JP6044363B2 - Computer, NAS access method and NAS access program - Google Patents

Description

  The present invention relates to a computer system such as a transaction system using NAS.

  In general, a transaction system including a plurality of computers includes a NAS (abbreviation of “Network Attached Storage”, also referred to as “network-attached storage”) that can be accessed in each business process executed on each computer. ing.

  This NAS is used, for example, to exchange data between business processes. That is, a certain business process outputs data to be transferred to another business process to a file that the NAS has, and the other business process can acquire the data from the file.

  The NAS can be used as a shared disk shared by a plurality of computers, like a normal file server.

  In such a transaction system, in order to guarantee the atomicity and persistence of the transaction, there are rules as described below in the configuration of the directory in the NAS.

  That is, a directory in NAS that can be uniquely determined by a business process that operates a transaction that outputs data to NAS, a transaction that outputs the data, and a business process that operates a transaction that extracts and processes the data from NAS. The data output from one transaction must be output to the directory as at least one data file.

  Further, all data output from one transaction must be output to one NAS, and cannot be output across a plurality of NAS.

  Here, the transaction system executes a plurality of tasks in parallel in one business process, and a transaction is executed in each business.

  A directory structure satisfying such rules will be described with reference to FIG. FIG. 14 is a diagram showing a directory structure created by hierarchizing directories in the NAS.

  In NAS1 of FIG. 14, Process-A, Process-B, and Process-N indicate business process identification names of business processes that process data output to the NAS. Process-1, Process-2, and Process-n indicate business process identification names of business processes in which transactions that output data to NAS operate. Furthermore, TX1, TX2, and TXm indicate transaction identification names of transactions that output data to the NAS.

  That is, the operator or the system creates a directory directly under NAS 1 for each business process that inputs and processes data from the NAS, using the business process identification name of the business process in which the input processing transaction operates. .

  By creating a directory using such a configuration, the operator or system creates data that is processed by a transaction that operates in a certain business process from NAS, and transactions that operate in a plurality of other business processes simultaneously. Even when outputting to NAS, each transaction can be output to NAS in parallel without contention.

  Furthermore, since each transaction outputs data to each directory corresponding to each transaction, a plurality of transactions can be prevented from being output to the same directory at the same time, and each transaction can access a file in parallel. .

  FIG. 15 is a diagram showing a directory structure in which a directory is created by using a name using a delimiter without layering the directory in the NAS.

  Here, FIG. 14 and FIG. 15 describe NAS1 as an example, but when using a plurality of NAS, it is necessary to create a directory in each NAS as well.

  In addition, in order to indicate a state in which a transaction is being executed and output has not been completed for all data output in the transaction, the operator or system indicates that a directory in which data is being output is being output. It is created using a name (for example, TXm_temp in FIG. 14).

  As a result, it is possible to prevent the data input side from the NAS from referring to the data file that is in the process of being output.

  A transaction for outputting a data file to the NAS outputs the data file to a directory having a name indicating that the data file is being output. When the execution of the transaction is completed, the operator or the system changes the directory from the name indicating the middle of output to the normal name.

  In FIG. 14, the system outputs data to TXm_temp during transaction execution, and changes the directory name from TXm_temp to TXm when the transaction is completed.

  By changing the directory name during output and when output is completed, a business process that operates a transaction that outputs a data file to the NAS and a business process that operates a transaction that is input from the NAS are included in each corresponding data file. It becomes possible to access.

  When the number of processes running simultaneously increases or the amount of data increases with the increase in business and processing volume, the number and size of data files to be output increases, so the operator can add the NAS where the data is stored. Capacity and network connection speed must be expanded.

  For this reason, it is conceivable to add a new NAS or replace an existing NAS with a larger capacity NAS.

  By the way, in Patent Document 1, in a system that writes data from an active system to a shared disk, a standby system that recognizes the occurrence of a failure in the active system blocks a file path that accesses the shared disk from the active system. Discloses a method for preventing double writing.

  Patent Document 2 describes a method related to memory reduction when using a prefetch function of a database management system.

JP 2006-189963 A JP 2012-014739 A

  Computer systems that perform mission-critical tasks are required to remain undisturbed 24 hours a day, 365 days a year.

  When a new NAS is added and the output destination is changed to a new NAS, data that has been output to the NAS but has not been processed yet after being input from the NAS remains. Sometimes.

  Here, when a plurality of output destination NASs are not specified in the system definition information, the transaction recognizes or accesses both the existing output destination NAS and the new output destination NAS at the same time. I can't.

  Therefore, in order to process without losing data that has not yet been processed, does the operator stop the side that outputs data to the NAS and waits until the side that inputs all the data output to the NAS performs processing? Alternatively, after stopping the entire system, it is necessary to move a file that has been output during that time and has not been subjected to input processing to a new NAS.

  Alternatively, there is a method in which the operator designs the added NAS and the existing NAS so as to be seen as one file device by virtualization of the file system. However, with this method, the actual file storage location is determined by the file system, so the rule that data output in the same transaction must be stored in one NAS cannot be observed. .

  Further, it is conceivable to change the definition information specifying a plurality of directories as a data storage destination directory so that a transaction can simultaneously recognize and access a plurality of directories. In this case, it is necessary to consider the definition change execution timing so that no deviation occurs between computers or processes.

  Here, a case where inconvenience occurs due to a shift in the definition change execution timing will be described with reference to FIG. FIG. 16 is a diagram for explaining a problem when the definition is changed.

  The system shown in FIG. 16A includes a computer 1, a computer 2, and a computer 1 and NAS-1 accessed from the computer 2.

  In FIG. 16A, it is assumed that process 1 of computer 1 and process 2 of computer 2 are defined to use NAS-1. At this time, the data output to the NAS-1 by the process 1 of the computer 1 can be processed by the process 2 of the computer 2.

  Next, a case where NAS-2 is added to the definition in FIG. 16A will be described with reference to FIG.

  As shown in FIG. 16B, it is assumed that only the computer 1 is changed to the definition with NAS-2 added first. In this state, the data output to NAS-2 in process 1 of computer 1 cannot be processed by inputting NAS-2 data because process 2 of computer 2 cannot recognize NAS-2.

