JP2009110477A - File management system and file management program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To enable a file that time has not passed so much after its deletion to be restored. <P>SOLUTION: A deletion file management module 24 deletes a file of which deletion is required from application from name space, and registers the file in a queue table 24e together with the deletion date, but disk block in which data of the file is stored is kept protected. The deletion file management module 24 releases the disk block in which data of the file is stored whenever prescribed event occurs if there is a file whose preservation period exceeds prescribed upper limit value, and erases registration of the file to the queue table 24e. Further, when file path of that time of the file is appointed and restoration of the file is required, the deletion file management module 24 generates directory entry of the file if the file is registered in the queue table 24e. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an apparatus and a program for managing files.

  As is well known, there are computers that have a trash can icon on their desktop. An operator can discard a file simply by dragging and dropping an icon of an unnecessary file to the trash can icon. However, the file was simply transferred to the trash directory from the usual directories and discarded. Therefore, the data of the file is stored in the data area of the disk in the last updated state, and the disk block in which the data of the file is stored in the data area is not overwritten. Protected. For this reason, simply discarding a file in the trash does not increase the free space on the disk.

In addition, as is well known, the file system deletes a file name from the name space that is a system of files and their hierarchical classification (directory) even if the delete [delete] instructed by the application is executed, Only the association between the file name and the disk block storing the data is canceled, and the file data is not erased from the data area (see, for example, Non-Patent Document 1). The disk block storing the data of the file to be deleted is switched to the overwritten release state (unused state) by the file system, but the file data remains stored. Since the released disk block is included in the free space of the disk, the free space of the disk is increased unlike the discarding to the trash box.

In recent years, a file restoration tool that restores a deleted file using such characteristics of the file system has been developed. The file restoration tool basically searches the data area for data that remains without being erased, identifies the data that should be grouped together to form the same file, and identifies the identified file. Associate a disk block of data with a file name, and generate the file name in the namespace. Therefore, the deleted file can be restored as long as the data remains in the data area without being overwritten.

  The file is restored only when the need for restoration arises. Eventually, even deleted files will be forgotten if there is no need for restoration. The need for restoration often occurs shortly after the file is deleted. Therefore, in practice, it is desirable that the file be stored in the disk for a while after being deleted.

  However, the file system overwrites data in a disk block that is in a released state in the data area without considering the circumstances of data remaining in the disk block. Therefore, it is possible to restore a file that has been deleted and not much time has passed even though a considerable amount of time has passed since the deletion and the data of the file that is hardly necessary to be restored still remains. Some or all of the data in the file may have already been erased by overwriting the data in another file and cannot be restored.

Patent Document 1 discloses a tool for securing a free area of a storage medium. This tool saves a file with the oldest reference date from the large-capacity storage area to the storage medium if there is not enough free space on the large-capacity storage medium when writing any file. To secure a free space in a large-capacity storage medium, and if the file does not exist in the large-capacity storage medium when any file is referenced, the file is stored in the large-capacity storage medium from the evacuation storage medium. It returns to the medium. According to this tool, it is possible to write a file to a large-capacity storage medium without interrupting the program, but it is not possible to restore a file that has not been deleted so much time.

“What is a file (including directory)”, [online], Kagawa University, [May 24, 2007 search], Internet <URL: http://stwww.eng.kagawa-u.ac.jp/ ~ imai / H17ComputerSystem / file-directory.html> Japanese Patent Laid-Open No. 03-078042

  The present invention has been made in view of the problems of the prior art as described above, and an object of the present invention is to be able to restore a file that has been deleted and has not passed so long.

