JPH01173140A - File control method - Google Patents

Abstract

PURPOSE:To quickly and accurately grasp the file abnormality by saving the result of detection when the file abnormality is detected in an on-line state and checking the saved information for identification of an incomplete file in an off-line state. CONSTITUTION:When the file abnormality is detected in an on-line state, the result of this detection is saved. Then the saved information is checked in an off-line state for identification of an incomplete file. When a power supply is applied or reset, the saved information is confirmed, deleted and checked newly. Thus the file abnormality is detected at the start of a system and therefore a hang-up phenomenon of the system due to the file abnormality can be avoided. Furthermore a user can know the file abnormality caused in a working mode and can perform the due countermeasure to avoid the abnormal run of a program. Then the file abnormality can be checked and corrected at an optional time point. Thus it is possible to prevent the abnormal run due to the file abnormality for a program which works later in an on-line state.

Description

[Detailed Description of the Invention] [Industrial Field of Application] The present invention provides a system for dynamically managing a collection of data (files) on an auxiliary storage medium of a computer by checking the correctness of the management information, This article relates to a file management method that allows for repair of abnormalities.

[Conventional technology]

Management information for managing files on auxiliary storage media including disks includes a file allocation table (hereinafter referred to as
Also called FAT. ), root directory table (hereinafter also referred to as RDiRT), and subdirectory table (hereinafter also referred to as 5DiRT). FAT is a table that stores the usage status of a storage medium, such as in which cluster (the smallest unit for managing files) files and 5DiRT are located on the storage medium, the connection status of the clusters, and the location of free clusters. a! D, RDiRT is a table containing information such as the file name, the number of the first cluster indicating the location of the file on the storage medium, and the file size.

Furthermore, 5DiRT has the same structure as RDiRT, and is a table for fc management of a "collection of child files" called a subdirectory, which is further created within a collection of files called a root directory.

FIG. 6 is an explanatory diagram for explaining the hierarchical structure of directories. In the same figure, 1 is RDiRT12A, 2B
indicates 5DIRT, and 3A to 3D indicate the file body. A directory is a collection of files, and the root directory contains not only files but also subdirectories that are child directories. The directory table contains the file name NA, the cluster number CN indicating the first cluster position on the recording medium, and the file Cf'.
) or BJ(d) is stored, and this configuration is used for RDi.
RT and 5DiRT are the same. Therefore, to know where the files are located on the storage medium, you can specify the name of the IJ group in the hierarchy and look at the files in the directory one number below. .

FIG. 7 is an explanatory diagram for explaining the parent-child relationship of directories, and this is an example in which the subdirectory table 2C is a child. Generally, in order to search for files in a child's subdirectory, only the data in the subdirectory table is saved as necessary, but the parent's directory table is discarded. Therefore,
After locating the subdirectory, you can locate the parent directory even after the search is complete.
By looking at the first cluster number of the subdirectory itself, the cluster link of the subdirectory,
In other words, it is possible to know which of the multiple clusters it is linked to.

Figure 8 is an explanatory diagram for explaining the link between FAT and cluster.
iRT or 5DiRT)t", Figure (B) shows the FAT, and Figure (C) shows the location of the file on the storage medium. In other words, in Directory Table 2 of Figure (A), the corresponding The cluster number of the beginning of a file or subdirectory (the beginning position in the storage medium,
1010''), then search FAT4 in the same figure (b) in order to find out the cluster combination (= link).The corresponding cluster number position of this FAT4 has the next linked Also, the last cluster number position contains a final mark, and knowing the last one allows you to access the program's topmost program (i.e., a program that cannot be called directly by a general user). Yes, it is centrally managed.For this reason,
Although it is almost impossible for the storage medium to be destroyed other than by physical destruction, this information may be revealed during the development process of a high-level program comparable to the highest-level program. - Once the computer system is corrupted, when the necessary files are loaded into the main memory and executed, the system itself may hang up when the computer system is started, and its execution may no longer be normal. This may cause serious malfunctions such as crashing. Alternatively, even in the event of a minor failure where an abnormality in a data file is simply detected, rather than a major failure, check the one-to-one correspondence between the file name and file location and check the name or file location. A great deal of time will be wasted making corrections. The specific matters of these defects will be explained below.