  As described above, when the data output side outputs data with a new definition before the input side, an error occurs on the input side where the new definition is not reflected.

  If there is only a change related to one output destination, it is not difficult to control the definition change timing. However, the control between the processes at the definition change timing for changing a plurality of output destinations at a time becomes complicated.

  Furthermore, if the process to be changed exists on multiple computers, if the change is to be made without stopping the system, the change must be divided into multiple times, which increases the complexity of managing definition information. End up.

  In such a situation, even if it is necessary to add or replace a NAS, it is possible to add or replace a NAS without stopping the transaction system or changing the definition. It is expected to provide a way to make it possible.

  Furthermore, as a second problem, a problem that occurs at the time of a failure will be described below.

  Consider a case where a partial failure occurs in the transaction system and data cannot be input from the NAS for processing. In this case, it is conceivable that a large amount of data stays in the NAS if operation is performed without stopping input from the outside during the period from the occurrence of the failure to the recovery.

  Further, at the time of failure recovery, new data is additionally output even while the accumulated data is being processed.

  In order to eliminate the accumulated data, the process of inputting data from the NAS must be performed faster than the output of new data to the NAS.

  When operating normally, the file information of the data output to the NAS is notified by communication between business processes.

  However, during failure occurrence, depending on the nature of the failure, notification may not be received, or the table storing the file information may be lost due to the restart of the business process, and the file information of the data output to the NAS may be lost. There is.

  If this information is lost, after recovery, the business process searches for the file in the NAS where the data is stored to obtain a list of files, and uses the obtained file list as file information to store the data. To process.

  When a file is output during the search after the recovery, the file search process includes the file output after the recovery in the search result. In addition, the file search process takes extra time due to an increase in search objects.

  In order to prevent double processing, it is necessary to exclude from the obtained search results the file information of data output after recovery from a recognized failure by notification between business processes.

  Patent Document 1 discloses only a method for preventing double writing to a shared disk when an active system failure occurs, and does not disclose any method for solving this problem.

  Patent Document 2 only discloses a method of reducing the memory used when using the look-ahead function of the database management system, and does not particularly disclose this problem.

  SUMMARY OF THE INVENTION Accordingly, it is a primary object of the present invention to provide a computer or the like that switches files stored in a NAS used in a computer system without stopping the computer system or changing a definition.

  In order to achieve the above object, a computer according to the present invention comprises the following arrangement.

That is, the computer according to the present invention is
When outputting data to the NAS shared by a plurality of computers, an output permission state operation means for enabling determination as to whether or not the state of the first directory that is the output destination of the data is output permission; A directory having a serial number shown is created in the plurality of computers, and a serial number generating means for storing the symbolic link to the data output destination first directory of the NAS in the second directory, and created in the computer File path generation means for generating a file path used for accessing the directory, file access means for accessing the NAS using the file path, and data input / output to / from the NAS by the file access means A processing means is provided.

  This object is also achieved by the NAS access method in the computer having the above configuration.

  According to the present invention described above, it is possible to switch the file in the NAS used in the computer system without stopping the computer system or changing the definition.

It is a block diagram which shows the structure of the computer which concerns on the 1st Embodiment of this invention. It is a hardware block diagram of the 1st Embodiment of this invention. It is a block diagram which shows the structure of the computer which concerns on the 2nd Embodiment of this invention. It is a system configuration | structure figure of the system which adds NAS in the 2nd Embodiment of this invention. It is a figure which shows the directory and symbolic link of each NAS (NAS600 and NAS601) and each computer (computer 200 and computer 300) in the 2nd Embodiment of this invention. It is a figure which shows the directory and symbolic link of each NAS (NAS600 and NAS601) and each computer (computer 200 and computer 300) in the 2nd Embodiment of this invention. It is a figure which shows the serial number management table 1921 in each business process (The business process 210, the business process 310, and the business process 320) in the 2nd Embodiment of this invention. It is a figure explaining the process of the serial number management part in the 2nd Embodiment of this invention. It is a system block diagram after adding NAS to the 2nd Embodiment of this invention. It is a flowchart which shows the process of the file output part in the 2nd Embodiment of this invention. It is a flowchart which shows the process of the file input part in the 2nd Embodiment of this invention. It is a block diagram which shows the structure of the computer which concerns on the 3rd Embodiment of this invention. It is a figure which shows each NAS (NAS600 and NAS601) and the directory and symbolic link of each computer (computer 200 and computer 300) in the 3rd Embodiment of this invention. It is a figure which shows each NAS (NAS600 and NAS601) and the directory and symbolic link of each computer (computer 200 and computer 300) in the 3rd Embodiment of this invention. It is a figure which shows the serial number management table 1921 in each business process (The business process 210, the business process 310, and the business process 320) in the 3rd Embodiment of this invention. It is a flowchart which shows the process of the file search part in the 3rd Embodiment of this invention. It is a figure which illustrates illustartively the hardware constitutions of the computer (information processing apparatus) which can implement | achieve the 1st, 2nd and 3rd embodiment of this invention. It is a figure which shows the directory structure produced by hierarchizing the directory in NAS. It is a figure which shows the directory structure which created the directory using the name which used the delimiter, without hierarchizing a directory in NAS. It is a figure explaining the subject at the time of a definition change.

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

<First Embodiment>
FIG. 1 is a block diagram showing a configuration of a computer according to the first embodiment of the present invention.

  The computer 1 according to this embodiment includes an output permission state operation unit 11, a serial number generation unit 12, a file path creation unit 13, a file access unit 14, and a processing unit 15.

  When outputting data to the NAS shared by a plurality of computers, the output permission state operation unit 11 makes it possible to determine whether or not the state of the first directory that is the output destination of the data is output permission.

  The sequence number generation unit 12 creates a second directory having a name including a sequence number indicating the output order in the plurality of computers, and stores a symbolic link to the first directory in the second directory.

  The file path creation unit 13 generates a file path used for accessing the second directory.

  The file access unit 14 accesses the NAS using the file path.

  The processing unit 15 processes data input / output for the NAS accessed using the file access means.

  FIG. 2 is a hardware configuration diagram of the first embodiment of the present invention. It is composed of at least two computers and at least one NAS.