  The file management device devised to solve the above problems is a management of a file and a name space that is a hierarchical classification system, a file name and a disk block in a data area in which data is stored. In order to perform management of association and control of file access from the application, a first reception unit that receives deletion of any file from the application, and when the first reception unit receives deletion of the file, When the first deletion unit that deletes the file name from the name space and the first reception unit accept the deletion of the file, a record that associates the file identification information for identifying the file, the file path, and the deletion date and time Registration part registered in the queue table, when a predetermined event occurs, from the queue table to the point in time Extraction unit that extracts records whose retention period exceeds a predetermined upper limit value, and when the extraction unit extracts records whose retention period exceeds a predetermined upper limit value, it is identified by the file identification information in that record A second unit that releases the association between the file to be recorded and the disk block in the data area, releases the disk block storing the file data, and deletes the record extracted by the extraction unit from the queue table The deletion unit, the second reception unit that receives the restoration and file path of the deleted file from the operator, and the second reception unit receives the restoration and file path of the deleted file from the operator. Using the directory path in the file path as a search condition, the search unit that searches the queue table, and when the search unit detects a record, The file identified by the file identification information in the record, further comprising: a restoring unit for generating a directory specified by the directory path is characterized.

  When configured in this way, the file requested to be deleted by the application is deleted from the namespace, but instead, the disk block registered with the deletion date and time in the queue table and storing the file data Remain protected. When a predetermined event occurs, the existence of a file whose retention period exceeds a predetermined upper limit value is confirmed. If there is a file whose retention period exceeds a predetermined upper limit value, the data of the file is stored. The disk block is released and the file is deleted from the queue table. Also, when the file path of the file at that time is specified and the restoration of the file is requested, whether the file is registered in the queue table based on the directory path in the file path is searched and detected. Then, the file is generated in the directory specified by the directory path.

In other words, the file requested to be deleted is deleted from the name space and is nominally deleted, but the data is protected for a certain period and can be overwritten after the certain period. In addition, if restoration is requested within the certain period, the file is restored to the name space.

  The above operation can also be realized by a file management program or a file management method. That is, the present invention may be a file management program that causes a computer to operate as a plurality of means that function in the same manner as each unit of the file management apparatus described above, or the computer executes functions equivalent to those units as a plurality of procedures. It may be a file management method.

  Therefore, according to the present invention, it is possible to restore a file that has been deleted and has not passed so long.

  Hereinafter, embodiments for carrying out the present invention will be described with reference to the accompanying drawings.

<Configuration>
As shown in the block diagram of FIG. 1, the computer 10 of this embodiment includes a display unit 10a such as a liquid crystal display, an input unit 10b such as a keyboard and a mouse, and a main body to which these units 10a and 10b are connected. . The main body includes a storage unit 10c, a CPU [Central Processing Unit] 10d, and a memory unit 10e. The storage unit 10c is a unit that stores various programs and data on a disk in a readable manner. The CPU 10d is a unit that performs processing according to a program in the storage unit 10c. The memory unit 10e is a unit in which the CPU 10d caches programs and data and develops a work area.

Further, the computer 10 of the present embodiment stores various applications (software) 11, operating system software 12, and various device drivers (programs) 13 in the storage unit 10c. The application 11 is a set of programs and data for realizing a certain function. The operating system software 12 provides API [Application Programming Interface] and ABI [Application Binary Interface] to various applications, management of storage areas of the storage unit 10c and the memory unit 10e, management of processes and tasks, file management, and various types. This is software for providing various applications 11 such as setting tools and editors, and for assigning windows to a plurality of tasks for multiplexing screen output. The device driver 13 is a program for operating peripheral devices in accordance with instructions from the operating system software 12.

  The operating system software 12 includes a virtual file system 12a, a plurality of local file systems 12b, and an interface (program) 12c, as shown in the memory development view of FIG. Among these, the local file system 12b is a program for managing data recorded on the disk in the storage unit 10c in units of files. The local file system 12b manages at least one storage disk, and performs name space management, disk block management, and access control, which are the basic three functions of the file system, for the files in the storage disk.

