JPH05108444A - Journal extraction system - Google Patents

Abstract

PURPOSE:To reduce journal capacity and to shorten data restoration time when data is destroyed in a file device by extracting only the final record content of the alteration request as against a record in the file device, which is outputted from a program, as a journal. CONSTITUTION:When there is the alteration request of the record as against the data file device 25 from the program 21, buffer management judges whether it can be processed in a buffer 22 or the record of the data file device 25 is to be really updated or not. At the time of really updating the record, journal management decides the storage place of a before journal and executes the instruction of the output of the before journal and the alteration of the record in the file device 25 to input/output management. The journal management decides the place where an after journal is stored after all the records which the program is altered are updated, and executes the instruction of the output of the after journal and the alteration of the record in the file device 25 to input/output management to release the before journal.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a journal collection method used for file device failure recovery.

[0002]

2. Description of the Related Art In the conventional journal collection method, when a program issues a request to change the contents of a record in a file device, the processing of the change request is completed at the time when the capacity of the input / output buffer reaches its limit. Each time I did, I wrote a journal as a record of record changes.

[0003]

In the conventional journal collecting method, journals are collected as a history of change requests for records in the file device issued from a program, and data in the file device is destroyed due to a program error or device error. When a problem occurred, the data in the file device that was destroyed was restored by restoring the data just before the journal was collected and then writing back the collected journal records in order.

However, in this journal collection method, the journal record is additionally output even when a change request for the same record is issued, so the capacity of the journal increases. In proportion to this, the data recovery time when the data in the file device is destroyed is extended.

The present invention reduces the journal capacity, which is a drawback of the conventional journal collecting method, by collecting only the final record contents of the change request for the record in the file device issued from the program as a journal, The purpose is to shorten the data recovery time in the event of data corruption.

[0006]

The journal collecting method of the present invention comprises means for deciding a storage location of a journal record when a record in a file device is changed, means for managing a journal file for storing the journal record, and And means for retrieving a journal record corresponding to a changed record when data destruction occurs in the file device.

[0007]

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

FIG. 1 is a flow chart showing an embodiment of the present invention, and FIG. 2 is an explanatory diagram showing an example of operation.

From the program 21 to the data file device 2
When the record reference request for 5 is issued, the record is not changed for the data file device 25, so no journal management is performed.

From the program 21 to the data file device 2
When a record change request for 5 is issued, the buffer management 11 determines whether processing can be performed in the buffer 23 or whether the record of the data file device 25 is actually updated. As a result, if the buffer can be processed, the journal management does nothing, but if the buffer cannot be processed and the actual update is performed, the journal management determines the storage location of the pre-update journal (12), and the input / output is performed. The management is instructed to output the pre-update journal and change the record in the file device 25 (13, 14). In this state, the buffer management can release the block 23 in the buffer 22.

From the program 21 to the data file device 2
When the instruction to complete the record change request to 5 is issued, the journal management decides the place to store the updated journals of all the records changed by the program (15), and outputs the updated journals to the I / O management. , After instructing to change the record in the file device 25,
The journal before update is released (16, 17, 18).

FIG. 3 is an explanatory diagram showing an example of a journal management method.

Data record 3 in the data file 33
If the data record 38 is actually updated in the state where the change is not completed when changing 8, the journal before the update is required. Also, after the data record change is completed, the updated journal is required.

Therefore, in the journal file 32, the post-update journal 35 is written from the other side, and the pre-update journal 36 is written from the rear side. The journal management manages the storage range of the block 34 of the journal file by the post-update journal number / pre-update journal number in the journal management table 31 with the journal number added to each journal file 32.

The journal record 37 in the journal file 32 and the data record 38 in the data file 33 indicate whether the journal block in the journal management table 31 is valid or not, the logical number of the journal block, and the journal. The physical address of the block stored in the file is associated with the journal information in the data record 38. This determines the journal record storage location and retrieves the journal record corresponding to the changed data record.

By these journal management, even if the data in the file device is destroyed due to a program error, device error, etc. in any state while changing the data record in the data file, the data from the journal record before or after the update is updated. Can be restored.

[0017]

As described above, according to the present invention, since only the contents of the changed final record are collected in the journal, it is possible to reduce the capacity of the updated journal and reduce the file resource. be able to.
Further, since the capacity of the updated journal can be reduced, the data recovery processing time from the updated journal can be shortened.

[Brief description of drawings]

FIG. 1 is a flow chart showing an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing an example of operation.

FIG. 3 is an explanatory diagram showing an example of a journal management method.

[Explanation of symbols]

 11-18 Steps of processing 21 Program 22 Input / output buffer 23 Block 24 Journal file device 25 Data file device

Claims (1)

[Claims] 1. A means for determining a storage location of a journal record when a record in a file device is changed, a means for managing a journal file for storing a journal record, and a means for changing when a data corruption occurs in the file device. And a means for retrieving a journal record corresponding to the selected record.