JPH0756791A - Method for backing up and recovering file in distributed system - Google Patents

Abstract

PURPOSE:To provide a method for preparing a backup magnetic tape for a daily work file and a method for restoring destruction or the like generated in the work file by means of the magnetic tape in a distributed system for integrating many data recording media in a recording medium. CONSTITUTION:The system is provided with a management table 20 for storing information such as status, file numbers, processing dates, and magnetic tape numbers in order to manage the working states of respective data files, and at the end of work, required contents in a working file 21 are recorded in a data file MT 23 by a backup process and then the contents of the table 20 are recorded and stored in a system file MT 24. In the case of recovering destructed data, a recovery process is started, a recovery information table is prepared from the information stored in the MT 24 and the data of a specified header in the backup MT 23 which corresponds to a number specified by the information of the prepared table are successively inputted to the file 21 to recover them.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention creates a backup magnetic tape for daily processing files when a plurality of media recording a large number of data are integrated in a predetermined order and output to one storage medium. Method and when the data on the storage medium is lost due to an accident such as the destruction of the working disk,
Data recovery work using backup magnetic tape
A quick and easy recovery method.

[0002]

2. Description of the Related Art A large number of individual information such as a huge number of documents submitted every day are accumulated, and processed in a predetermined order.
For example, in order to increase the processing speed when outputting to one recording medium such as a magnetic tape in the order of submission date and submission number, a distributed processing system consisting of multiple terminals connected by a local area network (LAN) is used. In some cases, processing steps may be performed in parallel.

Each terminal reads a plurality of media and inputs them to a common work magnetic disk file, or calls data from the file to perform processing, and the processed work results are again used as work magnetic disk files. Returned to. Next, after the processing work is completed, the data recorded by the information processing device is read in a predetermined order for these many magnetic disk files for work,
Output to one magnetic tape and create an integrated data file.

In such a distributed system, the progress of work by each terminal, in other words, the processing state of individual information is managed, and it is confirmed that the processing of individual information to be output is proceeding in a predetermined order. Need to guarantee. Otherwise,
Even if a small amount of data processing is delayed, the integration work is not completed and the overall progress of the work is adversely affected. Therefore, it is necessary to manage so that the data processing of the integration target ends as soon as possible.

Along with the progress management of this parallel work, an important factor is the processing when the file of the work medium in which the individual information is recorded or the integrated file being written is destroyed due to some reason. It is desirable to eliminate such a situation, but in reality, the file of the working medium is destroyed due to the cause of failure of the magnetic disk as hardware, power failure accident, human error, etc. It is possible for some or all to be lost.

The loss of data due to the destruction of the disk or the like causes a serious result because the file integration work is delayed even if it is a part of the data. Therefore, it is necessary to restore the state before the destruction as quickly as possible. is there. In preparation for such a situation, a magnetic tape that copies the contents of a working magnetic disk is created and used for data recovery.It is called a backup magnetic tape, and data recovery is called file recovery. To call.

As an example, it is assumed that data updating work for various files is being executed. In order to deal with the destruction accident that occurred in such a case, information about a plurality of related work files that were being worked until then is acquired, and it is determined whether or not each work file needs to be restored.
A procedure is required to continue the update process after returning to the state before destruction. However, when a large number of work files are being updated in parallel, it is necessary to check the history for each work file one by one to determine from which part the recovery work is to be started. Operation is required.

[0008] When the amount of data is large and the number of work files to be processed is large, the amount of data required for checking alone is enormous. Even after the check is completed, it is necessary to confirm the work file required for the recovery work to execute the processing operation to recover the data and execute the processing promptly. The situation was confusing due to the unexpected situation, and it took time to make a decision.

[0009]

SUMMARY OF THE INVENTION The present invention uses a distributed system in which a large number of terminals are connected to perform parallel processing of processing operations such as inputting, editing, and updating for enormous amounts of media data, When creating an integrated tape, a method of creating a backup file in preparation for recovery in the event of a system work file corruption, and in the event of damage, use the recovery process and backup file to quickly create a work file. It is an object of the present invention to provide a recovery method that restores a state before the destruction occurs and enables recovery processing in a short time.

