JPS62107350A - Data managing system for auxiliary storage device - Google Patents

Abstract

PURPOSE:To restore a data file to a previous state of an updating processing by saving only a previous state of the change of an object data file to be changed, when the updating processing is halted due to an accident of a service interruption, etc. in the course of the updating processing of a data file in an auxiliary storage device, and resetting the saved information after the updating processing is halted. CONSTITUTION:In an auxiliary storage device for storing and managing a data file (DF) by a list structure by a pointer, a saving area for an updating flag for showing that the DF is being updated, and a managing data group which is subjected to a change by an updating processing is provided, and after the updating flag is set, the DF is updated, and in case when it is a set state,the contents of the saving area are returned to the original managing data area, the updating flag is reset, and thereafter, the usual processing is executed.

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to a data management method for an auxiliary storage device, and the present invention relates to a data management method suitable for an auxiliary storage device that is directly incorporated into a device and whose storage medium is difficult to replace. Regarding the method.

[Background of the invention]

Conventionally, in auxiliary storage devices that store and manage data files in a list structure, if an accident such as a power outage occurs while updating one data file and the update process is interrupted, the list structure may be destroyed or data blocks may be destroyed. Problems such as logical disappearance of data occurred, and it was necessary to spend a great deal of effort to restore it. A management table is a way to solve this problem.
By duplicating all the pull and pointer groups and using the selector to select the valid and latest management table and pointer, all the previous states are saved during the update, and when the update process is interrupted, the saved previous state can be used to recover. There are known ways to do this.

However, if the number of data files to be stored and managed together with the names and attached information of each data file is large, this is not taken into account. As the storage area for storage devices increases, a problem arises in which the usage efficiency of the storage device decreases.

[Purpose of the invention]

An object of the present invention is to prevent the data file from being destroyed (before the update process) when the update process is interrupted due to an accident such as a power outage during the update process of the data file in the auxiliary storage device. It is an object of the present invention to provide a data management method for an auxiliary storage device that can restore the original state and reduce the storage area for a group of management data used for this purpose.

[Summary of the invention]

The method of managing data files in an auxiliary storage device using a list structure using pointers is widely used because the data files can be separated from the physical structure of the storage device, but the method of managing data files in a list structure using pointers is widely used. If this happens, the list structure will be destroyed and it will be difficult to restore it.One way to solve this problem is to duplicate all data files such as pointers, and restore the state before the update. Completely retain the data and select the selector indicating the valid management data group ('# How to switch when processing is completed, i.e.,
If the update process is interrupted, there is a way to restore the list structure by using the management data group before update indicated by the selector. The management data that is requested to be changed is the list structure of the data files to be updated and the list structure of unused data blocks, and the others are not changed.Therefore, in the present invention, only the pre-change state of the target to be changed is saved. Also, updates are in progress!
! By providing information that indicates when the update process has been interrupted, it is possible to detect the change after the update process is interrupted, restore the evacuation report, and restore the state before the update. It is possible to realize a data management method with high storage area usage efficiency even for an increased number of IJ ports.

[Embodiments of the invention]

An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a data file management configuration diagram showing an embodiment of the present invention. The data file group includes the name NAME, attribute AT, size SZ of each data file, and pointer DBHf to the first data block (directory group DIRO to DIR5 to store, and data block group DBO to DB15 to store the contents of the data file). , two data block pointer groups DBP0.DBP, which are a collection of pointers Pa to P15 that correspond one-to-one to each data block.
1, a selector DBP5 indicating which of these two data block pointer groups is currently valid, the number of currently stored data files DFN, and a pointer EDBH to the beginning of an unused data block. It is stored and managed by DMT. In addition, in order to perform the update process, there is an update 7-lag UPF that indicates that the update process is in progress, a saved data management table DMTS that saves the data management table before the update, and a pointer to the update data file (IJ number and pre-update directory contents). ? A save directory number DI几N and a save directory DIR8 to be saved are provided. The data block is
The contents of a data file are divided and stored, and the number of them and the size of one data block can be arbitrarily selected depending on the specifications of the storage device and system. Furthermore, each pointer in each data block pointer group is provided in a one-to-one correspondence with a data block, and when data is divided into multiple data blocks and stored, the corresponding pointer is Next, a final information (in the example, the data block number) is set to indicate the data block in which the data is stored, and end information (in the example, the data block number) is set in the pointer to the last data block to indicate that the data block ends there. , end pointer mark EP) are set. Thereby, the contents of each data file are configured as a chain list. Note that, as in the case of data files, information is set for each pointer group for unused data blocks so as to form a chain list. Regarding each chained list & data file mentioned above,
The pointer to the first data block of the corresponding chained list is
It is stored as a directory along with its name, attributes, and size, and for unused data blocks, a pointer to the first data block in the chain list is stored in the data management table along with the number of files currently stored. Pointer selector DBP8 is a temporary selector that indicates the data block pointer group that holds the currently valid chained list structure. For example, if this is set to 0', the first data block pointer group DBPo is valid. Therefore, this is used to read the data file, and when updating, the changed or stored structure is transferred to the second data block pointer group DBPI.
It shows that it is stored in .

Next, the operation of the embodiment will be explained using figures.

In the embodiment of FIG. 1, the name is 'A' +l
Three data files B + and + CI are stored, and the first
Since data block pointer group DBP0 is selected, the contents of each data file are stored in data file IA.
1 has data at, a in data blocks DBO, DB7.
2 is stored in data file /I/'B', data blocks DB1, DB! 1. DB6. DBIO,D
Data b1 to b5 are stored in BH, and data file I
C is data block DB8. DB9. DB12゜DB
Data C1 to C4 are stored in DB2.14, and unused data blocks are stored in DB2. DB4, DB5.1)8
13. It is DB+s.

