JPH0863435A - File Recovery Method - Google Patents

Abstract

(57) [Summary] [Purpose] To reduce the amount of journals when a file error occurs and to recover to the state before the error occurred in a short time. [Configuration] After creating a backup file of a file to be recovered, a process of extracting the latest update information from journal information during the predetermined time and saving it in the update information storage file is repeated every predetermined time, When a failure occurs in the recovery target file, the recovery target file is recovered to the state before the failure occurrence based on the update information storage file and the journal information and the backup file up to immediately before the failure occurrence.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for recovering a file when a file failure occurs in an online system or the like.

[0002]

2. Description of the Related Art Conventionally, for example, when a file failure occurs in an online transaction system, the status before the failure occurs is recovered based on various history information (journal information) acquired during transaction operation. The following two methods are generally known.

(1) Transaction information is acquired as journal information, and the online process from backup acquisition is reproduced to restore the file to the state before the failure.

(2) Extracting the journal of file updates in preparation for abnormal termination of transactions and system shutdown,
By sequentially reflecting the backup files, the files are restored to the state before the failure.

[0005]

However, in the file recovery method of (1), when re-executed from the input transaction information, there is a guarantee that the processing will be performed in exactly the same order as the original online processing and that the data can be restored correctly. There is a problem that resources such as all business programs and files are required to reproduce online processing even if some file failures occur.

In the file recovery method (2), the recovery is accurate because the journals are acquired according to the update order of the files, and only the relevant file and the journal information in which the error occurred during the recovery processing. Although there is an advantage that it is not necessary, since journal information for recovery must be retained until the time when the next backup file is created, the amount of journals will increase and the recovery processing time will inevitably increase. There is a problem. Particularly, in a 24-hour online system, the amount of journal information becomes enormous and the recovery processing time also becomes long.

An object of the present invention is to provide a file recovery method capable of reducing the journal amount when a file failure occurs and recovering the state before the failure occurrence in a short time.

[0008]

In order to achieve the above-mentioned object, the present invention creates a backup file of a file to be recovered, and then, every time a predetermined time elapses, the latest journal information is obtained from the journal information during the predetermined time. The process of extracting the update information and saving it in the update information storage file is repeated, and when a failure occurs in the recovery target file, the recovery target file is created based on the update information storage file and journal information up to immediately before the failure and the backup file. It is characterized by recovering the state before the failure.

[0009]

According to the above means, the process of extracting the latest update information from the journal information during the predetermined time and saving it in the update information storage file is repeated every time the predetermined time elapses. As a result, the journal information before being stored in the update information storage file becomes unnecessary. That is, journal information for a predetermined time until it is stored in the update information storage file may be stored.

When a failure occurs in the recovery target file, the recovery target file is recovered to the state before the failure occurrence based on the update information storage file and journal information just before the failure occurrence and the backup file.

Therefore, in the online system,
It is not necessary to save all the journal information from the start of online work to the time of failure, and the amount of journal information is reduced. As a result, it is possible to reduce the journal file capacity and the file recovery time.

[0012]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below in detail with reference to the embodiments shown in the drawings.

FIG. 1 is a time chart showing an embodiment of the present invention. This time chart shows backup file 1, journal file 2, update information file 3
The relationship between a and 3b is shown by taking time on the horizontal axis.

In FIG. 1, a backup file 1 of a file 4 to be recovered when a failure occurs at the start of the online service is created.

After this, time t until a predetermined time elapses
Up to ₁ , the journal information in the transaction processing is sequentially stored in the journal file 2.

At time t ₁ , the latest update information is extracted from the journal file 2 and the update information storage file 3a is created.

Further, when a predetermined time elapses and time t ₂ is reached, a new update information file is created from the journal information of the journal file 2 and the update information of the update information storage file 3a from time t ₁ to time t _2. Create 3b.

After that, the same process as the process at time t ₂ is repeated every predetermined time.

If a file failure occurs in the middle of this repetition, for example, if a file failure occurs at time t ₃ , the update information file 3b created at time t ₂
Then, based on the journal information of the journal file 2 up to time t ₃ and the backup file 1, the contents of the file 4 in which the failure has occurred are recovered.

As described above, only the latest update information is extracted from the journal information and stored alternately in the update information storage files 3a and 3b every time a predetermined time elapses. The journal information is unnecessary and it is not necessary to save all the journal information until the time of failure.

That is, even for an online system that operates for 24 hours, it is sufficient to prepare a journal file capable of accumulating journal information for a predetermined time. In this case, two update information storage files 3a and 3b are required, but the update information storage files 3a and 3b are for extracting the latest information of the record of the file 4 from the journal whose transaction processing is completed. The update information storage files 3a and 3b are less than or equal to the capacity of the backup file 1, and the entire file capacity is smaller than in the case where all journal information is stored.