[0010]

In order to solve the above problems, according to the present invention, (1) data of a medium in which a large number of data are recorded is targeted, and data is edited, updated, and coded in a predetermined order. In order to execute the processing steps such as conversion, the recording medium is processed in parallel by a plurality of terminals, and a plurality of processed data are integrated and output to one storage medium in a distributed system. The process of executing the update process while exchanging the conflict due to the access from the process, and the management table having the file number, the status recording area, the file extraction date recording area, and the output destination MT number recording area are provided, and the daily processing is performed. At the stage of recording required information about the file in the management table and at the time of file backup, the status is detected and the backup MT A file backup method comprising a step of checking whether or not it has been written and selectively writing the latest data file or update file associated with the same file to a backup file, and (2) data editing and updating, When the data on the storage medium is lost due to an accident such as the destruction of the working disk when executing the processing steps such as code conversion, the step of starting the file recovery process and the system recorded on the backup magnetic tape of the extraction date. The steps of reading the file and generating a table of recovery information files, and specifying the date to be recovered and writing the file name of the recovery target file and the backup magnetic tape number in the specified area of the table. , The backup magnetic tape number recorded in the table Read the information header, compare the magnetic tape number and the file name with the file backup date and the backup magnetic tape number and the file name that are flagged as the recovery target, and read the matched file And writing to the work disk by using the recovery method.

[0011]

The present invention employs the above method to systematically create a backup file magnetic tape as part of daily work. Even when the work file is destroyed, the file recovery process is started, the number of the backup file required for recovery is displayed in the recovery table using the contents of the operation management table, and based on the instruction, By reading the contents of the required magnetic tape into the work file and executing the procedure for returning to the state before the occurrence of the destruction, the recovery processing is performed in a short time and efficiently.

[0012]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a diagram for explaining the role of a management table in an update processing process in an embodiment of the present invention, in which 11 is a business process, 12 is a management process, 13 is a data file, and 14 is a management. It's a table. here,
The business process 11 is assumed to execute data update processing. In this case, since data is changed during processing, access from other business processes causes contention, which is not preferable. Therefore, the management process 12 is requested by the route from the business process 11 to perform exclusive processing of the processing target data.

While performing the exclusive processing, the business process 11 executes the update processing of the files 2 and 3 recorded in the file 13 of the data to be processed by the route of. When this process ends, the exclusive process ends from the business process 11 to the management process 12 by the route of. The management process 12, which has received the completion of the exclusive processing from the business process, changes the status of the file in the management table 14 via the route.

Before the management table 14 was changed, the status 141 was OFF for all files, indicating that there was no file update. But,
Since the business process 11 has updated the contents for file numbers 2 and 3, the statuses of 2 and 3 are OFF.
Is changed to ON, which is a flag meaning "updated". By referring to this flag and the processing date, even if a plurality of files having the same file name are accumulated, it is possible to trace the history of when the data was updated.

FIG. 2 is a diagram for explaining the principle of the backup method of the present invention, in which 20 is an operation management table and 21
Is a data file group, 22 is a processing process, 23 is a data file magnetic tape, and 24 is a system file magnetic tape. The data file group 21 is a mass storage device 211 such as a magnetic disk, and is composed of a large number of files according to the required capacity.

Although not shown, individual data stored in a relatively small-capacity storage medium such as a flexible disk has already been read by another input terminal process and transferred to the data file group 21. It is assumed that it has been stored. These data are stored in the storage medium with one unit of data corresponding to the type as one unit, and even when transferred to a data file, a file number is assigned to each data for identification.

In the operation management table 20, the column of file number 201 is set, and the file number for identifying the file to be processed is recorded. Corresponding to the file number, there are provided storage areas of status 202 for displaying the change status of the file, extraction date 203, and MT number 204.

The contents of the operation management table 20 are read into the processing process 22 by the route shown by in FIG. As a mark for recognizing the file to be backed up, the file number for which the status flag of ON is set in the status 202 is extracted, the data of that number is searched from the data file group 21, and the route of To read the extraction target file into the process 22.

Next, the data of the read file is processed for recording and sent to a magnetic tape device (not shown),
Through the route in the figure, the information header recording the file name and the extraction date and the data are sequentially output to the data file MT23 and written. The update process is executed by such a procedure, and the predetermined data stored in the data file group 21 is recorded as backup data in the data file MT23 to which the specific magnetic tape number is assigned.

