JPS62210552A - File recovery system - Google Patents

Abstract

PURPOSE:To attain a file recovery process in a short time, by performing a file rollback process using a journal before updating. CONSTITUTION:A user process group 10, when accessing a user file group 20, informs an access to a file recovery controller 40, and the controller 40 accumulates the journal before updating at a journal file group before updating 30. On completing an operation by the process group 10, it is informed to a system failure deciding part 50, and it is decided whether a system program failure is generated or not in a completion process. When the failure is generated, a file recovery process 60 is generated by the deciding part 50, then a process to be recovered is indicated. The process 60 rollbacks the file group using the file group 30 regarding an indicated process.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a file recovery method, and in particular detects that a system program failure has occurred during process termination processing and generates a file recovery process. The present invention relates to a file recovery method in which a file recovery process performs incomplete file roll pack processing using only a pre-update journal without referring to a system control table.

(Prior art) Conventionally, when a system program failure occurs during process termination processing, in order to prevent the scope of the system failure from expanding, the subsequent processing that should be continued, including the seven-file roll-up process, is omitted, and the failure occurs. A known method is to improve the reliability of the entire system by separating the process in which the error occurred from the system, and to perform the omitted file recovery process using the updated journal.

(Problems to be Solved by the Invention) In the conventional file recovery method described above, while accessing the file to be recovered during file recovery processing,
This method has the disadvantage that file recovery processing takes a long time because it is necessary to stop all processes and use the updated journal to start recovery from the start of the business.

An object of the present invention is to isolate the faulty process from the system when a system program fault occurs during process termination processing, improve the reliability of the entire system, and prevent the newly created process from causing the system program fault. The object of the present invention is to provide a configuration and a recovery method that eliminates the above drawbacks by recovering files without referring to system control tables or process-specific information, and allows file recovery processing to be executed in a short time.

(Means for Solving the Problems) The file recovery method according to the present invention allows a group of user processes to access the same group of user files in a storage device, while maintaining a pre-update journal before updating a file in preparation for a process failure. The file is configured to be accumulated in a group of pre-update journal files via a file recovery controller, and when a deadlock occurs or a process terminates abnormally, the accumulated pre-update journal is used to roll pack the file. This is a file recovery method for a computer system, and is configured to include a first processing means and a second processing means.

The first processing means detects that a failure has occurred in a system program during process termination processing, generates a file recovery process, eliminates the failed process, and notifies the file recovery process of an identifier of the failed process. It is something.

The second processing means is that the generated file recovery process searches for a group of pre-update journal files based on the notified failed process identifier and creates a pre-update journal file for the failed process without referring to the system control table. This is for file cool pack processing using only the pre-update journal with a heading.

(Example) Next, the present invention will be explained with reference to the drawings.

FIG. 1 is a block diagram showing the concept of an embodiment of the file recovery method according to the present invention. In FIG. 1, IO is a user process group, 20fl is a user file group accessed by the user process group 10, 3G is a pre-update journal file group for storing pre-update journals, and 4G is a pre-update journal. A file recovery controller for accumulating in the file group 30 or performing callback processing for the user file group 20 using the pre-update journal; 50 indicates whether a system program failure occurred during the user process group IO or termination processing; A system failure determination unit 60 is generated by the system failure determination unit 50, and when the user process group IO accesses the user file group 20 using the pre-update journal, The file recovery controller 40 stores the unupdated journal in the updated journal file group 30. When the user process group IO is terminated, the system failure determination unit 50 is notified, and it is determined in the termination process whether or not a system program failure has occurred. When a system program failure occurs, the system failure determination unit 5G generates a file recovery process 60 and instructs the process to be recovered.

The file recovery process 60 is configured to use the pre-update journal file group 30 for the instructed process and roll-pax the user file group 20t.

FIG. 2 is a block diagram showing in detail an embodiment of the file recovery method according to the present invention, and parts equivalent to those in FIG. 1 are designated by the same reference numerals.

In FIG. 2, 13 is a user process, 21-23 are user files, 31-33 are new journals, 41 is a control table section,
42 is a recovery execution unit, O1d target process selection unit, 6
2 is a process state determination unit, 63 is a recovery execution unit, 64
is the Resource Liberation Department.