  FIG. 2 includes a computer 8001, a computer 8002, a computer 8003, and a NAS 8004.

  In FIG. 1, the output permission state operation unit 11, the serial number generation unit 12, the file path creation unit 13, the file access unit 14, and the processing unit 15 are realized by a computer 8001, a computer 8002, and a computer 8003. .

  As described above, according to the first embodiment, it is possible to switch files in the NAS used in the computer system without stopping the computer system and without changing the definition. effective.

  The reason is that the computer system according to the present embodiment accesses a file stored in the NAS using a symbolic link created in a directory indicating a serial number and an output permission state.

<Second Embodiment>
Next, a second embodiment based on the computer 1 according to the first embodiment described above will be described.

  FIG. 3 is a block diagram showing a configuration of a computer according to the second embodiment of the present invention.

  In FIG. 3, the transaction system 7000 according to this embodiment includes at least one NAS 3000, a plurality of business processes 1000, one NAS control process 1950, an external system 5000, and definition information 2010.

  The NAS 3000 exchanges data between business processes.

  The definition information 2010 is various information of the entire system designated by the user, such as the association between the business process and the business, the NAS information including the file used in the business process, and the like.

  The definition information 2010 is taken into the business process 1000 by the transaction control unit 1010 and is referenced by each unit in the business process 1000.

  The business process 1000 includes a transaction control unit 1010, a business 1020, an external input unit 1030, a process communication unit 1040, a file information table 1041, a file input unit 1050, an input buffer 1051, a file output unit 1060, The output buffer 1061, an input file path generation unit 1900, an output file path generation unit 1910, a serial number management unit 1920, and a serial number management table 1921 are included.

In the business process 1000, a plurality of business operations 1020 can be executed in parallel.
In the business 1020, transactions are executed one by one, and are not executed in parallel.

  The transaction control unit 1010 controls the entire business process 1000 such as execution and re-execution of the business 1020 and database commit or rollback.

  In response to data input from the external system 5000, the external input unit 1030 notifies the transaction control unit 1010.

  Upon receiving the notification, the transaction control unit 1010 notifies the business 1020 of the start of the transaction.

  Here, the transaction represents a series of inseparable processes in the computer.

  The business 1020 performs processing such as referring to and updating a database or the like based on data input by the external input unit 1030.

  Further, when there is data to be transferred to another business 1020, the business 1020 outputs data corresponding to the other business 1020 to the output buffer 1061.

  The file output unit 1060 outputs the data output to the output buffer 1061 to a file (hereinafter referred to as a data file) included in the NAS 3000. The output file path generation unit 1910 generates a file path for accessing the data file.

  Here, the data output to the data file is in an unconfirmed state because the execution result of the transaction has not yet been confirmed. The transaction control unit 1010 performs commit in order to confirm the update of a database or the like (not shown) performed in the process when the process of the business 1020 is completed. As a result of the commit, the execution result of the transaction output to the data file is confirmed.

  When the transaction control unit 1010 succeeds in the commit, the process communication unit 1040 notifies the output data file information to other business processes. The data file information to be notified is, for example, information indicating where the data file is, such as the identifier of the transaction that output the data and the file path of the data file.

  On the other hand, the process communication unit 1040 that has received the data file information notifies that the received file information has been stored in the file information table 1041 and that the file information has been received by the transaction control unit 1010.

  Upon receiving the file information reception notification, the transaction control unit 1010 requests the file input unit 1050 to read data.

  The file input unit 1050 requested to read data reads data from the corresponding file in the NAS 3000 into the input buffer 1051 based on the file information stored in the file information table 1041. An input file path generation unit 1900 generates a file path for accessing the data file.

  After reading the data into the input buffer 1051, the transaction control unit 1010 notifies the business 1020 of the start of the transaction.

  The NAS control process 1950 includes an output permission state operation unit 1960 and a serial number generation unit 1970.

  The output permission state operation unit 1960 creates a directory in the NAS 3000 that can determine the output permission state based on the directory name, and controls whether the output permission state is output permission or output prohibition using the created directory. .

  Furthermore, the transaction system 7000 according to the present embodiment includes a serial number directory that stores a symbolic link to the NAS on a local disk in a computer on which the transaction system 7000 operates. This system adds a serial number generation unit 1970 that generates a serial number used as a part of the name of the serial number directory, a serial number management unit 1920 that manages the latest serial number, and a serial number management table 1921 to the business process 1000. , To solve the problems described in the “Background Technology” column.

  The serial number generation unit 1970 and the output permission state operation unit 1960 make the computer recognize the added NAS without changing the definition information. Further, the serial number generation unit 1970 and the output permission state operation unit 1960 give the business process file output unit 1060 and the file input unit 1050 an opportunity to switch the output destination to the added NAS.

  The file output unit 1060 of the business process recognizes and generates the path to the added NAS without depending on the definition information by the output path generation unit 1900, the serial number management unit 1920, and the serial number management table 1921, and sets the output destination. Can be switched.

The business process file input unit 1050 outputs the unprocessed data output to the existing NAS and the new NAS without stopping the system by the input path generation unit 1910, the serial number management unit 1920, and the serial number management table 1921. FIG. 4 is a system configuration diagram of a system to which a NAS is added according to the second embodiment of the present invention.

  A transaction system 900 shown in FIG. 4 includes an external system 100, a computer 200, a computer 300, a NAS 600, and a NAS control process 999.

  In FIG. 4, two computers and one NAS are shown, but the number of each is not limited.

  The computer 200 executes a business process 210 that controls the business 211.

  The computer 300 executes a business process 310 that controls the business 311 and a business process 320 that controls the business 321.

  The NAS 600 can be referred to and updated from the computer 200 and the computer 300.

  The transaction system 900 inputs data from the external system 100 and passes the input data to the business 211 operating in the computer 200.

  The business 211 determines whether the business 311 or the business 321 performs processing based on the contents of the input data.

  If the business to be processed is the business 311 based on the determined result, the business 211 creates a file in a directory corresponding to the business 311 in the NAS 600 and outputs data to the file. If the business to be processed is the business 321, the business 211 creates a file in a directory corresponding to the business 321 in the NAS 600 and outputs data to the file.

