JP2526726B2 - Multiplexed file recovery method - Google Patents

Description

The present invention relates to a multiplexed file recovery system, and more particularly to a multiplexed file recovery system for a primary file and one or more secondary files used in a host computer. .

[Conventional technology]

Conventionally, files used by a single host computer have been multiplexed for the purpose of improving reliability, and business can be continued even if one file becomes an obstacle.

The following is an example of a conventional multiplexed file recovery method.

Multiplex the files and prepare a sub-file with the same contents as the normal file.

The memory in the host computer has failure information (information about a portion where input / output can be normally performed with respect to the primary system file and the secondary system file and a portion where input / output cannot be performed due to some failure).

At the time of executing input / output, the file to be accessed is determined based on this failure information. Then, when referring to data, input data from one normal or subordinate file and output data to all normal or subordinate files when updating data. To do.

When a failure occurs, the failure information on the memory is updated to reflect the failure content. (After that, I / O is no longer performed for the part where the error occurred.) At this time, I / O is executed for other files (primary file or secondary file) where the error has not occurred, and the work is continued. To do.

Then, there are the following two methods as a recovery method when the above-mentioned failure is a failure of a part of the file.

(1) The work is temporarily stopped, the input / output to / from the file is stopped, the alternative area is allocated in the external storage device, the data is copied from the normal file, and then the work is restarted.

(2) Continuing the work, allocating an alternative area in the external storage device, copying the data from the normal file, and at the same time updating the failure information about the copied portion.

[Problems to be Solved by the Invention]

The conventional multiplexed file recovery method described above has the following problems when applied to a file shared by a plurality of host computers.

The latter (2) restoration method can be realized by a single host computer system because the failure information is held in memory, but restoration is possible for files shared by multiple host computers. There is a drawback in that there is a possibility that the reflection of the failure information may differ between the host computers during operation, resulting in data inconsistency. (For example, in one host computer, the normal system file is considered to be a failure and data is written only to the secondary system file.In another host computer, the recovery operation is considered complete, and the data is written from the normal system file. Read, for that,
Data updates on a host computer that is regarded as an error in the primary file may not be reflected in the operations of other host computers. ) On the other hand, although the conventional multiplex file recovery method in which the failure information is held in the external storage device that can be shared between the host computers does not cause data inconsistency, the failure information in the external storage device is rewritten each time input / output is performed. Will be required,
It has a drawback that performance is extremely deteriorated. (Generally, it takes several thousand times more time to refer to an external storage device than to refer to a memory.) In the former method (1), recovery is possible, but work is suspended once. It has the disadvantage that it is necessary and cannot be applied to work such as 24-hour operation.

[Means for solving the problem]

A multiplexed file recovery method of the present invention is a multiplexed file recovery method of an information processing system having a plurality of host computers and an external storage device shared by the plurality of host computers, wherein A host system file that stores various data, and one or more subsystem files that have the same contents in the system file, and the host computer stores data stored in the external storage device. And a file access unit that is started by the business program and starts access to the primary file and the secondary file, and is started by the file access unit when data is referenced. Data from the normal one of the normal file or the secondary file Input / output means for inputting and outputting data to all normal files of the normal system file and the secondary system file when updating the data; and the normal system file and the secondary system file when referred to by the input / output device. Fault information storage means for returning information indicating whether a normal input / output is possible or an input / output is impossible due to some failure, and communication for transmitting / receiving information between the host computers Means and, when a failure occurs during input / output to / from the external storage device, is started by the input / output means, to the other host computer through the communication means, to the failure occurrence part in the external storage device. Input / output partial stop request means for requesting temporary stop of input / output, and the external storage by all the host computers After stopping the I / O to the faulty part of the device,
After being started by the input / output means, the information of the newly failed external storage device is added to the failure information storage means of each host computer with respect to the other host computers through the communication means, Input / output restart requesting means for requesting restart of input / output, and input / output to / from a corresponding part of the external storage device temporarily in response to an input / output stop request from the other host computer through the communication means. In addition to the input / output temporary stop / resume means for stopping and adding the information of the corresponding part of the external storage device to the failure information storage means in response to the subsequent input / output restart request, Continuation of business processing even in the event of a failure for the multiplexed file of the primary file and one or more secondary files shared between the host computers In order to make it possible, an alternative area allocation means for allocating an alternative area from the external storage device when a failure occurs in a part of one of the files, and first, the input / output partial stop means. It is started to temporarily stop the input / output to / from the other part from the host computer, and secondly, the input / output unit is started to activate one normal file in the area allocated by the alternative area allocation means. Data is copied, thirdly, the failure information for the relevant part is deleted from the failure information stored in the failure information storage means, and fourthly, the input / output resumption requesting means is activated to activate the other host computer. File partial recovery means for deleting the failure information for the corresponding part from the above failure information storage means and restarting input / output, and the corresponding file when an overall failure of any file occurs And a whole file restoring means for sequentially restoring the corresponding file by activating the alternative area allocation means and the file partial restoring means for each of the divided portions. Even if an overall file failure occurs, the entire failed file can be dynamically restored without interrupting the continuation of operations and with almost no impact on operations during recovery to ensure normal operation. It is composed by returning.

