JP2000222265A - Restoring method for file using extracted journal - Google Patents

Abstract

PROBLEM TO BE SOLVED: To shorten a total journal integration time and to efficiently and easily restore a faulty file by sorting and extracting database update information (FJ) temporarily as an extracted journal to an update information storage file. SOLUTION: In response to a journal swap at time t1, an FJ extraction program 22 is started to perform storage in an update information accumulation file 23 through sorting into A1-1, A1-2, and A2-1. Further, storage in an update information storage file 24 is performed through sorting into B1-1, B1-2, and B2-1, At this time, pieces Z-1 and Z-2 of journal information which are not necessary for file restoration are not saved in the update information storage files. At the time of a journal swap at time t2, storage in an update information storage file 26 is similarly performed through sorting. At the end te (system end time) of the operation day, an FJ integration program 28 is started and only the latest update information is extracted from the update information storage files 23 and 26, and 24 and 27 and saved in integration journal files 29 and 20.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for acquiring journal information in an online system and a method for recovering a file in the event of a file failure. The present invention relates to an acquisition method and a file recovery method using an extraction journal.

[0002]

2. Description of the Related Art Conventionally, in an online system or the like,
When a file failure has occurred, the journal information is sequentially overwritten on the backup file of the file to be recovered, thereby recovering the state immediately before the failure occurred.

A system has been put into practical use that provides a file recovery method that can reduce the amount of journal used for the recovery processing and recover the state before the failure in a short time.

As described in Japanese Patent Application Laid-Open No. H8-63435, for example, this type of file recovery method creates a backup file of a file to be recovered and then stores only database update information necessary for volume failure recovery. Is accumulated for each journal swap, and if a failure occurs in the file to be recovered, the database update information (accumulation journal) up to the time immediately before the failure occurs is overwritten in the backup file, and the failure occurs in the file to be recovered. Recover to the previous state.

[0005]

According to the above method, when storing the database update information, the database update information output during the current journal swap is sorted into the database update information (accumulated journal) created during the previous journal swap. -Merge to create the latest database update information (aggregated journal). The process of creating the database update information (accumulated journal) (hereinafter, FJ integrated program) is performed for each journal swap. However, in a system that updates a large amount of files with a large capacity, the number of journal swaps is enormous. Further, since the sort / merge is performed for each journal swap, there is a problem that the efficiency is low and the journal accumulation time is too long.

[0006] The object of the present invention is the following two points.

(1) To provide a method for reducing the total journal integration time.

[0008] (2) When a file failure occurs, the failure file can be efficiently and easily recovered by using the accumulated journal created before the previous operation day and the update information accumulation file of the current operation day. Provide a recovery method.

[0009]

In order to achieve the above object, the present invention provides (1) a method in which FJs are once extracted and sorted into an update information storage file as an extraction journal;
A method of shortening the total journal accumulation time is provided by performing sort / merge of J once per one working day.

(2) A method is provided in which the total journal integration time is further reduced by activating the FJ integration program multiple times.

(3) When a file failure occurs during a certain operation period, the backup file is stored in the backup file as an intermediate file of the accumulation journal before the previous operation day and the intermediate file of (1) without waiting for the creation of the accumulation journal on the current operation day. Provided is a file recovery method capable of efficiently recovering the state before the failure by reflecting the extracted journal in the created update information storage file of the current operation day and the subsequent journal information.

[0012]

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

FIG. 1 illustrates the flow of operation when an integrated journal is acquired at the end of an operation day without creating an extraction journal. Data A1 and A1 are stored in files 11 and 12 to be backed up, respectively.
2. There are data B1 and B2.

First, the update information A of the data A1 is stored in the journal file 13 at the time of journal swap at time t1.
1-1, A1-2. Update information A2-1 of data A2. Update information B1-1 and B1-2 of data B1. Update information B2-1 of data B2. Unnecessary journal information (other than FJ) Z-1 and Z-2 exist at the time of file recovery at random.

Next, the journal file 15 at the time of journal swap at time t2 contains the update information A of the data A1.
1-3, A1-4. Unnecessary journal information Z-3 exists at the time of file recovery. And the last te of this operation day
At the time of system termination, the FJ integrated program 17 is started, and integrated journal files 18 and 19 are created by extracting only the latest update information of the files 11 and 12.

The FJ integrated program 17 first sorts and merges only the update information necessary at the time of file recovery using the data name and the update time as keys. After that, only the latest update information A1-4, A2-1, B1-2, B2-1 of each data is extracted, and the accumulated journal files 18, 19 are respectively obtained.
To save.