Namespace management refers to management of a name space, which is a system of files and their hierarchical classification (directory), and disk block management refers to a storage disk in which a file name and data constituting the file are stored. The management of the association with the disk block inside. The management information used for name space management and disk block management is generally referred to as metadata. This metadata 12d is stored in an area other than the file data area in the storage disk by the local file system 12b. Stored. The access control refers to execution control of reading, writing, and deletion of a file with respect to a disk block based on file access (request for file operation) from the application 11 and exclusive control of access right.

  There are several types of the local file system 12b depending on the form of the metadata and the form of the file data, or the difference between the management methods. Different types of local file systems 12b have different formats on their storage disks, and therefore cannot manage the same storage disk. In this embodiment, a plurality of types of local file systems 12b are implemented in the operating system software 12 in order to manage each of the plurality of storage units 10c. However, the application 11 performs system calls according to the types of the local file systems 12b. The operating system software 12 is provided with a virtual file system 12a so that it is not necessary to use the two separately. The virtual file system 12a accepts a file access request from the application 11 with a unified system call, identifies a unique process of the local file system 12b that manages the file for which access is requested, and determines the local file system. 12b is handed over. The device driver 13 has a different command and data format depending on the manufacturer and product type of the storage unit 10c. Therefore, the operating system software 12 has an interface 12c. The interface 12c bridges information for I / O [Input / Output] between the local file system 12b and the device driver 13 of the storage unit 10c.

  In the present embodiment, the functions according to the present invention are incorporated in each local file system 12b. As shown in the explanatory diagram of FIG. 3, each local file system 12b includes an access control module (program) 21, a name space management module 22, and a disk block management module 23 as conventional functions. As a function related to the above, a deletion file management module 24 is included. The access control module 21, the name space management module 22, and the disk block management module 23 are module programs that respectively perform the above-described access control, name space management, and disk block management.

In order to more specifically describe the operations of the access control module 21, the namespace management module 22, and the disk block management module 23, an ext [Extended File System] disk format is illustrated in FIG. In the disk format illustrated in FIG. 3, the disk in the storage unit 10c is divided into several partitions, and one partition is divided into a boot area and several block groups. The boot area is an area in which a boot loader that performs partition table inspection or the like is stored, and the block group is an area occupied by one block group when the blocks in one partition are divided into predetermined groups. . Further, one block group is divided into a metadata area 25 and a file data area 26. The file data area 26 is an area in which normal files and directories (files storing information related to them) are recorded. The metadata area 25 is an area where the metadata 12d is recorded. In this way, dividing one partition into several block groups and setting the metadata area 25 and the file data area 26 in each block group means that the ext file system exemplified in FIG. The UFS [UNIX File System] (UNIX is a trademark of the Ziopen Group), which is a major file system of the operating system called Solaris (trademark of Sunsoft, USA). The present invention is not limited to a (local) file system that divides one partition into a plurality of block groups (cylinder groups in UFS), and includes a boot area, a metadata area, and a file data area in one partition. It can also be applied to (local) file systems. An example of such a file system is PRIMECLUSTER GFS (trademark of Fujitsu Limited).

  The ext-type disk format illustrated in FIG. 3 will be further described. As shown in the layout diagram of FIG. 4, the metadata area 25 includes a super block, a group descriptor area, a block bitmap area, an i-node bitmap area, and The i-node table area. The i-node table area is an area in which attribute information of a plurality of files (including files related to directories) stored in the file data area 26 is stored. In this i-node table area, a predetermined number of i-nodes for storing file attribute information are prepared, and the i-nodes are arranged side by side. Accordingly, each i-node in this area conceptually constitutes an i-node table 25a. Each file in the file data area 26 is assigned one i-node from this i-node table 25a. The file attribute information stored in the i-node includes i-node number (ID [Identification]), permission, owner, file size, modification date, pointer (corresponding file (or directory) in the file data area 26). The number of the disk block in which is stored). However, in the i-node table 25a shown in FIG. 4, the field in which the i-node number (ID) and the pointer are recorded is shown, and the other fields are not shown. In addition, an i-node having an i-node number (ID) of 2 is assigned to the root directory, and an i-node of No. 3 or later is assigned to the subdirectory. It should be noted that a directory (a file in which information related to it is stored) is associated with the name of a subdirectory or file belonging to that directory and the i-node number (ID) of the i-node assigned thereto. Stored as a directory entry. The i-node bitmap area is an area in which presence / absence of file allocation for each i-node on the i-node table 25a is recorded. The block bitmap area is an area in which the presence or absence of data storage for each disk block in the file data area 26 is recorded. The group descriptor area is an area possessed only by the 0th block group, and is an area in which information relating to the block group is recorded. The super block is an area in which information regarding basic attributes of the local file system 12b, such as the number of i-nodes, the number of blocks, the number of free blocks, the number of free i-nodes, and the like is recorded.