When the file backup is completed and the data file MT23 is created, it is no longer necessary to read the updated files in duplicate, and if duplicate recording files are created, an error occurs. Status 2 of the operation management table 20 in the route of
Change the status flag in the 02 column to OFF, set the MT extraction date and MT number to the date of the extraction work,
The operation management table 20 is updated.

Next, the file backup processing is executed, the contents of the operation management table 20 have been changed, and since a part of the information of the operation management table is used at the time of restoration work, the contents are stored in the system file. It is necessary to record on 24 and make a backup. Therefore, the processing process records the contents of the operation management table 20 in the system file 24 by the route of and executes the system backup processing. When the creation of both the data file MT23 and the system file 24 is completed, the file backup process ends.

FIG. 3 shows the creation of a file in the present invention.
It is a figure explaining the relationship of the update process and the backup file MT created by the file backup method by this invention. In FIG. 3, reference numeral 31 is a work date column, and illustrates the month / day when the processing work was executed. In this example, on September 1, the work files are created and stored with the file names of the files F1 and F2 as indicated by reference numeral 32.

After the processing operation on this day is completed, the created data of F1 and F2, which are the contents of these data files, are written on the magnetic tape indicated by reference numeral 33, as indicated by the broken line in the figure, and called MT-01. Make a backup tape of number. This is the data file M as the backup tape described in FIG. 1 and FIG.
Corresponds to T23. In this case, the extraction date is 9 of the work date.
It will be the 1st of the month. Similarly, on September 2, a file of F3 is created and the contents of F3 are written on the magnetic tape MT-02 to create a backup tape.

On September 3, the update work of the file F1 is executed and the contents are written to the magnetic tape MT-03 to create a backup tape. Here, note that the magnetic tape MT-01 and the magnetic tape MT-
In 03, the file with the file name F1 is recorded. However, even if a file with the same file name F1 is recorded in MT-01 and MT-03, the content of F1 of MT-03 has been updated and the data has been changed. It is different from F1 of the contents of -01. In order to identify this, the data of the work date (extraction date) together with the magnetic tape number are important keys.

In the example of FIG. 3, file F on September 4th
The relationship between the data F2 of the magnetic tape MT-04 for the update data of No. 2 and the F2 of MT-01, the data F3 of the magnetic tape MT-05 for the update data of the file F3 on September 5, and the F3 of MT-02. Has the same relationship, and has the same file name F2 or F3, but the contents are different. Therefore, as described above, the operation management table is used to record and manage the data regarding the update and the like, and the data is written in the information header of the magnetic tape and used for identification.

Referring to the contents of the system file dated September 5, from the file creation and update status of FIG. 3,
F1 created on 9/1 has been updated to 9/3 to MT-03,
F2 updated to 9/4 to MT-04, F3 to 9
It is updated to / 5 and recorded on MT-05 respectively, and as a result, files F1 to F4 are MT-03,
It can be seen that the data is recorded on the magnetic tapes 4 and 5. Thus, when the same file name exists, the updated file can ignore the previous file, recognize the valid file to be adopted, and create the table.

Next, a file recovery method which is the second function of the present invention will be described. On one working day,
If a file breaks down in the middle of a job and it becomes necessary to recover it, the work done between the morning of the day and the failure cannot be recovered, but the backup tape According to the procedure of the present invention, the file can be restored to the last state of the file on which the work has been normally performed, that is, the state at the start of work on the morning of the day. Various file recovery methods have been provided in the past, but if data is updated frequently and frequently, a large number of backup files will be created. Requires a very complicated processing procedure.

In the present invention, the extraction date, the file name and the MT name of the information header recorded on the data MT are compared with the date and the MT name on the recovery table generated by the recovery process so that they match. Perform recovery work only when you do so. For example, when the file name F1 is restored, there are two data MTs MT-01 and MT-03. When the data is read from MT-01, it does not match the extraction date and MT name on the recovery table, so it is not a recovery target. By performing the above checks, erroneous recovery of data files is avoided.