As described above, the method of sorting and merging a large amount of randomly arranged update information at once for one operating day is enormous in the time required for the sort and merge, and is inefficient.

FIG. 2 explains the flow of operation when an extraction journal is created from the journal file of FIG. 1 and an integrated journal is obtained. Figure 1
Update information storage file 23 corresponding to file 11 of
26, update information storage file 2 corresponding to file 12
4, 27 are prepared. The FJ extraction program 22 is started at the time of the journal swap at the time t1, and is sorted and stored in the update information storage file 23 as A1-1, A1-2, A2-1. The update information storage file 24 stores B1-1, B1-2, and B2-1.
At this time, unnecessary journal information Z-
1, Z-2 are not stored in the update information storage file.

Similarly, at the time of journal swap at time t2, the update information of the file 11 is sorted and stored in the update information storage file 26 in the order of A1-3 and A1-4. Then, at the end of the current operation day (at the end of the system), the FJ integrated program 28 is started, and the update information storage file 2
Only the latest update information A1-4 and A2-1 are extracted from the files 3 and 26 and stored in the integrated journal file 29. Similarly, from the update information storage files 24 and 27, B
1-2, B2-1 is extracted and the collected journal file 2
Create 0.

As described above, by sorting update information in advance in small units called journal swaps and removing unnecessary information, sort / merge processing can be performed efficiently, and journal integration time can be reduced. it can.

FIG. 3 illustrates a file recovery embodiment in the case where a failure occurs in a file whose backup is acquired every few days. First, when the system is closed on the operation day d1, a backup file 32 of the file 31 to be recovered is created.

On the next operation day d2, triggered by the journal swap, only the FJs are extracted from the journal file 33 and sorted and stored in the update information storage file 34. And at the end of the system, the accumulated journal file 3
Create 5.

At time t1 on the next operation day d3, file 3
If a failure occurs in the first backup file 3
2, the failed journal file 35 is sequentially overwritten with the accumulated journal file 35 created on the previous day (operating day d2), the update information accumulation file 37 on the current day (operating day d3), and the journal file 36 until time t1. 31 is restored to the file 38.

[0024]

According to the present invention, the following effects can be obtained.

(1) Reduction of journal accumulation time: Since information unnecessary for file recovery is deleted in advance and only information necessary for recovery is extracted, it is possible to reduce the time for reading unnecessary information during journal accumulation.

To sort when extracting journals,
It is possible to reduce the sorting time when collecting journals.

The journal recovery time can be further reduced by dividing the recovery target file into groups in advance and activating the FJ integration program multiple times.

(2) Reduction of File Recovery Time The file recovery time can be reduced by using the update information storage file created at the time of journal extraction for recovery in the event of a file failure.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram of a method of acquiring an integrated journal without creating an extraction journal according to the present invention.

FIG. 2 is an explanatory diagram of a method of creating an extraction journal and acquiring an integrated journal according to the present invention.

FIG. 3 is a diagram showing a file recovery embodiment of the present invention.

[Explanation of symbols]

11, 12, 31, 38 ... files, 23, 24, 2
6, 27, 34, 37 ... update information storage file, 18,
19,20,29,35 ... Collecting journal file, 3
2. Backup file.

Continued on the front page (72) Inventor Takehiro Shoji 890 Kashimada, Saiwai-ku, Kawasaki-shi, Kanagawa Prefecture F-term in Information Systems Division, Hitachi, Ltd. 5B082 AA01 DB04 DD06 DD07 EA10 GA04 GC04

Claims (3)

[Claims] 1. During a certain operation period (hereinafter, operation day) set by a user, only database update information (hereinafter, FJ) necessary for volume failure recovery is extracted from journal information and update information is stored. The process of sorting and storing the information in a file is repeated, and the FJs in the update information storage file are further sorted and merged after the operation is completed, thereby shortening the time required to create an integrated journal having only the latest update information. How to recover files using the journal. 2. The total journal accumulation time is further shortened by dividing the backup target file into groups in advance, preparing update information accumulation files for each group, and multiplexing the processes for acquiring accumulation journals. File recovery method using extraction journals. 3. In the event of a file failure, a backup file contains, in the backup file, an integrated journal acquired before the previous operating day,
The feature is that the file recovery time is reduced by restoring the recovery target file to the state before the failure occurred by reflecting the extracted journal accumulated until just before the failure occurred and the subsequent journal information acquired during the operation period. File Recovery Method Using Extract Journal.