  When the application 11 requests a file read or file write, the access control module 21 shown in FIG. 3 stores a file to be processed and data of the file in the name space management module 22 and the disk block management module 23. The number of the current disk block is inquired, and file data is read from and written to the specified disk block. Further, when the application 11 requests deletion of a file, the access control module 21 notifies the deletion file management module 24 to that effect.

  When the namespace management module 22 receives a file inquiry from the access control module 21, the name space management module 22 creates a file based on the i-node table 25a in the metadata 12d and each directory pointed to (a file storing information on the directory). The file to be processed is specified from the path. In addition, when the access control module 21 instructs the name space management module 22 to change the file name or directory name or to change the file affiliation (directory), the name space management module 22 performs a process of reflecting the change in the name space. It is like that.

When receiving a disk block inquiry from the access control module 21, the disk block management module 23 notifies the access control module 21 of the pointer of the processing target file in the i-node table 25a in the metadata 12d. When the access control module 21 is instructed to write a file, the disk block management module 23 associates the name of the file with the disk block in the storage disk storing the data constituting the file. The process to change is performed.

  Generally, the metadata 12d is cached on the memory unit 10e from the metadata area 25 of the disk in the storage unit 10c. When the application 11 performs file access, the name space management module 22 and the disk block management module 23 update the metadata 12d on the memory unit 10e. The metadata 12d on the memory unit 10e is reflected on a request from the application 11 or periodically on the metadata area 25 of the disk in the storage unit 10c.

  Further, the deleted file management module 24 according to the present invention shown in FIG. 3 controls the delay and execution of the deletion of the file and determines whether or not the file can be restored upon receiving the instruction to restore the file, as will be described in detail later. Is a module program that controls execution and execution. As shown in the configuration diagram of FIG. 5, the deletion file management module 24 includes a deletion reception program 24a, an empty block securing program 24b, a block release control program 24c, a restoration program 24d, a queue table 24e, and an empty block number storage area 24f. , Including.

  The deletion reception program 24a is a program for receiving a notification from the access control module 21 that has received a file deletion instruction from the application 11 and registering the deletion file in the queue table 24e. The contents of the process executed by the CPU 10d according to the deletion acceptance program 24a will be described later with reference to FIG.

  The free block securing program 24 b is a program for securing a free block in the file data area 26 in response to a notification from the access control module 21 that has received a file write instruction from the application 11. The contents of the processing executed by the CPU 10d according to the empty block securing program 24b will be described later with reference to FIG.

  The block release control program 24c is a program for determining whether or not the disk block storing the file data registered in the queue table 24e can be released and executing it. The contents of processing executed by the CPU 10d according to the block release control program 24c will be described later with reference to FIGS.

  The restoration program 24d is a program for receiving a notification from the access control module 21 that has received a file restoration instruction from the application 11 and deleting the registration of the file in the queue table 24e. The contents of the process executed by the CPU 10d according to the restoration program 24d will be described later with reference to FIG.