In Figure @2, when the user process 1l-13 in the user process group 1O accesses the user files 21 to 23 in the user file group 20, it sends a message to the recovery execution unit 42 in the file recovery controller 40. The recovery execution unit 42 notifies the user of the access, and stores the pre-update journals of the user files 21 to 230 in the pre-update journal files 31 to 33, and also registers the management information in the control table unit 41.

When a file roll pack request is sent from the user process 11-13, the recovery execution unit 42 in the file recovery controller 1O retrieves the control table unit 41 and the pre-update journals in the pre-update journal files 31-33. rollpack the user files 21 to 23 using

When the execution of the user processes 11-13 is completed, the system failure determining unit 50 is notified of the termination, and it is determined whether a system program failure has occurred in the process termination process. If a system program failure occurs,
The system failure determination unit 50 performs the file recovery process S
0, instructs the process to be recovered, and entrusts the file recovery/<IJ processing. Thereafter, user processes 11-13 are disconnected from the system.

The file recovery process 60 finds the pre-update journal file 31 used by the process instructed by the target process selection unit 61, for example, the process 11.

When the user process 11 should terminate normally or should terminate abnormally, the process status determination unit 62 determines whether a failure has occurred in the system program and starts the recovery execution unit 63. Determine whether or not.

When the process 11 should have terminated abnormally, the recovery execution unit 63 is activated and roll packs the user file 21t using the pre-update journal file 31 without referring to the control table unit 41. After that, Resource Release Department 6
Step 4 releases the system resources used by the user process 1'l and ends all processing. However, when the process 11 should be normal, the recovery execution unit 63 is not activated, and the resource release unit 64 releases the system resources used by the user process 11 and terminates all processing. .

Next, an overview of the operation will be explained with reference to the schematic flow shown in FIG.

When a user process accesses a user file,
The pre-update journal storage step 100 is executed in the pre-update journal file. When the user process is terminated, a system program failure determination step 110 is executed, and if there is no failure, a user process termination step l 20 iL is executed.

If a system program failure has occurred, a file recovery process generation step 130 is executed, and a user process deletion step 140 is executed. On the other hand, the generated file recovery process executes the step 160 of securing the journal file before update,
A step 160 of determining the termination status of the user process is executed.

If the user process should have terminated normally, execute step 12 (l) of releasing system resources and terminate.

If the user process is about to terminate abnormally, the user file role pack step 17 (1) is executed, the system resource release step 180 is executed, and the process ends.

(Effects of the Invention) As explained above, the present invention isolates the faulty process from the system when a system program fault occurs during process termination processing, improves the reliability of the entire system, and By recovering the file without referring to the system control table or process-specific information that caused the system program failure, the process reduces the possibility that the system program failure will occur again, and also reduces the possibility that the system program failure will occur again. The effect is that the recovery time can be shortened, since the amount of journals required for recovery is smaller than that of the journal before the update.

[Brief explanation of drawings]

FIG. m1 is a block diagram showing the concept of an embodiment for realizing the file recovery method according to the present invention. FIG. 2 is a block diagram showing in detail an embodiment of the file recovery method according to the present invention. FIG. 8 is a schematic flowchart showing an outline of the operation of FIG. 2. lO・S11 User process group 11A-13 @-User process 20...User file group 21-23...User file 30...Pre-update journal group 31-33...Pre-update journal 40.・File recovery controller 41@・Cloud control table unit 42 ・Recovery execution unit 50 ・Dispute/System failure determination unit 60 ・File recovery process 61 ・Target process selection unit 62 ・Fist/Process status determination Section 63...Recovery execution section 64...Resource release section

Claims (1)

[Claims] While a user process group accesses the same user file group in a storage device, a pre-update journal before updating the file is accumulated in a pre-update journal file group via a file recovery controller in preparation for a process failure; A file recovery method for a computer system configured to roll back a file using the accumulated pre-update journal when a deadlock occurs or a process terminates abnormally. a first processing means for detecting the occurrence of a failure, generating a file recovery process, extinguishing the failure process, and notifying the file recovery process of an identifier of the failure process; and the generated file recovery process; searches the pre-update journal files based on the notified failed process identifier without referring to the system control table, finds the pre-update journal file for the failed process, and uses only the pre-update journal. and a second processing means for performing the file roll pack processing.