JP2013025773A - File management device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve reliability of a file system without duplexing data with an inexpensive configuration.SOLUTION: A file management device 1 includes a controller (a backup manager 100) that executes backup processing and reboot processing. In the backup processing, a non-volatile storage area 30 holding management information containing the file name and write data of an opened file is secured and management information is stored in the non-volatile storage area 30 when a data write request is received from outside. In the reboot processing, in the recovery from power interruption during data write, the write data stored in the non-volatile storage area 30 is written back to a storage device 20 by referring to the non-volatile storage area 30.

Description

  The present invention relates to a file management apparatus that reads / writes data for each unit area via a file system having a cache area when a file operation is performed using a storage device.

  A method for managing files recorded in a file device on a computer and software for performing such management are called a file system. The file system receives a file operation request from an application program and efficiently saves data in the file device. In general, various items related to file management including a file creation method, a file management method on a file device, a storage area of a management area and a data area, and the like are determined depending on the type of the file system.

  For example, a file system called a FAT (File Allocation Table) file system manages file devices by dividing them into a management area for storing file management information and a data area for storing file contents. The management area includes a FAT and information called a directory entry, manages the storage location of data constituting the file by the FAT, and manages the file name, directory name, attribute, etc. by the directory entry (for example, Patent Document 1).

JP 2009-282275 A

  As disclosed in Patent Document 1, in a file system, file operation processing is interrupted when the power is suddenly shut down or the system crashes, and the file system can be recovered when a problem occurs in the contents of the file device. In addition, it is required to have a failure recovery function. Generally, a file system provides a high-speed file operation process for an application program by having a cache area. Therefore, if the power is suddenly shut down or the system goes down, the contents of the file operation stored in the cache area are not reflected on the file apparatus, or the management information on the file apparatus and the file in the data area are not reflected. There was a problem with reliability from the viewpoint of reliability.

  In the worst case, problems such as FAT chain fraud and cross-links in which two or more files use the same unit area occur, and all files stored in a storage device that is external storage are accessed. There is a risk that it will not be possible. Conventionally, in order to prevent this, reliability has been ensured by duplicating data. However, in this case, it is necessary to secure data equal to or larger than the capacity of the storage device, which is costly. Therefore, it is difficult to apply to a relatively small computer system such as a PC (Personal Computer). there were.

  SUMMARY An advantage of some aspects of the invention is that it provides a file management apparatus that improves the reliability of a file system with a low-cost configuration without duplicating data.

  In order to solve the above problems, the present invention is a file management apparatus that reads and writes data for each unit area via a file system having a cache area when performing file operations using a storage device, A non-volatile storage area for holding management information including a file name and write data of a file opened in a writable state is secured, and when a data write request is received from the file system, the management is stored in the non-volatile storage area. In the backup process for holding information, and when recovering from a power failure or system crash during data write, the write data held in the nonvolatile storage area is written to the storage device with reference to the nonvolatile storage area Having a restarting process and a control unit for executing And butterflies.

  According to the present invention, the control unit executes a process of writing back the write data to the storage device based on the management information held in the nonvolatile storage area at the time of recovery from a power failure or a system crash during the data write. . In this way, the management information is held in the non-volatile storage area secured during data write, and in the event of an unexpected situation such as a power failure, the restart process that writes back the data held in the non-volatile storage area to the storage device By executing this, data consistency can be maintained. Therefore, since it is not necessary to duplicate data, it can be expected to maintain and improve reliability while effectively using memory resources with a low-cost configuration.

  In the present invention, when the control unit writes back the write data, the controller assigns the cache area data in which a flag assigned to each unit area of the cache area is valid to an empty area of the nonvolatile storage area. And when the writing of the data in the cache area to the storage device is completed, the flag is invalidated and the write data held in the nonvolatile storage area is deleted to increase the space. . According to the present invention, it is possible to efficiently manage the capacity of a nonvolatile storage area by holding and deleting write data by flag control assigned to each unit area of the cache area. Moreover, high-speed temporary storage of write data can be realized by managing the nonvolatile storage area as a database.