(B) Problem I: A cluster is a collection of bytes of data, and file management is done in raster units. Also, the file size F in the directory table is in bytes, but if the number of cluster links is N, then the total number of bytes of cluster links is NX
Although it is smaller than C (C is the number of bytes per cluster), it is never smaller than the number of bytes for (N-1) clusters, that is, (N-1)C<F≦NxC-- -----(1) must be satisfied. Therefore, if the consistency between the file size and the number of cluster links is incorrect, it is an incorrect file.

(B) Problem No. ■ If the cluster link in the FAT shown in FIG. 8 (B) overlaps with multiple files, the relevant file is incorrect.

(c) Problem # ■ In the FAT shown in Figure 8 (b), some link data other than an unused mark or final mark is placed in the FAT position corresponding to a cluster that is not linked from any file or subdirectory. It may be included. This is a result of the previous FAT being erased by overwriting the FAT, and means that the first cluster number indicated by the directory table for at least one file is incorrect, resulting in a problem. Often accompanied by.

(D) Malfunction (Sub-buildf shown in FIG. 7) This is a case where an abnormality occurs in the first and second directories from the top of the IJ table 2C. In other words, the best one is your own 5DiRT
However, if this is broken, control will be transferred from the parent directory to the subdirectory, and incorrect results will be produced during the process of searching for the link mode of the subdirectory itself. Also, if there is an error in the second item, it will not be possible to return to the parent directory. Such a malfunction can lead to a complete system hang-up, which can lead to a serious malfunction.

Therefore, it is an object of the present invention to enable quick and accurate identification of file abnormalities and to ensure safe system operation.

[Means for solving problems]

When an abnormality is detected in a file while online, the results are saved, and when offline, the saved information is checked to identify incomplete files, and the process includes repairing errors using a man-machine interface. On the other hand, when power is turned on or reset, processing including confirmation, deletion, and new checking of the saved information is performed.

[Effect]

In order to solve the above problems (a) to (d), FAT
, RDiRT is checked, and 5DJRT is checked and repaired. However, in the FAT and RDiRT checks, the consistency between the file size and the number of cluster links is checked, the cluster link is checked for sloppiness, and the usage that is not linked from anywhere is checked. In the 5DiRT check, the 5DiRT's own cluster number and the cluster number of the parent directory are checked.

Each check checks for abnormalities during online operation,
If there is an abnormality, the abnormality information is saved in a storage medium. When the power is turned on or reset, that is, when the system starts, there is a possibility that the abnormal file does not already exist, or it may have been deleted and then created as a new normal file in a different location on the storage medium. Check whether the abnormal information matches the current situation, taking into consideration the fact that there is fluidity in At the same time, other files are also checked, and if any abnormality is found, the abnormality information is saved in the storage medium.

On the other hand, you can also check offline at any time.
Abnormality information saved in order to deal with the abnormality as soon as possible CR
The display is displayed via a man-machine interface device such as a CRT to let the user know which file or storage medium has an abnormality, and the file is deleted through interactive processing with the man-machine interface device such as a CRT. This will allow you to replace the files later.

Finally, since files change dynamically, it is possible to check them at system startup, online and offline, and if there is an abnormality, it can be repaired in a safe manner. This is intended for operational purposes.

〔Example〕

FIG. 1 is an explanatory section for explaining the invention in detail. This shows the processing contents to be executed each time when online, when the power is turned on, and when offline.

That is, when online, the check program is called from the user program to perform the processes 1) to 3) in Section "Arts and Crafts." The flowchart is shown in FIGS. 2A to 2C. Figure 2A shows the consistency check between the file size and the number of cluster links (whether the formula (1) is satisfied or not), Figure 2B shows the cluster link redundancy check, and Figure 2C shows performs subdirectory table checking and repair operations, respectively.

Note that the details will be omitted as they are redundant.

In addition, when the power is turned on or reset, depending on the program that is automatically started at that time, 1) ~
4). The flowchart is shown in Figure 3 through Figure 3F.