The queue table 24e is a table for registering a file requested to be deleted from the application 11. As shown in the table structure diagram of FIG. 6, each record in the queue table 24e has fields of “queue number”, “i-node number”, “file path”, and “deletion date”. The “queue number” field is a field in which a queue number that specifies the order of the files is recorded. The “i-node number” field is a field in which the i-node number (ID) of the i-node assigned to the file is recorded. The “file path” field is a field in which a file path immediately before deletion of the file is recorded. The “deletion date and time” field is a field in which the date and time when the application 11 instructs to delete the file is recorded.

  The empty block number storage area 24f is an area in which the number of empty blocks is recorded, and is generated on the memory unit 10e. The empty block number storage area 24f may be present in the metadata 12d copied by the local file system from the disk in the storage unit 10c onto the memory unit 10e. When the local file system 12d is the ext file system exemplified in FIG. 3, the free block number storage area 24f is generated in the super block of the metadata 12d copied (cached) in the memory unit 10e.

  In addition, two types of empty block numbers are recorded in the empty block number storage area 24f. On the other hand, the number of internal free blocks is generally the number of free blocks recorded in the metadata area 25 and matches the number of unused disk blocks in the file data area 26. The number of external free blocks on the other side is the sum of the number of disk blocks unused in the file data area 26 and the number of disk blocks storing files registered in the queue table 24e. ,Match.

<Processing>
When the programs 12a to 12c related to the file system in the operating system software 12 are activated, the execution of the block release control program 24c among the programs 24a to 24d shown in FIG. At the same time, a queue table 24e and an empty block number storage area 24f are generated on the memory unit 10e. The programs 24a, 24b, and 24d other than the block release control program 24c are executed in accordance with a request from a so-called user process that is activated based on an instruction from the operator.

(Deletion process)
First, when the access control module 21 receives a delete instruction from the application 11 (via the virtual file system 12a), the access control module 21 starts the delete reception program 24a. After the start of the deletion acceptance program 24a, as shown in the flowchart of FIG. 7, the CPU 10d receives the file path of the file designated to be deleted from the access control module 21 in step S101. Note that the CPU 10d that executes step S101 corresponds to the first reception unit and the first reception unit described above.

  In the next step S102, the CPU 10d reads the directory path from the file path received in step S101, and deletes the directory entry of the file designated for deletion from the directory specified by the directory path. Instruct module 22. The CPU 10d that executes this step S102 corresponds to the first deletion unit and the first deletion unit described above.

  In the next step S103, the CPU 10d updates the number of external free blocks recorded in the free block number storage area 24f by increasing the number of disk blocks storing files designated for deletion. .

  In the next step S104, the CPU 10d registers a record relating to the file designated to be deleted at the end of the queue table 24e in FIG. The CPU 10d that executes this step S104 corresponds to the above-described registration unit and registration means.

  In the next step S105, the CPU 10d instructs the block release control (based on the program) process 24c to start execution. Thereafter, the CPU 10d ends the deletion acceptance process according to FIG.

  When the deletion request program 24a requests the application 11 to delete a file, the file is deleted from the name space (step S102), registered in the queue table 24e (step S104), and the number of free blocks is stored. The number of external free blocks in the area 24f increases by the amount of the file (step S103). However, the number of internal free blocks in the free block number storage area 24f does not increase.

(Free block allocation processing)
Further, when the access control module 21 receives a write instruction from the application 11 (via the virtual file system 12a), the free block reservation program 24b is activated in response to the write instruction. After the start of the free block securing program 24b, as shown in the flowchart of FIG. 8, the CPU 10d inquires of the disk block management module 23 about the number of blocks necessary for writing the file in step S201. And CPU10d will advance a process to the following step S202, if the number of blocks required for writing is acquired from the disk block management module 23. FIG.

  In step S202, the CPU 10d reads the number of empty blocks for the inside from the empty block number storage area 24f.