  FIG. 5A and FIG. 5B are diagrams showing directories and symbolic links between each NAS (NAS 600 and NAS 601) and each computer (computer 200 and computer 300) in the second embodiment of the present invention.

  FIG. 5C is a diagram showing a serial number management table 1921 in each business process (business process 210, business process 310, and business process 320) in the second embodiment of the present invention.

  The operator creates “/ export” as a public directory in the NAS 600.

  The operator creates a directory corresponding to the business process for extracting data from the NAS in the created “/ export” directory. That is, in the (a) line of FIG. 5A, directories are created using the names “A310” and “A320” as directories corresponding to the business process 310 and the business process 320, respectively.

  Furthermore, the operator creates a directory “vv” indicating that the output permission state is output permission in these two directories.

  The output permission state indicates whether file output to the directory is permitted, and includes two states: a state where file output is permitted and a state where file output is prohibited.

  The output permission status can be identified by the directory name.

  That is, the change from the state in which the file output is permitted to the state in which the file output is prohibited is performed by the output permission state operation unit 1960 changing the path of the output destination directory at the time of file output.

  Here, as an example, the name of the directory is “vv” when output is permitted and “ii” when output is prohibited.

  If the output permission status of the access target directory is output permission, the business process succeeds in file output.

  On the other hand, when the output permission status of the access target directory is prohibited, the business process fails to output the file.

  The file output unit 1060 of the business process recognizes that the output destination must be changed at the time of subsequent file output by detecting the error due to the access failure.

  A directory corresponding to the business process for outputting data to the NAS is created in the directory “vv”. Here, a directory is created using the name “A210” as a directory corresponding to the business process 210.

  Through a series of operations so far, the operator creates the “/ export / A310 / vv / A210” directory. A transaction operating in the business process 210 creates a directory for each transaction in the directory, and outputs data to be passed to the business process 310 to the created directory.

  Similarly, a transaction operating in the business process 210 creates a directory for each transaction in the “/ export / A320 / vv / A210” directory, and outputs data to be passed to the business process 320 to the created directory.

  The operator can access the files in the NAS 600 by mounting the public directory “/ export” in the NAS 600 to “/ nas1” using the mount command on the computer 200.

  In the computer 200, a serial number directory “X1” exists in the “/ data” directory of the local disk, and “/ data / X1 / A310” and “/ data / X1 / A320” exist in the serial number directory. These directories store symbolic links to the directory “/ export / A310” and “/ export / A320” in the NAS 600.

  For example, according to “/ data / X1 / A310-> / nas1 / A310” shown in line (a) of FIG. 5A, “/ data / X1 / A310” is a symbolic link, and “/ nas1 / A310” is changed to “/ nas1 / A310”. You can see that it points.

  By using a symbolic link, a transaction operating in the business process 210 executed on the computer 200 outputs data desired to be passed to the business process 310 by specifying a file in the “/ data / X1 / A310” directory. Actually, the file is output to a file in the “/ export / A310” directory in the NAS 600.

  The name of the serial number directory consists of a fixed name and a number that is a serial number.

  The serial number is used to simultaneously identify and access a plurality of directories storing data processed by one business process.

  The serial number is uniquely determined from the current value to the next value, and has regularity that does not overlap between the start of operation of the system and the stop of use.

  The serial number exists for each business process so that the storage location of the file can be changed for each business process. The serial number management unit 1920 manages the latest serial number in association with the business process by using the serial number management table 1921.

  The serial number management table 1921 exists for each business process. This is because the serial numbers constituting the paths of a plurality of files output from one transaction must all be the same, and therefore must not be affected by the update to the serial number management table 1921 by other business processes. .

  FIG. 6 is a diagram for explaining the processing of the serial number management unit in the second embodiment of the present invention.

  The operation of the serial number management unit 1920 that references and updates the serial number management table 1921 will be described with reference to FIG.

  The serial number management unit 1920 performs a serial number acquisition process and a serial number count process.

  In the serial number acquisition process, a serial number for a unique identifier such as the name of a business process that inputs and outputs data in the NAS is acquired from the serial number management table 1921 (S14).

  In the sequence number counting process, the number part of the sequence number acquired by the sequence number acquisition process is added (S12), and the sequence number management table 1921 is updated using the added sequence number (S13).

  The serial number directory X1 is created using the names of X2 and X3, with the numeric part incremented when the serial number is counted up.

  This serial number directory is created in each computer (computer 200 and computer 300) by the serial number generation unit 1970 of the NAS control process 1950 when a new NAS is added by the operator, and symbolic to the NAS is included in the serial number directory. Generate a link.

  Depending on the serial number directory, there are symbolic links to the storage destinations of data processed by the same business process in different directories, so that each computer can simultaneously recognize a plurality of data storage destinations. As a result, the business process can input unprocessed data output to the NAS that was the output destination before adding the NAS and output to the added NAS simultaneously or almost simultaneously.

  Further, since the serial number is a serial number, the next serial number can be inferred from the current serial number, and the business system can recognize the storage destination added to the business process without the definition information.

  Then, the business process counts up the serial number currently recognized by the business process from the serial number management table 1921 when the file access error is detected due to the change of the output permission status, and the serial number after the count-up is added to the file path By using it for generation, the added NAS or directory can be accessed.

  Further, since the generation of the symbolic link is performed in advance before the output permission state is changed, the problem of reflection deviation that occurs when the definition information is used as described in the problem does not occur.

  The business process 210 in the computer 200 starts a transaction related to the business 211 in response to data input from the external system 100. Thereafter, the business 211 distributes and outputs the data to the business 311 or the business 321.

  The business process 210 stores the output data as a file in the NAS 600.

  The file information is sent from the business process 210 to the business process 310 that controls the business 311 and the business process 320 that controls the business 321.

  When a business process outputs data to a file, it creates a directory for each transaction. For example, if the transaction identifier is “trns1”, directories “/ data / X1 / A310 / vv / A210 / trns1_temp” and “/ data / X1 / A320 / vv / A210 / trns1_temp” are created.

  These are actually directories “/ export / A310 / vv / A210 / trns1_temp” and “/ export / A320 / vv / A210 / trns1_temp” in the NAS 600. Files that store data are created in these directories.