  In the present invention, when there is no free space in the non-volatile storage area, the control unit refers to a time stamp further included in the management information and writes back old data preferentially to the storage device to increase the free space. It is characterized by that. According to the present invention, when there is no free space in the non-volatile storage area, the time stamp included in the management information is searched, the old data is preferentially written back to the storage device, and a free area for holding new data is created. In addition, efficient capacity management of the nonvolatile storage area and high-speed temporary storage can be realized.

  In the present invention, the control unit deletes the write data from the nonvolatile storage area based on a checksum value further included in the management information. According to the present invention, it is possible to efficiently delete write data from the nonvolatile storage area by managing the data held in the nonvolatile storage area based on the checksum value included in the management information.

  In the present invention, the control unit holds the file name between the file write function call and the end of the file write function call when the cache mode is write-through, and the file open function when the cache mode is copy back. The file name is held between a call and a file close function call. According to the present invention, the file name backup timing is controlled by the cache mode, so that the file name currently being written can be reliably held, and the location where the data inconsistency occurs can be localized. it can.

  In the present invention, the control unit refers to the nonvolatile storage area at the time of recovery from a power failure or a system crash during the data write, and if the write data is retained, the retained data Is written back to the storage device, and the write data held in the non-volatile storage area is deleted, and if the write data is not held, the file name that is not closed is acquired based on the management information Then, the process for repairing the acquired file name is executed. According to the present invention, it is possible to localize a location where data inconsistency occurs by holding the file name currently being written, and refer to the non-volatile storage area when a power shutdown or system crash occurs. Since the file consistency can be determined, the time required for the restart process can be shortened.

  According to the present invention, it is possible to provide a file management apparatus that improves the reliability of a file system with a low-cost configuration without duplicating data.

It is a block diagram which shows the structure of the file management apparatus which concerns on embodiment of this invention. It is a flowchart which shows operation | movement of the backup process of the file management apparatus which concerns on embodiment of this invention. It is a flowchart which shows the operation | movement of the restart process of the file management apparatus which concerns on embodiment of this invention.

  Hereinafter, an embodiment for carrying out the present invention (hereinafter simply referred to as the present embodiment) will be described in detail with reference to the accompanying drawings.

(Configuration of the embodiment)
FIG. 1 is a block diagram showing the configuration of the file management apparatus according to this embodiment. As shown in FIG. 1, a file management apparatus 1 according to this embodiment includes a host device 10 such as a PC (Personal Computer), an SD (Secure Computer), and the like.
It comprises a storage device 20 such as a Digital) memory card or a USB (Universal Serial Bus) memory. Characteristically, a non-volatile storage area 30 is allocated to a partial area of the storage device 20, and is managed and used as a backup DB (Data Base) of management information described later. The non-volatile storage area 30 may be assigned by attaching an external non-volatile storage device such as a flash memory, for example, without being assigned to a partial area of the storage device 20.

The host device 10 is shown in FIG.
As shown in FIG. 3, the program executed by the (Unit) is expanded and shown as an application program 101 (hereinafter simply referred to as application 101), file system 102, cache library 103, backup manager 100, and block device driver. 104. Here, for convenience, the program structures are listed in order from the upper layer.

  The application 101 is a program that performs file operations, and the file system 102 is the FAT described in the background art item used in Windows (registered trademark), for example. In the case of FAT, a storage device 20 is divided into logical units called clusters, and one or more clusters are allocated according to the data size, and a file is configured at this time. Information about whether a cluster is used or whether the cluster is used or not is recorded and managed in a management area.

  The cache library 103 has a cache area composed of a high-speed and small-capacity volatile memory. When the file system 102 issues a read request to a certain sector, the cache library 103 determines whether the data corresponding to the sector address exists in the cache area, and if it does not exist in the cache area, the storage device 20 The data of the target area is read from the data, and the data is held in the cache area. On the other hand, if it exists in the cache area, the data is transferred to the file system 102.