FIG. 4 is a conceptual diagram for explaining the file recovery method of the present invention. As the first step, the recovery process 41 is started, and among the backup steps described above,
System file MT24 created by the process of FIG.
Is read into the system as in, and information necessary for recovery such as the file name, extraction date, and magnetic tape number recorded in the management table is written in the work file 42 of the process.

In the second step, the extraction date, that is, the date to be restored is designated, and the operation management table read from the system file MT is searched as shown by the route of, and the date of the designated restoration date is set. The system file that has been read is read and the recovery information file 43 is created. Here, the specified extraction date is usually the previous day when the work is continuously performed every day, but when the previous day is a holiday or a day when the work is not performed, , The date when the latest normal file was created.

Next, as a third stage, the recovery process 41
As shown by the route of, the recovery recovery information file 43 is referenced and the data file MT23 having the same date as the extraction date is read. As shown in FIG. 4, the format of the contents of this MT is such that a file name and an extraction date are recorded as an information header, and this information header and data are written in pairs. The contents of the information header are read at the root of, and in the comparison process indicated by the reference numeral 44, when the contents are compared with the contents of the recovery recovery information file 43, the files belonging to the information header are Write under 21 specific directories.

In the fourth stage, when the writing of data is completed, as shown by the route of, the ON flag indicating that the restoration is completed is written in the area of the restoration flag provided in the recovery restoration information file 43. If there is a blank in the area of the recovery flag, it means that there is a file name that needs to be recovered. Therefore, referring to the magnetic tape number of the data file MT23 from the file name recorded there, the specified MT Are read and processed in the same procedure as described above. In this way, as indicated by the arrow in the column of the recovery flag in FIG. 4, when the work is repeated and the areas of the recovery flag in FIG. Is completed.

[0033]

INDUSTRIAL APPLICABILITY The present invention makes it possible to systematically create a backup file magnetic tape as part of daily work in an information processing system in which a vast amount of processing work is integrated by parallel processing. Further, even when the working file is destroyed, the file recovery process is started, the backup MT is sequentially read according to the recovery recovery information file generated from the system file MT, and the file is returned to the state before the destruction. Since the procedure is executed, it is possible to perform recovery processing efficiently in a short time.

[Brief description of drawings]

FIG. 1 is a diagram illustrating a role of a management table in an update processing process.

FIG. 2 is a diagram illustrating the principle of the backup method of the present invention.

FIG. 3 is a diagram illustrating a relationship between a file creation / update process and a backup file according to the file backup method of the present invention.

FIG. 4 is a conceptual diagram illustrating a file recovery method of the present invention.

[Explanation of symbols]

 13 data file 31 work date column 32 file 33 magnetic tape 41 recovery process 42 work file

Claims (2)

[Claims] 1. Data of a medium in which a large number of data is recorded is targeted, and data is edited, updated, and recorded in a predetermined order.
In order to execute a processing step such as code conversion, a recording medium is processed in parallel by a plurality of terminals, and a plurality of processed data are integrated and output to one storage medium in a distributed system. By providing a management table that includes a stage for executing update processing while exclusive processing of conflict due to access from a business process and a file number, a status recording area, a file extraction date recording area, and an output destination MT number recording area, At the stage of recording necessary information about the processing file in the management table and at the time of file backup, the status is detected and the backup M is already executed.
A file backup method comprising a step of checking whether data is written in T and selectively writing the latest data file or update file of the same file to a backup file. 2. The data of a medium in which a large number of data are recorded is targeted, and the data is edited, updated, according to a predetermined order.
In order to execute the processing steps such as code conversion, the recording medium is processed in parallel by a plurality of terminals, and in a distributed system in which a plurality of processed data are integrated and output to one storage medium, When the data on the storage medium is lost due to an accident such as the destruction of the working disk when executing the processing steps such as editing, updating, and code conversion, the file recovery process is started and the backup magnetic tape on the extraction date is written. The stage of generating the table of recovery information file by reading the recorded system file and the date to be recovered are specified. And the stage of writing
The backup magnetic tape number recorded in the table is selected, the information header is read, and the magnetic tape number and the file name are flagged with the recovery target as the backup date and the backup magnetic tape number and the file name. A step of comparing, a step of reading out the matched file and writing it to a working disk,
Recovery method with.