JPH0259838A - File restoring system applying optimization of storing position for file restoring information - Google Patents

Abstract

PURPOSE:To start the restoration at and after the latest log file and to shorten the restoring time when a file fault occurs by preparing a restoring file information part, a restoring file information retrieving part, a restoring file update request part, a log file swap event deciding part, and a log file swap event informing part. CONSTITUTION:A file restoring information output part 10 includes a restoring file information control part 11, a restoring file information retrieving part 12 and a restoring file update request part 13. A log file access control part 20 includes a log file control part 21, a log file swap event deciding part 22, a log file writing part 23 and a log file swap event informing part 24. A file access control part 30 includes a file updating part 31 and a log file writing request part 32. Then a log file 40 is added. When a file #1 has a fault, the record address updated in the file #1 and an update record image are extracted out of a log file #2. Then the file #1 is updated with the update record image. Thus the file #1 is restored and this restoring time is shortened.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a file recovery method when a file failure occurs, and particularly to a file recovery method that optimizes the storage location of recovery information.

[Prior art] Conventionally, this type of file recovery method saves file storage information before the file is updated, and then logs the updated information when the file is updated by a program. ing. Then, when a failure occurs in the file, the file is set on a normally accessible magnetic disk, the previously saved information is stored in that file, and the file update information up to the time the failure occurs is stored. By doing so, the file was restored to an updated state.

[Problem that the invention aims to solve] In the conventional file recovery method described above, files are recovered using all of the file update information from the time the file storage information was saved until the time a file failure occurred. Therefore, the disadvantage is that the larger the amount of file update information, the longer it takes to restore the file.

[Means for Solving the Problems] The file recovery method of the present invention by optimizing the storage location of file recovery information writes the record to the file when a file write or reference request is received from the program.
Or a file access control means that reads from a file and requests to write information of the record to a log file, and when a write request is received from the file access control means to the log file, writes data to the log file and the log file becomes full. a log file access control means that notifies a certain program of a full event when the log file becomes full; a log file that stores and accumulates data when there is a write request to the log file from the log file access control means; When a log file becomes full due to the file access control means, the file is read once in order to accept the full event and store the recovery information of the file to be restored in the next log file.
and file recovery information output means for instructing the file access control means to issue a write request based on the image.

[Example] Next, the present invention will be explained with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of one embodiment.

The file recovery method according to the present invention includes a file recovery information output section 10 that includes a recovery file information management section 11, a recovery file information search section 12, and a recovery file update request section 13.
, a log file access control unit 20 consisting of a log file management unit 21, a log file swap event determination unit 22, a log file write unit 23, and a log file swap event notification unit 24, a file update unit 31, and a log file write request unit. 32, and a log file 40.

Referring to FIG. 2, the recovery file information management unit 11:
For each file to be restored, the file name and the address of the record to be restored in the file are set. In this example, file #1. #2゜#3 is the file to be restored, file #1 has an al record, file #
2 indicates the b code, and file #3 indicates that the C1 and 02 records are the records to be recovered.

Referring to FIG. 3, in the log file management section 21, a file name, capacity of each log file, writable capacity of the log file, and start address at which the log file can be written are set for each log file.

FIG. 4 shows the record format of the log file 40, which is composed of updated file information, the address of the updated record, and an image of the updated record. This example shows that record a of file #1 has been updated.

Returning to FIG. 1, when a certain log file 40 becomes full, the recovery file information search unit 12 receives a notification to that effect from the log file access control unit 20, and refers to the recovery file information management unit 11. It has a function to obtain the file to be restored and the address of the record to be restored in the file. The recovery file update request unit 13 reads the record stored at the record address obtained by the recovery file information search unit 12, and sends a request to write to the file at the same address in order to store the image of the record in the log file 40. have the ability to do so.

The file update unit 31 has a function of recognizing the corresponding file when there is an update request for a file and storing the record at the specified address. The log file write request unit 32 has a function of requesting to write an image of the updated record to the log file 40 when a file is updated.