  “_Temp” at the end means that output is in progress, and prevents the business processes 310 and 320 from erroneously reading the file before the completion of execution of the transaction is confirmed.

  When all the files with the transaction identifier “trns1” are output, the midway output directory “trns1_temp” is changed to “trns1” in order to make it a fixed directory.

  That is, “/ data / X1 / A310 / vv / A210 / trns1_temp” is changed to “/ data / X1 / A310 / vv / A210 / trns1”. Similarly, “/ data / X1 / A320 / vv / A210 / trns1_temp” is changed to “/ export / A320 / vv / A210 / trns1”.

  Next, the business process 310 reads a file stored in the NAS 600 according to the information notified from the business process 210. The path at this time is generated from the serial number directory “X1” at the time of storing the file included in the notified information, the transaction identifier “trns1”, and the file name, and “/ data / X1 / A310 / vv / A210 / trns1 / File name ".

  The business 311 performs processing based on the data read by the business process 310.

  In addition, the business process 320 and the business 321 operate in the same manner as the business process 310 and the business 311.

  Here, let us consider a case where it is assumed that the amount of processing in the business 311 and the business 321 increases due to the service modification and the capacity of the NAS 600 becomes a bottleneck in the entire system. In this case, the operator adds a new NAS 601 to expand the capacity of the entire NAS, and changes the storage destination of data processed by the job 321 from the NAS 600 to the NAS 601 using the expanded storage area. Disperse.

  FIG. 7 is a system configuration diagram after the NAS is added to the second embodiment of the present invention.

  With reference to FIG. 7, a process for adding the NAS 601 will be described.

  First, when making a change, the operator creates a directory in the NAS 601 and mounts the NAS 601 on each computer.

  As a result, the directories and symbolic links of each NAS and each computer are as shown in line (b) of FIG. 5A. Further, the serial number management table 1921 managed by each business process is as shown in the (b) line of FIG. 5C.

  The directory to be created corresponds to the directory for the business process 320 to be changed, and is “/ export / A320 / vv / A210” as in the NAS 600.

  In the computer 200 and the computer 300, the operator mounts the public directory “/ export” of the NAS 601 to “/ nas2” by using a mount command.

  At this point, the path used by each business process for file input / output is not changed.

  Next, the NAS control process 999 starts processing for changing the output destination in the business process 320.

  As shown in the (c) line of FIG. 5A, the serial number generation unit 1970 of the NAS control process 999 has the serial number directory name “/ data / X1 / A320” in which the data processed by the business process 320 is stored as “ X1 ", the number part of this name is counted up and a directory is created using the name" X2 ".

  Further, the serial number generation unit 1970 of the NAS control process 999 generates a symbolic link to the directory “/ export / A320” in the NAS 601 in the “X2” directory, as shown in the (c) line of FIG. 5A. To do. This is performed in each computer (computer 200 and computer 300).

  Next, the output permission state operation unit 1960 of the NAS control process 999 has a directory “/ export / A320” in which data to be processed by the business process 320 in the NAS 600 is stored, as shown in line (d) of FIG. 5A. The name of the directory in the output permission state is changed from the permission state “vv” to the prohibition state “ii”.

  Here, the operation at the time of file output to the NAS will be described with reference to FIG.

  FIG. 8 is a flowchart showing the processing of the file output unit in the second embodiment of the present invention.

  When the business process 210 outputs data to the business process 320 as a file, the file output unit 1060 first performs serial number acquisition processing by the serial number management unit 1920 to acquire the latest serial number corresponding to the business process that is the output destination (S21). ).

  Next, the file output unit 1060 specifies the acquired latest serial number and the output permission state, and the output file path generation unit 1910 generates the output file path “/ data / X1 / A320 / vv / A210 /”. (S22).

  The file output unit 1060 opens a file using the generated output file path (S23), and determines whether the file has been successfully opened (S24).

  As shown in the NAS 600 in the row (d) of FIG. 5A, the path of the directory in which the data processed by the business process 320 in the NAS 600 is stored is “/ export / A320 / ii / A210 /”. Opening the file fails because it is different from the path used to open the file.

  If the file fails to open (if “NO” in S24), it is the first file in the directory (that is, the first data output in the transaction to output data) (“YES” in S27) ), The file output unit 1060 performs the serial number counting process of the serial number management unit 1920 (S29).

  Here, the serial number management unit 1920 adds the serial numbers, changes the serial number directory from “X1” to “X2”, and updates the location corresponding to the business process 320 in the serial number management table 1921 of the business process 210 to “X2”. To do. As a result, the serial number management table 1921 becomes as shown in line (e) of FIG. 5C.

  Then, the file output unit 1060 designates the latest serial number after the serial number counting process and the output permission state, generates an output file path (S22), and performs the processes after S23.

  The regenerated file path is “/ data / X2 / A320 / vv / A210 /” from “X2” obtained as a serial number corresponding to the business process 320 from the serial number management table 1921 and “vv” in the output permission state. Therefore, the file output unit 1060 succeeds in opening the file.

  If the file is successfully opened, the file output unit 1060 writes data to the output file (S25) and closes the file (S26).

  If the file has failed to be opened (in the case of “NO” in S24) and is not the first file in the directory, the file output unit 1060 has already output data in the transaction to output data. Recognize Therefore, the file output unit 1060 designates the latest serial number and the output prohibition state, and after the output file path is generated by the output file path generation unit 1910 (S28), the processing after S23 is performed.

  In this way, the file output unit 1060 detects the added output destination without changing the definition information. In addition, files output in the same transaction can be stored only within one NAS, and the directory configuration rules can be observed.

  Next, file input from NAS will be described with reference to FIG.

  FIG. 9 is a flowchart showing the processing of the file input unit in the second embodiment of the present invention.

  The business process 320 tries to read the file according to the file information of the output data notified from the business process 210.

  First, the file input unit 1050 has the same serial number in the notified file information as the latest serial number corresponding to the business process 320 stored in the serial number management table 1921 of the business process 320 in the row (e) of FIG. 5C. It is determined whether it exists (S31).

  In the case of data output before the NAS control process 999 changes the output permission state, the output serial number included in the file information of the notified output data is “X1”, which is the same as the latest serial number. Yes (in the case of “YES” in S31). Therefore, the file input unit 1050 generates the input file path by the input file path generation unit 1900 with the latest serial number and the output permitted state (S32). The generated file path is “/ data / X1 / A320 / vv / A210 /”.