Figure 3A shows the checking and operation of abnormality information for saved files regarding the consistency between file size and number of cluster links, Figure 6B shows the new checking operation for all files, and Figure 3C shows the operation. Regarding the cluster link double check, the check operation for saved files is also shown in Figure 3D for all files,
A check operation is performed on all subdirectory tables, a check operation is performed on used clusters that are not linked from anywhere in Figure 3E, and a check and repair operation is performed on the IJ table (in Figure 3F). The details are also omitted.

Furthermore, when off-line, a command is issued as appropriate to start a predetermined batch processing program, and processes such as 1) to 4) in (2) in FIG. 1 are performed.

Figure 4 shows the consistency checking operation between the file size and the number of cluster links mentioned in 1), and 2)
For items 3) and 4), please refer to ■2-a respectively.
), 3) and 4), so the flowchart has also been omitted.

Each mode (on/off, power on, offline)
Figure 5 illustrates the various check points in the process. A) Checking consistency between file size and number of cluster links 0 First, if there is an error while online, save the error information.

Next, check this abnormality information when the power is turned on and when offline. At this time, if the file does not already exist or there is no abnormality, the abnormality information is rewritten (rewritten).

When the oiIC power source is turned on, all files are checked.

B) Cluster link redundancy check 0 First, check all files when the power is turned on, and if there is an abnormality, save the abnormality information. We also check old abnormal information and update it to suit the current situation.
Delete or rewrite.

0 Check this old error information when online or offline, and delete or rewrite it to match the current situation.

C) Check and repair subdirectory table 0IE
When S is turned on and offline, all subdirectory tables are repaired.

0 When online, only the corresponding subdirectory is repaired.

D) Check for used clusters that are not linked from anywhere 0 Due to its nature, this is not performed when online, but only when the power is turned on and offline.

〔Effect of the invention〕

According to this invention: a) File abnormalities are detected at the time of system startup, so it is possible to prevent system hang-up due to file abnormalities (effect at system startup).

b) As a result of the user being aware of file abnormalities that occur during operation and being able to take appropriate measures, abnormal program execution can be prevented (online effect).

C) File abnormalities can be checked and corrected at any time, so it is possible to prevent abnormal running of a program that will later run online due to file abnormalities (offline effect).

d) Since file abnormalities can be checked at the three timings mentioned above, dynamic file generation,
It is possible to adequately cope with extinction (effect on system change).

e) Since file abnormalities can be checked at the six timings mentioned above, it is possible to significantly reduce the amount of time required to detect and correct abnormal files), and the system throughput can be increased.

[Brief explanation of the drawing]

FIG. 1 is an explanatory diagram for explaining the invention in detail, and FIG.
Figures A to 2C are flowcharts for explaining various operations when online, Figures 3A to 3F are flowcharts for explaining various operations at power-on or reset, and Figure 4 is a flowchart for explaining various operations when offline. An explanatory diagram to explain the operation, Fig. 5 is an explanatory diagram to explain various check points in each mode) 1 - An explanatory diagram to explain, Fig. 6 is an explanatory diagram to explain the hierarchical structure of the directory, Fig. 7 8 is an explanatory diagram for explaining the parent-child relationship of directories, and FIG. 8 is an explanatory diagram for explaining the link between the file allocation table and the cluster. Code explanation 1...Root directory table (RDIR)
T), 2.2A to 2D...subdirectory table (SDiRT), 3A to 3D...file body, 4...7 aisle table (FAT), DE...Identification symbol, NA...
...File name, CN...Start cluster number. Agent Patent Attorney Akio Namiki Agent Patent Attorney Kiyoshi Matsuzaki Figure 1 Figure 2A Figure 2C Figure 3A Figure 3B Figure 30 Figure 3C Figure 4

Claims (1)

[Claims] A management method for dynamically managing a data collection (file) on an auxiliary storage medium of a computer system based on its management information, wherein when an abnormality in a file is detected in an online state, The results can be saved, and while offline the saved information can be checked to identify incomplete files and can be processed in the form of a man-machine interface, including repairing errors, while on power-up or reset. A file management method characterized by performing processing including confirmation, deletion, and new checking of the saved information.