In this embodiment, the operation of reading the data file ICI will be explained. First, search the directory QDIR,2)2 of the data file 1C, and use the pointer DBH to the first data block to retrieve the data block D.
Contents of B8 C1i reading page. Read the contents C2 of the data block DB9 indicated by the corresponding pointer P8 of the data block pointer group? : Read and similarly read the contents C6 of the data file DB12 indicated by the pointer P9, and also read the contents C6 of the data file DB12 indicated by the pointer P12.
Contents of a C4'f! :read out. At this point turtle pointer P
Since the end pointer mark EP is stored in 14, the read operation ends there.

Next, the data file update process will be explained using FIG. 2, FIG. 6, and FIG. 4. FIG. 2 is a flowchart showing the processing procedure for updating a data file. below,
According to the procedure shown in Figure 2, the contents of the data file I A I? :al'1a2°, the operation when updating to 35' will be explained.

First, the directory number 几1 of the data file I A I and its directory number (1 in the example) are
``Evacuation directory DIR8 and evacuation pointer respectively)
Stored in IJ number DIRH and data management table DM
Store T in the saved data management table DMTS and set the 7-lag UPFχ set state (for example, "1') during update. Next, connect the chain list of one data file IAI after the unused data block list. In the example,
Pointer P1 of first data block pointer group DBP0
5 contains the content IQ+ of the pointer DBH to the first data block of directory DII (,1), and the second
This is done by moving to data block pointer group DBP1. Then, the updated data a1', a2', aS''If are stored in DB2, DB4, and DB5 from the beginning of the unused data block, and the updated size S2 of the data file IA and the first data block are The directory DIR1 is updated with the pointer 121 to .The updated state so far is shown in Fig. 3.Subsequently,
The content 1B+ of the pointer P5 corresponding to the last data block pressure of the update data file °A° is moved to the unused data book head pointer EDBH of the data management table, and the end pointer mark EP is set in the pointer P5.

Thereafter, the pointer selector DBP5y+r31 is switched to 11', and the update processing flag UPF is reset. Management configuration diagram after being updated by the above update process?
It is shown in Figure 4.

Next, consider a case where the process is interrupted due to an accident such as a power outage, for example in the state shown in FIG. 3, before the above update process is completed. In this case, the 7-lag UPF is in the set state during the update process, and the pointer iJh of the update data file before update [and the data management table saved]
After the accident is recovered, by performing the recovery process according to the procedure shown in Figure 5, it is possible to recover to the state before the update process shown in Figure 1, and the update process can be restarted. When there is no evacuation report or update processing interruption report, the FL data block pointer group is set to the DB in Figure 6.
Because it is held in P1, when you try to read the data file, a data block larger than the size exists, or when you update it, a circular error occurs in the chained list, causing inconsistencies in processing. It becomes impossible to read the data file.

In the above embodiment, the number of data files to be managed was set to six, but this is not limited to six. (You may create as many directories as the number of data files you want to manage in the system. Also, the number of data files to be managed is six. As a refresher, attributes are given as an example, but this does not need to be limited, and various additional information may be added.In addition, in this example, in order to demonstrate the update process of data files, we will use the example of saving directories. Although only one area is provided, there is no need to limit it, and multiple areas may be provided depending on the processing content.

〔Effect of the invention〕

According to the present invention, even if an accident such as a power outage occurs while updating a data file in an auxiliary storage device and the update process is interrupted, the information before the update process and the information indicating that the update process has been interrupted can be combined. is retained, so after the accident has been recovered, this information can be used to quickly restore the state before the update process1.
The next update process etc. can be restarted. Moreover, only a small amount of management data is required for this purpose, and is the storage space usage efficient? There is no need to lower it.

[Brief explanation of drawings]

FIG. 1 is a data file management configuration diagram showing one embodiment of the present invention, and FIG. 2 is a flowchart showing a data file update procedure in the invention. FIG. 3 is a management configuration diagram of a data file during update processing in an embodiment of the present invention, and FIG. 4 is a management configuration diagram of a data file after update processing is performed for the embodiment shown in FIG. , FIG. 5 is a flowchart showing the data file management recovery procedure in the present invention. UPF: Update processing flag, DBP5: Pointer selector, DMT: Data management table, DF'
N...Number of stored data files, EDBH...Unused data block start pointer, DMT S -, -
Evacuation data management table, DII (, N... Evacuation directory number, DIR8... Evacuation directory, DI
Ro, DIRt, , DIR5-0, directory, NAME... file name, AT... attribute, S
Z...Size, DBH...Pointer to the first data block of the data file, DBPo, DBP1...Data block pointer group, PO, Pl. ..., P15... pointer, DBo, DB+,
..., DB15...data block.

Claims (1)

[Scope of Claims] 1. An updating flag for indicating that a data file is being updated in an auxiliary storage device that stores and manages data files in a list structure using a pointer, and a save area for a management data group that is changed by the update process. A data management method for an auxiliary storage device characterized by providing the following. 2. In the data management method for an auxiliary storage device as described in claim 1, when performing an update process, a group of management data to be changed by the update process is saved to a save area, and an updating flag is set. A data management method for an auxiliary storage device characterized in that a data file is updated after being set, and an updating flag is reset when the update is completed. 3. In the data management method of the auxiliary storage device according to claim 1, the updating flag is checked at initial startup such as when the power is turned on, and if it is in the set state, the contents of the evacuation area described in 1. A data management method for an auxiliary storage device characterized in that normal processing is performed after returning to the original management data area and resetting an updating flag.