  When a write request is issued to a certain sector of the file system 102, the cache library 103 determines whether or not data corresponding to the sector address exists in the cache area, and if there is not, the storage device 20 To read the data of the target area to the cache area and write the corresponding data to the cache area. If there is data in the cache area, the data in the cache area is written to the storage device 20 at a time near the end of the file close process or write process after the data is written in the cache area.

  When the paging method is adopted, the cache library 103 is assigned a flag for each page in the cache area, so the page flag of the written area is turned ON, and data is stored only in the area where the flag is ON. The corresponding data is written to 20 at once.

  Note that the period during which the cache library 103 holds file names in the storage device 20 differs depending on the cache mode. Specifically, in the “write-through mode” in which data is written to the storage device 20 at the same time as data is written to the cache area, the data is held only from the time of the write function call to the end of the write function call, and cached in the free time. In the “copy back mode” in which the data written in the area is written in the storage device 20, the time is from the time when the file open function is executed until the time when the file close function is executed. Further, when file operations are performed using the storage device 20 as described above, each data in the file is finally read / written by the sector address by the block device driver 104.

  The backup manager 100 operates in cooperation with the file system 102, the cache library 103, and the block device driver 104, thereby “non-volatile holding management information including at least the file name and write data of the opened file” When the storage area 30 is secured and a data write request is received from the file system 102, a backup process for holding the management information in the nonvolatile storage area 30 and a non-volatile state at the time of power failure during data write or recovery from a system crash The control center functions as a “control unit that executes a restart process for referring to the storage area 30 and writing back the write data held in the nonvolatile storage area 30 to the storage device 20.

  Here, the information backed up in the non-volatile storage area 30 includes, in addition to the “file name” and “write data” that are the target of data write, the total value of data for checking data inconsistency “A checksum value” indicating “a write start sector address of the cache area”, “time stamp” which is time information at the time of data writing, and the like are included. In addition, for accessing the nonvolatile storage area 30, SQL (Structured Query Language), which is a relational database management system, is employed to realize efficient data management. Note that the files to be managed by the backup manager 100 are files (write processing) that may control the cache area flag in the cache library 103. For files that are opened for reading only, an unexpected power failure, etc. Even if this occurs, the management information of the file system 102 is not destroyed and is not managed. The reason for not being a management target is that the management target is not added.

(Operation of the embodiment)
Hereinafter, the operation of the file management apparatus 1 according to the present embodiment will be described with reference to the flowcharts of FIGS. First, the backup processing operation will be described with reference to FIG.

  First, when receiving a data write request from the file system 102, the cache library 103 writes data to the corresponding cache area and sets the corresponding flag to the ON state (step S101). Subsequently, at the timing when the flag is turned ON, the cache library 103 inquires the backup manager 100 about the free information in the nonvolatile storage area 30 and acquires the free state of the backup area (step S102).

  When the cache library 103 receives a response of free information from the backup manager 100 (step S103), if it is determined that “there is a free space” in the cache area (“YES” in step S104), the cache whose flag is ON The data in the area is deleted (step S108), and the backup data is inserted into the nonvolatile storage area 30 via the backup manager 100 and held (step S109). The data held at this time includes the “file name”, “write data”, “checksum value”, “cache area write start sector address”, and “time stamp” that are the targets of data write. Management information included.

  On the other hand, if it is determined in step S104 that there is no space (“NO” in step S104), the cache library 103 flushes the backup data stored in the non-volatile storage area 30 via the backup manager 100. Processing is executed (step S105). Subsequently, the backup manager 100 searches the write data held in the nonvolatile storage area 30 based on the time stamp included in the management information, and executes a process of writing to the storage device 20 in order of old data (step S106). ), The process of deleting the written data in order to increase the free area of the nonvolatile storage area 30 is executed, and the process returns to the free state acquisition process of the backup area (nonvolatile storage area 30) in step S102 (step S107).