  Next, the file input unit 1050 opens the file using the generated input file path (S34), and determines whether the file has been successfully opened (S35).

  As shown in the NAS 600 in the row (e) of FIG. 5B, the path for accessing the directory storing the data processed by the business process 320 in the NAS 600 is “/ export / A320 / ii / A210 /”. Opening the file fails because it is different from the path used to open the file.

  Since the file open has failed (in the case of “NO” in S35), the file input unit 1050 performs the serial number counting process of the serial number management unit 1920 (S38).

  Here, the serial number management unit 1920 increments the serial number “X1” to “X2” and updates the part corresponding to the business process 320 in the serial number management table 1921 of the business process 320 to “X2”. As a result, the serial number management table 1921 becomes as shown in line (f) of FIG. 5C.

  Here, by counting up the serial number in the serial number management table 1921, the file input unit 1050 does not fail to access the file when the file is input next time. The file input unit 1050 can generate a correct file path from the beginning and succeed in file access using the generated file path.

  Then, the file input unit 1050 generates the input file path by designating the serial number before the serial number counting process and the output prohibited state (S33), and performs the processes after S34.

  Since the regenerated file path is “/ data / X1 / A320 / ii / A210 /” from the serial number “X1” before the serial number count processing and the output prohibited state “ii”, the file input unit 1050 , Succeeds in opening the file.

  When the file is successfully opened, the file input unit 1050 reads data from the file (S36) and closes the file (S37).

  For the data output after the NAS control process 999 changes the output permission state, the serial number at the time of output included in the file information of the notified output data is “X2”, which matches the latest serial number. (In the case of “NO” in S31). For this reason, the file input unit 1050 generates the input file path by the input file path generation unit 1900 using the notified serial number “X2” and the output permission state (S32), and then performs the processes after S34.

  In this way, the transaction system processes the unprocessed data output before the storage destination is added without stopping the system even when the NAS that is the data storage destination is added. At the same time, the data stored in the added storage destination can be processed.

  Further, also in other business processes, when normal file access becomes an error, the file access is re-executed and the serial number management table 1921 in each business process is updated.

  As a result, the serial number management table 1921 managed by each business process becomes as shown in the (g) line of FIG. 5C.

  Thus, the processing for adding the NAS 601 and then changing the storage destination of the data processed by the business process 320 from the NAS 600 to the NAS 601 is completed.

  Note that, after the processing on the data processed by the business process 320 in the NAS 600 is completed, the operator can use the directory “/ export / A320 / ii /” in which the data processed by the business process 320 in the NAS 600 is stored, The symbolic link to the storage location of the data processed by the business process 320 of the serial number X1 that is no longer used in each computer (computer 200 and computer 300) can be deleted.

When these deletions are performed, directories and symbolic links in the NAS 600 and each computer (computer 200 and computer 300) are as shown in line (h) of FIG. 5B. Further, the serial number management table 1921 of each business process (business process 210, business process 310, and business process 320) is as shown in line (h) of FIG. 5C.

  As described above, in the second embodiment, in a transaction system that uses NAS for transferring data between businesses, the NAS needs to be used in response to an increase in the necessary capacity of the NAS due to an increase in business and processing amount. Even when it becomes necessary to add or replace the NAS, it is possible to add or replace the NAS without stopping the transaction system and without changing the definition. .

  The reason is that the computer system according to the present embodiment accesses a file stored in the NAS using a symbolic link created in a directory indicating a serial number and an output permission state.

<Third Embodiment>
Next, a third embodiment based on the transaction system according to the second embodiment described above will be described. In the following description, the characteristic part according to the present embodiment will be mainly described. At this time, the same reference numerals are assigned to the same configurations as those in the second embodiment described above, and the duplicate description is omitted.

  FIG. 10 is a block diagram showing a configuration of a computer according to the third embodiment of the present invention.

  10, the transaction system 7000A includes a file search unit 1052 and a search file path generation unit 1930.

  In response to a file information acquisition request from the transaction control unit 1010A, the file search unit 1052 searches for a file.

  The file search unit 1052 generates a file storage directory path by the search file path generation unit 1930, searches for a file using the generated path, and stores the file information acquired by the search in the file information table 1041.

  The directory and symbolic link in the NAS 600 and each computer (computer 200 and computer 300) and the serial number management table 1921 of each business process (business process 210, business process 310, and business process 320) are shown in FIG. The third embodiment will be described using the state shown in the row and the row (h) in FIG. 5C as an initial state.

  Assume that a failure has occurred in the computer 300 in the transaction system 900A shown in FIG.

  Since the input from the external system does not stop until the failure is recovered, the business process 210 continues to output data to be passed to the business processes 310 and 320 to the NAS 600 and NAS 601.

  During the occurrence of a failure in the computer 300, the business processes 310 and 320 do not perform processing for reading the data of the NAS 600 and NAS 601, so there is a possibility that a large amount of data files stay in the NAS 600 and NAS 601.

  Here, it is assumed that the cause of the failure of the computer 300 has been removed.

  Before resuming the operation of the computer 300, the operator newly sets a directory for storing data processed by the business processes 310 and 320 in the NAS 600 and 601 as a storage destination of data to be output to the NAS after failure recovery. create.

  FIG. 11A and FIG. 11B are diagrams showing directories and symbolic links between each NAS (NAS 600 and NAS 601) and each computer (computer 200 and computer 300) in the third embodiment of the present invention.

  FIG. 11C is a diagram showing a serial number management table 1921 in each business process (business process 210, business process 310, and business process 320) in the third embodiment of the present invention.

  As shown in line (a) of FIG. 11A, the directory created here is created by using, for example, a name with “#” added to the end of the name of the directory created before the failure occurred. .

  Next, the serial number generation unit 1970 of the NAS control process 999 generates symbolic links to the directories created in the NASs 600 and 601 in each computer (computer 200 and computer 300), as shown in line (b) of FIG. 11A. To do.

  That is, since the serial number corresponding to the business process 310 is “X1” based on the serial number management table of the business process 210, the serial number generation unit 1970 carries this up to “X2” and generates a symbolic link.