When a write request is made to the log file 40, the log file swap event determination unit 22 refers to the log file management unit 21 and determines a writable log file,
The size of the record requested for writing is compared with the writable capacity of the log file, and if writable, a writable start address is obtained. If the log file is not writable, the swap event determination unit 22 sets the writable flag of the log file management unit 21 to the effect that the log file is not writable, and indicates to the log file that the log file is full. It has the function of notifying the file swap event notification unit 24, and also recognizing other writable log files and obtaining the writable address of the log file. The log file writing unit 23 stores the record requested to be written to the log file writable start address obtained from the log file swap event determination unit 21 in the corresponding log file, and writes from the writable capacity of the log file. It has a function that subtracts the size of the record and changes the writable start address to the next record address of the written record. When the log file swap event notification unit 24 is notified by the log file swap event determination unit 21 that the log file is full, it determines the program to which the event should be passed and notifies the program. have

Next, the operation of the file recovery method by optimizing the file recovery information storage location of this embodiment configured as described above will be described.

When there is an update request for a file, the file update unit 3
1 updates the corresponding record at the specified address and notifies the log file write request unit 32 that the file has been updated. The log file write request unit 32 issues a write request to the log file access control unit 20 in order to store the updated file information, updated record address, and updated record image in the log file 40.

The log file swap event determination unit 22 refers to the log file management unit 21, obtains the addresses of the storable log file #1 and the log file to be written, and the log file writing unit 23 writes the file update information to the addresses. Write. In this way, there are file update requests one after another, and when the log file swap event determination unit 22 recognizes that it is no longer possible to write to log file #1,
It instructs the log file writing unit 23 to write to the writable log file #2, and notifies the log file swap event notification unit 24 that the log file #1 is full.

When the log file swap event notification unit 24 receives the notification, it notifies the file recovery information output unit 10 of the event.

Upon receiving the event that log file #1 is full, the recovery file information search unit 12 refers to the recovery file information management unit 11, obtains the record address of the file to be recovered, and sends the request to the recovery file update request unit 13. An instruction is issued to update the record. The recovery file update request unit 13 reads the record from the file and issues a request to write the read record image back to the file as is at the same address.

The file access control unit 30 writes the record in the file, and issues a request to the log file access control unit 20 to write the updated file information, update record address, and update record image in the log file.

The log file access control unit 20 writes the record requested to be written to the log file #2.

Therefore, when log file #1 becomes full, all the record information of the file to be restored will be written to log file #2.

A file recovery method when file #1 becomes a failure in this state will be explained with reference to FIG.

First, secure file #1 in a location where it can be accessed normally (100). Store the contents of file #1 before it is updated in the storage medium 130 in advance, and
The contents of 0 are temporarily stored in file #1 (110).

Next, in order to restore the updated record in file #1 to the latest image, extract the record address updated in file #1 and the updated record image from log file #2, and restore it to file #1. File #1 can be restored by updating (120) with the updated record image.

[Effects of the Invention] As explained above, the present invention includes a recovery 7-eye information section, a recovery file information search section, a recovery file update request section,
By providing a log file swap event determination section and a log file swap event notification section, even if a file failure occurs, the file can be recovered from the latest written log file, reducing recovery time. There is an effect.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention.
Figure 2 shows the configuration of the recovery file information management unit shown in Figure 1, Figure 3 shows the configuration of the log file management unit shown in Figure 1, and Figure 4 shows the configuration of the recovery file information management unit shown in Figure 1. The structure of the record format of the log file shown in FIG. 5 is shown, and FIG. 5 is a diagram for explaining file recovery means when file #1 becomes a failure. 10... File recovery information output unit, 11... Recovery file information management unit, 12... Recovery file information search unit, 13... Recovery file update request unit, 20... Log file access control unit, 21...Log file management unit, 22...Log file swap event determination unit, 23
...Log file writing unit, 24...Log file swap event notification unit, 30...File access control unit, 31...File update unit, 32...Log file write request unit, 40... logfile. Figure 2 Figure 3 Kuchi Shuko each payment record Figure 4 n

Claims (1)

[Claims] 1. A file access control means for writing or reading the record to or from the file and requesting writing information of the record to a log file when there is a file write or reference request from the program; A log file access control means that writes data to the log file when there is a write request to the log file, and notifies a certain program of the full event when the log file becomes full; The log file stores and accumulates data when there is a write request to the log file, and the above log file access control means accepts the full event and restores to the next log file when the log file becomes full. and a file recovery information output means for once reading the file in order to store the recovery information of the file to be restored, and instructing the file access control means to write a write request using the image. File recovery method by .