  In the next step S203, the CPU 10d determines whether or not the number of internal free blocks read in step S202 is equal to or larger than the necessary number of blocks acquired in step S201. If the number of internal free blocks is less than the required number of blocks acquired in step S201, the CPU 10d branches the process from step S203 to step S204.

  In step S204, the CPU 10d instructs the block release control (based on the program) process 24c to start execution.

  In the next step S205, the CPU 10d stands by until an execution end notification is received from the block release control process 24c. When the execution end notification is received, the CPU 10d returns the process to step S202.

  On the other hand, in step S203, if the number of internal free blocks read in step S202 is equal to or greater than the required number of blocks acquired in step S201, the CPU 10d advances the process from step S203 to step S206.

  In step S206, the CPU 10d instructs the access control module 21 to continue writing the file (and assign a disk block necessary for the writing). Thereafter, the CPU 10d ends the free block securing process according to FIG.

When writing of a file is requested by the application 11 by the free block securing program 24b, the writing of the file is interrupted, and during that time, the number of internal free blocks in the free block number storage area 24f is necessary for writing. It is confirmed whether it is more than the amount (step S203). If the number of internal free blocks is greater than or equal to the amount necessary for writing, the file writing is continued as it is (step S206), and if the number of internal free blocks is less than the amount required for writing, the block The release control process 24c is executed to secure the number of internal free blocks (step S20).
4, S205), the file writing is continued (step S206).

(Block release processing)
In the block release control process 24c, as shown in the flowcharts of FIGS. 9 and 10, first, in step S301, the CPU 10d determines whether or not there has been an execution start instruction from the deletion acceptance program 24a or the free block reservation program 24b. Determine. If there is no execution start instruction, the CPU 10d branches the process from step S301 to step S302.

  In step S302, the CPU 10d determines whether or not a predetermined time has elapsed since the start of the process 24c or after the end of the previous time measurement. If the predetermined time has not elapsed since the start of the process 24c or the end of the previous time measurement, the CPU 10d branches the process from step S302 and returns the process to step S301.

  During execution of the processing loop of steps S301 and S302, if there is an execution start instruction from the deletion acceptance process 24a or the free block securing process 24b, or if a predetermined time has elapsed since the start of the program 24c, the CPU 10d Then, the process proceeds to step S303.

  During the execution of the processing loop of steps S301 and S302, the CPU 10d advances the process to step S303 even when a predetermined time has elapsed after the start of the process 24c or after the end of the previous time measurement.

  In step S303, the CPU 10d searches the queue table 24e of FIG. 6 without a search condition.

  In the next step S304, the CPU 10d determines whether or not a record has been detected from the queue table 24e. If no record is detected (registered) in the queue table 24e, the CPU 10d branches the process from step S304 to step S313. On the other hand, when one or more records are detected from the queue table 24e, the CPU 10d executes the first processing loop L1.

  In the first processing loop L1, the CPU 10d executes the processing of steps S305 to S312 one by one with respect to the records detected from the queue table 24e in FIG.

  In step S305, the CPU 10d calculates a retention period from the deletion date and time in the processing target record and the date and time at this time.

  In the next step S306, the CPU 10d determines whether or not the storage period calculated in step S305 exceeds a predetermined upper limit value. If the retention period exceeds the predetermined upper limit value, the CPU 10d advances the process from step S306 to step S309. On the other hand, when the retention period is equal to or shorter than the predetermined upper limit value, the CPU 10d branches the process from step S306 to step S307.

  In step S307, the CPU 10d reads the number of internal free blocks from the free block number storage area 24f.

In the next step S308, the CPU 10d determines whether or not the number of internal free blocks read in step S307 is equal to or greater than a predetermined lower limit value. When the internal free block count is equal to or greater than the predetermined lower limit value, the CPU 10d leaves the first processing loop L1 without performing the first processing loop L1 for the remaining records after the processing target, The process proceeds to step S313 in FIG. On the other hand, if the number of internal free blocks is less than the predetermined lower limit value, the CPU 10d branches the process from step S308 to step S309. The CPU 10d that executes these steps S301 to S309 corresponds to the extraction unit and the extraction unit described above.