  Further, the serial number generation unit 1970 generates a symbolic link by incrementing the serial number corresponding to the business process 320 to “X3” according to the serial number management table of the business process 210.

  Furthermore, the NAS control process 999 needs to switch directories for storing data processed by the business processes 310 and 320. Therefore, the output permission state operation unit 1960 changes the output permission state of the directory that was the output destination before the failure occurred, among the directories in which data to be processed in the business processes 310 and 320 are stored. That is, the output permission state under the directory of “/ export / A310” of the NAS 600 and “/ export / A320” of the NAS 601 is changed from the permitted state “vv” to the prohibited state “ii”, and the line (c) of FIG. Change as follows.

  As a result, as in the second embodiment, the business process 210 fails to access a file for outputting data to the business process 310, resulting in an error. The file output unit 1060 according to the present embodiment counts up the serial numbers triggered by a file access error. Using the serial number after counting up, the output file path generation unit 1910 regenerates the file path and performs file access again based on the recreated file path.

  As a serial number count-up process, the serial number management unit 1920 increments the serial number “X1” to “X2”, and updates the location corresponding to the business process 310 in the serial number management table 1921 of the business process 210 to “X2”.

  The regenerated file path is “/ data / X2 / A320 / vv / A210 /” from “X2” obtained as the serial number of the business process 310 from the serial number management table 1921 and “vv” in the output permission state. File access is successful.

  The same applies to the data output from the business process 210 to the business process 320. The serial number is incremented to “X3”, and the output destination is switched.

  As a result, the directories and symbolic links of each NAS (NAS 600 and NAS 601) and each computer (computer 200 and computer 300) are as shown in line (d) of FIG. 11A. Further, the serial number management table 1921 of each business process (business process 210, business process 310, and business process 320) is as shown in line (d) of FIG. 11C.

  Further, in other business processes, when normal file access has become an error, the file access is re-executed and the serial number management table 1921 of each business process is updated.

  As a result, the directories and symbolic links of each NAS and each computer are as shown in line (e) of FIG. 11B. Further, the serial number management table 1921 managed by each business process is as shown in line (e) of FIG. 11C.

  On the other hand, because the business processes 310 and 320 cannot receive the file information of the data output to the business processes 310 and 320 by the business process 210 while the computer 300 has a failure in the business processes 310 and 320. It has been destroyed.

  In addition, the table (file information table 1041 and serial number management table 1921) held in the computer 300 may be lost due to failure or restart of the computer 300.

  Therefore, the serial number management unit 1920 creates a serial number management table 1921 after failure recovery.

  First, the serial number management unit 1920 creates a list of directories in the “/ data” directory.

  The serial number management unit 1920 performs the following processing in order from the directory with the smallest serial number among the directories in the obtained directory list.

  When the business process name obtained from the file name is not registered in the serial number management table 1921 for all the files in the serial number directory, the serial number management unit 1920 sets the business process name and serial number as a pair. Then, it is additionally registered in the serial number management table 1921.

  When the business process name is already registered in the serial number management table 1921, the serial number management unit 1920 does nothing in particular.

  In the case of the computer 300 in the (c) line of FIG. 11A, the serial number management unit 1920 obtains “X1”, “X2”, and “X3” as the directory list. Next, since the minimum serial number is “X1”, the serial number management unit 1920 creates a directory list of the directory “/ data / X1”. In the example, as a result of referring to the “/ data / X1” directory, “A310” is obtained as a directory list.

  When the business process 310 is not registered in the serial number management table 1921, the serial number management unit 1920 registers the business process 310 as the serial number “X1” in the serial number management table 1921.

  By repeating such processing for the “X2” and “X3” directories, the serial number management unit 1920 creates a serial number management table in the business process 310 in the line (c) of FIG. 11C.

  Next, the serial number management unit 1920 creates a file information table 1041.

  The business processes 310 and 320 search a directory storing data to be processed and collect file information.

  FIG. 12 is a flowchart showing the process of the file search unit in the third embodiment of the present invention.

  The file search unit 1052 generates a path from the latest serial number and permission state by the search file path generation unit 1930 (S41).

  Since the business process 310 has the serial number “X1”, the path generated by the search file path generation unit 1930 is “/ data / X1 / A310 / vv”.

  Next, the file search unit 1052 performs a file search using the generated path (S42).

  However, since these paths do not exist, the file search unit 1052 generates a prohibited path “/ data / X1 / A310 / ii” when the file search fails (in the case of “NO” in S43) ( (S44) The file search is performed again based on the generated path (S42).

  This time, the file search unit 1052 can access the file that has been output successfully (in the case of “YES” in S43).

  By adding a data storage destination directory in the event of a failure or the like by the processing of the file search unit 1052, the file output unit 1060, the output path generation unit 1900, and the serial number management unit 1920 can change the definition information without changing the definition information. It is possible to switch the data output destination after failure recovery.

  In addition, since the data output destination is switched, the data before and after the failure recovery are output to separate files, and it is not necessary to perform the process of excluding the data output after the recovery from within the file.

  While the business processes 310 and 320 are processing the output unprocessed data, the other business processes output files to the newly created output destination directory, and the file information of the data is stored in the business processes 310 and 320. Notify Therefore, the business processes 310 and 320 can acquire file information of all unprocessed data by executing the file search once.

  When the business processes 310 and 320 detect a file by file search, they usually input a file in the same manner as when receiving file information of data from another business process. This is the same as in the second embodiment, and the business processes 310 and 320 have the output permission state changed from the permission state “vv” to the prohibited state by the file input unit 1050, the input file path generation unit 1900, and the serial number management unit 1920. Data can be input from a file that has been changed to “ii”.

  After the business processes 310 and 320 complete the processing for all data output before the failure recovery, the operator can output “/ export / A310 / ii” and “ / Export / A320 / ii "can be deleted. Further, the operator can also delete the symbolic link in the old serial number directory where the data processed by the business processes 310 and 320 in each computer is stored.

  When deletion is performed, the directories and symbolic links of each NAS (NAS 600 and NAS 601) and each computer (computer 200 and computer 300) are as shown in line (f) of FIG. 11B. Further, the serial number management table 1921 of each business process (business process 210, business process 310, and business process 320) is as shown in line (f) of FIG. 11C.

