equal (YES 302), and if there is an update, they do not match (No 302). If the condition is YES in 502 with no update, 202.20 in FIG. 2 in the block buffer! The contents of the update counter updated in I and the contents of the data section are written to the block on the disk. In the condition of 302 ONO where there has been an update from another process, the process must be restarted from reading the family block, so the block is read (304). next,
In Fig. 2, check the execution results of the channel program in Fig. 3, and if the condition is YES in 503 in Fig. 3, there was no update from another process, so what should be done? End it, No
If the condition is , there has been an update in another process, so the process returns to step 202 to redo the process (206).

〔Effect of the invention〕

According to the present invention, when a file on the same disk is shared and updated by multiple processes, the period from reading to writing to update a block in the file is exclusive to stop other processes. This eliminates the need for control and improves the processing efficiency of the entire system.

[Brief explanation of the drawing]

FIG. 1 is an equipment configuration diagram of an embodiment of the present invention, FIG. 2 is a flowchart of a processing procedure performed by a central processing unit, and FIG. 3 is a flowchart showing a channel processing procedure. 1... Central processing unit, 2... Channel, 3...
Main storage device, 4...Block buffer, 5...
- Save area, 6...Disk, 7...Block, 8...Update counter section, 9...Data section. 7 Block 8 Saragen Tsuntamo Tosho 2 Figure q Chibu town name 3 M

Claims (1)

[Claims] 1. In a data processing system in which a block stored on a random access storage device may be simultaneously accessed and updated by multiple processes operating simultaneously on one or more central processing units, in a block to be updated. An area is provided to store information that uniquely indicates the update status of a block, such as the time of update or an update counter, and based on this information, double updates of blocks from multiple processes can be prevented and the block to be updated or this A disk file double update prevention method characterized by eliminating the need for exclusive control from block reading to block writing for a file containing the file.