  By the way, after the backup data insertion processing by the backup manager 100 (step S109), the cache library 103 executes the purge processing of the cache data held in the cache area (step S110), and the cache area data is transferred to the storage device 20. Write (step S111) and set the flag corresponding to the cache area for which writing has been completed to OFF state, and request the backup manager 100 to delete the backup data in the nonvolatile storage area 30 when this flag is turned OFF. Is performed (step S112).

  In response to this, the backup manager 100 refers to the management information, and executes a DB search for the backup data held in the nonvolatile storage area 30 based on the checksum value included in the management information (step S113). If there is data hit as a result of the DB search (step S114 “YES”), the corresponding backup data is efficiently deleted (step S115), otherwise (step S114 “NO”). End backup processing.

  Next, the operation of the restart process when the file management apparatus 1 according to the present embodiment is powered off or returned from the occurrence of a system crash will be described with reference to the flowchart of FIG. In this restart process, if backup data remains in the non-volatile storage area 30, there is a possibility that a power interruption or system crash may have occurred during data write. Data consistency between the storage device 102 and the storage device 20 is ensured.

  For example, when the power is restored after the power is shut down, the restoration work is executed in an upper layer such as an OS (Operating System), and the restart is performed (step S201), the backup manager 100 writes the write data to the nonvolatile storage area 30. An inquiry is made as to whether or not there remains, and the response is obtained (step S202). When the write data remains in the nonvolatile storage area 30 (step S203 “YES”), the backup manager 100 searches for the time stamp included in the management information, and writes and holds it in the storage device 20 in chronological order. A flash process for writing back all data is executed (step S204). The write data that is no longer necessary here is sequentially deleted from the nonvolatile storage area 30 and the process returns to the data acquisition process of step S202 (step S205).

  On the other hand, if there is no write data remaining in step S203 (step S203 “NO”), the backup data recovery processing is terminated. Next, the upper layer acquires the corresponding write data from the nonvolatile storage area 30 for the file name that has been closed without being closed (step S207). The upper layer checks whether or not the corresponding file exists (step S208), and if there is a corresponding file, performs a FAT chain repair and a cross-link check on the acquired write data, and writes the restored write data back to the storage device. File processing is executed (step S209).

  As described above, in the upper layer, after repairing data to the storage device 20, it is possible to perform a more reliable restart process by repairing a falchain fraud and a crosslink for a file name that has not been closed. Become. If there is no corresponding file in step S208 (step S208 “NO”), the above-described series of restart processing after return ends.

(Effect of the invention)
As described above, according to the file management apparatus 1 according to the present embodiment, the control unit (backup manager 100) stores the write data in the nonvolatile storage area at the time of recovery from power failure or system crash during data write. Data consistency can be maintained by executing a process of writing back to the storage device 20 based on the management information stored in the storage device 30. Therefore, since it is not necessary to duplicate data, it can be expected to maintain and improve reliability while effectively using memory resources with a low-cost configuration.

  Also, efficient storage management of the nonvolatile storage area 30 is possible by holding and deleting write data by flag control assigned to each unit area of the cache area. Further, by managing the nonvolatile storage area 30 as a database, high-speed temporary storage of write data can be realized. Furthermore, when there is no free space in the nonvolatile storage area 30, the free space is stored by referring to the time stamp further included in the management information and preferentially writing back old data to the storage device 20, increasing the free space, and holding new data. Therefore, efficient capacity management of the nonvolatile storage area 30 and high-speed temporary storage can be realized.

  Further, according to the file management apparatus 1 according to the present embodiment, the control unit (backup manager 100) deletes the write data from the nonvolatile storage area 30 based on the checksum value further included in the management information. As described above, the database management based on the checksum value for checking the data mismatch between the file system 102 and the storage device 20 enables efficient deletion of the write data from the nonvolatile storage area 30. Also, by controlling the timing of file name backup by the cache mode, the file name currently being written can be reliably held, and the location where the data inconsistency has occurred can be localized.

  Further, according to the file management apparatus 1 according to the present embodiment, the location where the data inconsistency has occurred can be localized by holding the name of the file currently being written, and non-volatile when a power shutdown or system crash occurs Since the consistency of the corresponding file can be determined by referring to the property storage area 30, the time required for the restart process can be shortened.