〔Example〕

Next, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of a multiplexed file restoration system of the present invention.

Referring to FIG. 1, in this embodiment, the host computers 1 and 2, the external storage device 3, the primary file 4, the secondary file 5, the communication means 6, the business programs 11 and 21, the file access means 12 and 22, Fault information storage means 13, 23, input / output means
14, 24, input / output partial stop requesting means 15, 25, input / output restart requesting means 16, 26, input / output pause restarting means 17, 27, file partial restoring means 18, 28, alternative area allocating means 19, 28, file It is composed of total recovery means 20, 30.

Next, the operation of this embodiment will be described with reference to the drawings.

The host computer 1 and the host computer 2 are
The external storage device 3 is shared, and the external storage device 3 is multiplexed with the primary file 4 and the secondary file 5.

Also, the host computer 1 and the host computer 2
And are connected by the communication means 6.

Since the operation in each host computer 1 and 2 is the same, the operation of the business in the host computer 1 will be described below.

A. Normal operation: The file is multiplexed in the primary file 4 and the secondary file 5 in the external storage device 3, and the failure information is held in the memory in the host computer 1.

The business program 11 requests the file access means 12 to access (reference or update) the record in the file.

The file access means 12 determines the storage position in the file of the record requested by the business program 11,
The input / output unit 14 is requested to refer to or update the block including the request record.

The input / output unit 14 refers to the fault information stored in the fault information storage unit 13, and whether the block requested by the file access unit 12 can be normally accessed in the primary file 4 and the secondary file 5 , Judge whether the access is impossible due to a failure.

The input / output unit 14 reads a block from one of the normal files in the case of a reference request (does not read from the file in which a failure has occurred), and in the case of an update request, blocks in all of the normal files. Write (do not write to the failing file).

B. Operation at the time of failure: When a failure occurs during input / output to / from the external storage device 3, the input / output unit 14 adds information on the newly failed part to the failure information storage unit 13, and Stop request means
15 is activated to request the other host computer 2 through the communication means 6 to temporarily stop the input / output to / from the faulty part in the file.

When the input / output from all the host computers 1 and 2 to the faulty part in the file is stopped, the input / output means 14 activates the input / output restart requesting means 16 and the communication means 6
Through, requesting another host computer 2 to restart the input / output after adding the information of the newly failed portion to the failure information storage means 23.

The input / output suspend / resume means 27 temporarily suspends the input / output to / from the corresponding portion in response to the input / output stop request from the other host computer 1 through the communication means 6, and responds to the input / output restart request to stop Fault information is added to the fault information storage means 23, and input / output is restarted.

After that, the operation returns to normal.

C. Recovery Operation When File Partially Fails: When a part of the file fails, the alternative area allocating means 19 allocates an alternative area from the external storage device 3.

The file partial recovery means 18 is an input / output partial stop means 15.
Is started to temporarily stop the input / output from the other host computer 2 to the relevant part.

Next, the file partial restoration means 18 activates the input / output means 14 to copy the data from the normal file to the area allocated by the alternative area allocation means 19.

Next, the file partial restoration means 18 deletes the failure information for the relevant portion from the failure information stored in the failure information storage means 13.

Next, the file partial recovery means 18 activates the input / output resumption requesting means 16, deletes the failure information for the relevant part from the failure information on the other host computer 2, and restarts the input / output.