  As described above, the third embodiment has an effect that the input of data output before failure recovery can be performed without being affected by the data output after failure recovery.

  The reason is that the computer system according to the present embodiment performs control using the above-described serial number directory so that data before and after failure recovery is stored in different directories.

(Hardware configuration)
In the above-described embodiment, each unit illustrated in FIGS. 1, 3, and 10 can be regarded as a function (processing) unit (software module) of a software program. However, the division of each part shown in these drawings is a configuration for convenience of explanation, and various configurations can be assumed for mounting. An example of the hardware environment in such a case will be described with reference to FIG.

  FIG. 13 is a diagram for exemplifying the hardware configuration of a computer (information processing apparatus) capable of realizing the first, second, and third embodiments of the present invention. That is, FIG. 13 shows the whole or a part of the system shown in FIG. 1, or the transaction system 7000 shown in FIG. 3, the transaction system 900 shown in FIG. 4, and the transaction system 900A shown in FIG. This is a configuration of a computer (information processing apparatus) such as a server capable of realizing the above, and represents a hardware environment capable of realizing each function in the above-described embodiment.

  An information processing apparatus 9000 shown in FIG. 13 includes a CPU (Central Processing Unit) 9001, a display 9002, a communication interface (I / F) 9003, a ROM (Read Only Memory) 9004, a RAM (Random Access Memory) 9005, a hard disk device ( HD) 9006, and these are connected via a bus 9007. The hard disk device (HD) 9006 stores a program group 9006A and various kinds of stored information 9006B. The program group 9006A is a computer program for realizing a function corresponding to each block (each unit) shown in FIG. 1, FIG. 3, and FIG. The various storage information 9006B is, for example, the file information table 1041 and the serial number management table 1921 shown in FIG. A communication interface 9003 is a general communication unit that realizes communication with an external device via the network 9100.

  The present invention described with the embodiment described above as an example is a block diagram (FIGS. 1, 3, and 10) or a flowchart (FIGS. 8, 9, and 12) referred to in the description of the embodiment. After the computer program capable of realizing the above functions is supplied, the computer program is read out to the CPU 9001 of the hardware and executed using the hardware resources of the information processing apparatus 9000 shown in FIG. The The computer program supplied to the information processing device 9000 may be stored in a readable / writable temporary storage memory (9005) or a non-volatile storage device (storage medium) such as the hard disk device 9006.

  In the above case, the computer program can be supplied to the apparatus by a method of installing it in the apparatus via various storage media such as a CD-ROM or by an external communication line 9100 such as the Internet. A general procedure can be adopted at present, such as a method of downloading more. In such a case, the present invention can be understood to be constituted by a code constituting the computer program or a computer-readable storage medium in which the code is recorded.

DESCRIPTION OF SYMBOLS 1 Computer 11 Output permission state operation part 12 Serial number generation part 13 File path creation part 14 File access part 15 Processing part 100 External system 200 Computer 210 Business process 211 Business 300 Computer 310 Business process 311 Business 320 Business process 321 Business 600 NAS
601 NAS
900 transaction system 900A transaction system 999 NAS control process 1000 business process 1010 transaction control unit 1020 business 1030 external input unit 1040 process communication unit 1041 file information table 1050 file input unit 1051 input buffer 1052 file search unit 1060 file output unit 1900 input file path Generation unit 1910 Output file path generation unit 1920 Sequence number management unit 1921 Sequence number management table 1930 Search file path generation unit 1950 NAS control process 1960 Output permission state operation unit 1970 Sequence number generation unit 2010 Definition information 3000 NAS
5000 External system 7000 Transaction system 7000A Transaction system 8001 Computer 8002 Computer 8003 Computer 8004 NAS
9000 Information processing device (computer)
9001 CPU
9002 Display 9003 Communication interface (I / F)
9004 ROM
9005 RAM
9006 Hard disk device (HD)
9006A Program group 9006B Various stored information 9007 Bus 9100 Network

Claims (10)

An output permission state operation means for enabling determination as to whether or not the state of the first directory, which is the output destination of data, is output permission when outputting data to the NAS shared by a plurality of computers;
Serial number generation means for creating a second directory having a name including a serial number indicating an output order in the plurality of computers and storing a symbolic link to the first directory in the second directory;
File path generation means for generating a file path used for accessing the second directory;
File access means for accessing the NAS using the file path;
And a processing means for processing data input / output to / from the NAS accessed using the file access means. The computer according to claim 1, further comprising: a file search unit that searches for data output to the NAS using the file path to the second directory.   The computer according to claim 1 or 2, wherein the output destination of the data is switched when the replacement of the NAS or the addition of a new NAS becomes necessary.   3. The computer according to claim 1, wherein the output destination of the data is switched before and after the failure recovery.   5. A computer system comprising a plurality of computers according to claim 1 so as to be communicable via said NAS. When outputting data to a NAS shared by a plurality of computers, a first directory that is an output destination of the data and that can be determined whether output is permitted is created.
A second directory having a name including a serial number indicating an output order is created in the plurality of computers, and a symbolic link to the first directory is stored in the second directory;
Generating a file path used to access the second directory, accessing the NAS using the file path;
A NAS access method characterized by processing input / output data for the NAS.   The NAS access method according to claim 6, wherein the data output to the NAS is searched using the file path to the second directory.   The NAS access method according to claim 6 or 7, wherein when the replacement of the NAS or the addition of a new NAS becomes necessary, the output destination of the data is switched.   8. The NAS access method according to claim 6, wherein the output destination of the data is switched before and after failure recovery. A computer program for controlling a computer that is connected to the NAS so as to be accessible by the computer program,
An output permission state operation function that makes it possible to determine whether or not the state of the first directory that is the output destination of data is output permission when data is output to the NAS shared by a plurality of computers;
A serial number generation function for creating a second directory having a name including a serial number indicating an output order in the plurality of computers and storing a symbolic link to the first directory in the second directory;
A file path generation function for generating a file path used for accessing the second directory;
A file access function for accessing the NAS using the file path;
A computer program for causing a computer to realize a function of processing input / output data for the NAS accessed by the file access function.