  In the file management apparatus 1 according to the present embodiment, the backup manager 100 is described as a program module that exists independently of the file system 102. However, the backup manager 100 is incorporated in the file system 102 and the same. The effect of can be obtained. In addition, the file system 102 having a temporary holding mechanism such as the cache library 103 can be applied without being limited to the FAT file system. Further, not only the file system but also the information to be retained that is handled inside the device, high reliability can be ensured by applying the present invention.

  In addition, with the mechanism that duplicates file data, by holding backup data for one storage device, if a problem specific to the storage device occurs on one side, restore the data held on the other side to ensure consistency. It is possible to take. Further, when a partial area of the storage device 20 capable of high-speed processing is allocated as the nonvolatile storage area 30, it is possible to perform higher-speed processing by treating it as an area for temporarily storing temporary data.

  As described above, according to the file management apparatus 1 according to the present embodiment, when the write data handled internally is held in the storage device 20, the non-volatile memory allocated externally or assigned to a partial area of the storage device 20 By efficiently holding only the portions that need to be changed in the storage area 30, it is not necessary to duplicate all the areas, so that it is possible to save a temporary storage area and improve data reliability. In addition, high-speed temporary storage can be realized by DB management of this temporary storage area. Further, even when a power interruption or a system crash occurs during data writing to the storage device 20, data consistency can be maintained by writing back the backup data in the nonvolatile storage area 30. Furthermore, by holding the name of the file currently being written in the non-volatile storage area 30, it is possible to localize the location where data inconsistency has occurred, and check only the consistency of the corresponding file By performing the above, it is possible to shorten the time for the return process.

  As mentioned above, although this embodiment was described, it cannot be overemphasized that the technical scope of this invention is not limited to the range as described in the said embodiment. It will be apparent to those skilled in the art that various modifications or improvements can be added to the above embodiment. Further, it is apparent from the scope of the claims that the embodiments added with such changes or improvements can be included in the technical scope of the present invention.

  DESCRIPTION OF SYMBOLS 1 ... File management apparatus, 10 ... Host apparatus (PC), 20 ... Storage device, 30 ... Non-volatile storage area (backup DB), 100 ... Backup manager (control part), 101 ... Application program, 102-File system, 103-Cache library, 104-Block device driver

Claims (6)

A file management device that reads / writes data for each unit area via a file system having a cache area when performing file operations using a storage device,
A non-volatile storage area for holding management information including a file name and write data of a file opened in a writable state is secured, and when a data write request is received from the file system, the management is stored in the non-volatile storage area. Backup processing for holding information, and when recovering from power failure or system crash during data write, refer to the non-volatile storage area, and write data stored in the non-volatile storage area to the storage device A file management apparatus comprising: a control unit that executes a restart process for writing back. The controller is
When writing back the write data, the cache area data in which a flag assigned to each unit area of the cache area is enabled is held in a free area of the nonvolatile storage area, and the storage device 2. The file management apparatus according to claim 1, wherein when writing of data in the cache area ends, the flag is invalidated and the write data held in the nonvolatile storage area is deleted to increase the space. The controller is
3. The storage device according to claim 2, wherein when there is no free space in the non-volatile storage area, the time stamp further included in the management information is referred to, and old data is preferentially written back to the storage device to increase the free space. File management device. The controller is
3. The file management apparatus according to claim 2, wherein the write data is deleted from the nonvolatile storage area based on a checksum value further included in the management information. The controller is
If the cache mode is write-through, the file name is retained between the file write function call and the end of the file write function call. If the cache mode is copy back, the file open function call to the file close function call The file management apparatus according to claim 1, wherein the file name is held in a file name. The controller is
When recovering from a power failure or system crash during the data write, referring to the non-volatile storage area, and if the write data is retained, write back the retained data to the storage device, When the write data held in the nonvolatile storage area is deleted, and the write data is not held, the file name that is not closed is acquired based on the management information, and the acquired file name is restored The file management apparatus according to claim 1, wherein the process is executed.

Images