After that, it is possible to continue the work in a normal state for both the primary file and the secondary file.

D. Recovery operation at the time of failure of the entire file: When the failure of the entire file occurs, the entire file recovery means 20 restores some of the files to such an extent that the time required for partial recovery of the file does not affect the business. Divide into parts.

Next, the whole file restoring means 20 includes an alternative area allocating means 19 and a file partial restoring means for each of the divided parts.
Start 18 and recover files sequentially.

〔The invention's effect〕

As described above, the multiplexed file recovery method of the present invention multiplexes files shared among a plurality of host computers to improve reliability, and when an overall file failure occurs, The relevant file is divided into several parts and the relevant files are restored in sequence so that the time required for partial restoration does not affect the work. It has an effect that it is possible to dynamically recover the entire failed file and return to normal operation with almost no influence of.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of a multiplexed file restoration system of the present invention. 1, 2 ... Host computer, 3 ... External storage device, 4 ... Primary file, 5 ... Subsystem file, 6 ... Communication means, 11 ... Business processing program, 12 ... File access means, 13 ... Fault information storage means, 14 ... input / output means, 15 ... input / output partial stop request means, 16 ... input / output restart request means, 17 ... input / output temporary stop / resume means, 18 ... file partial recovery means, 19 ... alternative area allocation means, 20 ... whole file recovery Means 21 ... Business processing program 22 ... File access means 23 ... Failure information storage means 24 ... Input / output means 25 ... Input / output partial stop request means 26 ... Input / output restart request means 27 ... Input / output temporary stop Resuming means, 28 ... File partial restoring means, 29 ... Alternative area allocating means, 30 ... Whole file restoring means.

Claims (1)

(57) [Claims] 1. A multiplexed file restoration method of an information processing system having a plurality of host computers and an external storage device shared by the plurality of host computers, wherein various data are stored in the external storage device. Normal file and one or more secondary files having the same contents as the normal file, and in the host computer, business processing is performed using data stored in the external storage device. And a file access unit that is started by the business program and starts access to the primary file and the secondary file; and a file access unit that is started by the file access unit and refers to the normal file or the Input data from one of the normal sub-files and update the data Is an input / output unit that outputs data to all normal files of the primary system file and the secondary system file, and normally refers to the primary system file and the secondary system file when referred to by the input / output system. Failure information storage means for returning information indicating whether the input / output is possible or the input / output is impossible due to some failure, communication means for transmitting / receiving information between the host computers, and the external storage When a failure occurs during input / output to / from a device, the input / output means is activated to temporarily input / output to / from another host computer through the communication means to the part in which the failure occurs in the external storage device. Input / output part stop requesting means for requesting stop, and a part where the failure occurs in the external storage device by all the host computers Of the external storage device newly activated in the fault information storage means of each host computer, which is activated by the input / output means after stopping input / output to / from After adding the information of the part, to the corresponding part of the external storage device in response to the input / output restart request means for requesting to restart the input / output, and the input / output stop request from the other host computer through the communication means. I / O temporary stop / resume means for temporarily stopping the input / output of the device, adding the information of the corresponding portion of the external storage device to the failure information storage means, and restarting the input / output in response to a subsequent input / output restart request. And a multiplexed file of one or more subsystem files and one or more subsystem files shared between the plurality of host computers. In order to enable the continuation of business processing even in the event of a failure, an alternative area allocation means for allocating an alternative area from the external storage device when a failure occurs in a part of any one of the files; , The input / output part stopping means is activated to temporarily stop the input / output to / from the other part from the host computer, and secondly, the input / output means is activated to allocate the alternative area allocating means. The data from one normal file is copied to the reserved area, thirdly, the failure information for the corresponding portion is deleted from the failure information stored in the failure information storage means, and fourthly, the input / output restart request means is provided. File partial recovery means for starting and deleting the failure information for the relevant part from the failure information storage means on the other host computer, and restarting input / output; In this case, the corresponding file is divided into several parts, and the alternate area allocation means and the file partial recovery means are activated for each of the divided parts, and an entire file recovery means for sequentially recovering the applicable file is provided. And if any file global failure occurs,
A multiplexed file recovery method that recovers the entire failed file and returns to normal operation without stopping the continuation of the operation and with little impact on the operation during recovery.