FIG. 2 is an explanatory view showing a method of extracting the latest update information from the journal information, and the update information extracting program 5 extracts the latest update information.

In this case, A, B, C, D, E, F ...
The file information management table 6 having the entry of is provided. Each of these entries is similar to a directory of files.

On the other hand, each record of the journal file 2 has a name corresponding to each entry as journal information.

When the latest update information is extracted from the journal information, it points as a journal record corresponding to the entry of the file information management table 6.

For example, entry A has A ',
A ″, B ′ for entry B, C ′ for entry C,
Point like C ″, C ″ ′.

The condition for pointing to a journal record is determined by the type of file, but if a general recovery journal for online transactions is used, the transaction has been completed (valid). .

Those in which the update of the file has not been canceled (there is no cancellation journal).

The order of points is the order in which the journal was acquired.

The latest update record.

According to these conditions, the journal record on the file management table 6 is pointed to, the latest update information is extracted, and registered in the second file management table 7.

At time t ₂ , when a new update information storage file 3b is created from the update information storage file 3a and the journal information, the update information storage file 3a is expanded in the memory and the second file management table 7 Add the latest update information registered in.

The present invention can be utilized for file recovery with respect to a file used for message waiting management in online control control (hereinafter referred to as a message waiting file).

This will be described below.

Message queuing management is a method for waiting for messages generated from a terminal to a central program in a online system, regardless of the state of terminals or lines or the state of a central program, until the central program can process them. It is a function that waits for a message from the center to the terminal until the destination terminal or line becomes available, or waits for a message between programs until the program at the message destination can process it.

The message waiting control described here is a function of assembling, registering, and reading (deleting) the message sent to the terminal and the central program among the above functions, and the message waiting file is assembled by the above function. This file temporarily stores (registers) the received messages.

Here, as a precondition of the message waiting control, messages are waited in units of a window, the substance of the message is acquired as a journal record before writing the message, and during the process of assembling the message after the journal is acquired. There are journals that will be retrieved if canceled for any reason, journals that will be taken when a message is assembled and written to a file, journals that will be retrieved when a message is read (deleted).

If the file recovery method described with reference to FIGS. 1 and 2 is used, (1) message recovery by the journal, message, and status extraction file becomes possible.

(2) The journals and messages can be extracted as follows.

A message is extracted by a valid journal up to the point immediately before the occurrence of a failure, and the message is registered as an update status file for each window.

The contact management table is expanded on the memory and the message entity is extracted from the journal for each contact.

A table for managing the process of assembling a message
Temporarily register in the (temporary contact table), and if there is an assembly completion journal, register it in the contact management table.

If there is a message cancellation journal during assembly, the message is discarded from the temporary contact table.

If there is a message deletion journal, the corresponding message is deleted from the window management table.

As described above, by repeatedly assembling, registering, and deleting a message in the counter management table, the message registered for each counter can be recovered. Register these messages in the extraction file.

The group of messages thus extracted can be called a new journal from which information unnecessary for queuing control is removed from the journal. That is, the journal effectively used for the message waiting control can be reproduced.

(3) Recovery of Message Waiting File The message waiting file can be recovered by applying the procedure of FIG. 1 by using the message extraction file and the journal immediately before the occurrence of a file failure after the extraction file is created. it can.

By carrying out the above (1) to (3),
Even if you do not have all the journals from the start of online business, you can recover the message waiting file only with the message extraction files that are acquired regularly and the journals acquired after that.

The present invention is not limited to the case of applying the message waiting control as described above, but can be applied to any file management for recovering from a failure.

[0050]

As described above, according to the present invention, after the backup file of the file to be recovered is created,
Every time a predetermined time elapses, the process of extracting the latest update information from the journal information during the predetermined time and saving it in the update information storage file is repeated, and when a failure occurs in the recovery target file, the update up to immediately before the failure occurs Since the recovery target file is restored to the state before the failure based on the information storage file and journal information and the backup file, the journal amount in the event of a file failure is reduced, and the status before the failure occurs in a short time. Can be restored to.

[Brief description of drawings]

FIG. 1 is a time chart showing an embodiment of a file recovery method of the present invention.

FIG. 2 is an explanatory diagram of a method of extracting the latest update information from journal information.

[Explanation of symbols]

1 ... Backup file, 2 ... Journal file,
3a, 3b ... Update information storage file, 4 ... Recovery target file, 5 ... Update information extraction program, 6, 7 ... File information management table.

Claims (1)

[Claims] 1. A process of, after creating a backup file of a file to be recovered, extracting the latest update information from journal information during a predetermined time and saving the file in an update information storage file at every predetermined time. Repeat,
When a failure occurs in the recovery target file, the recovery target file is restored to the state before the failure based on the update information storage file and the journal information and the backup file immediately before the failure occurred. .