  In step S309, the CPU 10d instructs the namespace management module 22 to cancel the allocation of the i-node to the file specified by the processing target record. Upon receiving this instruction, the name space management module 22 deletes information related to the file from the i-node table area of the metadata area.

  In the next step S310, the CPU 10d instructs the disk block management module 23 to release the disk block in which the data of the file specified by the processing target record is stored. Upon receiving this instruction, the disk block management module 23 updates the block bitmap area of the metadata area. The CPU 10d that executes step S310 corresponds to the above-described release unit and release unit.

  In the next step S311, the CPU 10d updates the number of internal free blocks in the free block number storage area 24f by increasing the number of disk blocks in which the file specified in the processing target record is stored. .

  In the next step S312, the CPU 10d deletes the processing target record from the queue table 24e in FIG. 6, and moves up the record that is after the processing target record. Thereafter, the CPU 10d ends the execution of the first processing loop L1 for the processing target record. The CPU 10d that executes this step S312 corresponds to the second deletion unit described above.

  When the CPU 10d finishes executing the processes in steps S305 to S312 described above for all the records detected from the queue table 24e in FIG. 6, the CPU 10d leaves the first processing loop L1 and performs step S313 in FIG. Proceed to the process.

  In step S313, the CPU 10d determines whether or not an execution end notification is necessary. If the execution end notification is not necessary, the CPU 10d branches the process from step S313, and ends the block release control process according to FIGS. On the other hand, when the execution end notification is necessary, the CPU 10d advances the process from step S313 to step S314.

  In step S314, the CPU 10d notifies the execution end to the process that has instructed the execution start. Thereafter, the CPU 10d ends the block release control process according to FIGS.

  The block release control program 24c starts execution immediately after completion, and waits until an execution start instruction is received or a predetermined time elapses.

The block release control program 24c confirms the storage period of the records registered in the queue table 24e in FIG. 6 in order from the beginning, and the storage period exceeds a predetermined upper limit value, or the storage period Even if is less than or equal to a predetermined upper limit value, if the number of internal free blocks in the free block number storage area 24f is below a predetermined lower limit value, the previous record is specified (steps S305 to S308). Then, the attribute information of the file related to the specified record is deleted from the i-node table 25a (step S309), and the disk block storing the data of the file is released (step S310). , Completely deleted. That is, when it is determined by the block release control process 24c that a file that has already been deleted from the name space by the deletion receiving program 24a is to be deleted (step S306; yes or step S308; no). The file data area 26 is also completely deleted.

(Restore process)
In addition, when the access control module 21 receives a restoration instruction from the application 11 (via the virtual file system 12a), the restoration control program 24d is started using this as an opportunity. After starting the restoration program 24d, as shown in the flowchart of FIG. 11, the CPU 10d receives the file path of the file designated to be restored from the access control module 21 in step S401.

  In the next step S402, the CPU 10d reads the directory path from the file path acquired in step S401, and inquires of the namespace management module 22 about the directory path.

  In the next step S403, the CPU 10d determines whether or not the result of the inquiry in step S402 is that there is a directory path. When the inquiry result indicates that there is no directory path, the CPU 10d branches the process from step S403, and ends the restoration process according to FIG. On the other hand, if the result of the inquiry is that there is a directory path, the CPU 10d advances the process from step S403 to step S404.

  In step S404, the CPU 10d searches the queue table 24e of FIG. 6 using the file path received from the access control module 21 together with the notification in step S401 as a search condition.

  In the next step S405, the CPU 10d determines whether or not a record has been detected as a result of the search in step S404. If no record is detected, the CPU 10d branches the process from step S405 and ends the restoration process according to FIG. On the other hand, if the record can be detected, the CPU 10d advances the process from step S405 to step S406. The CPU 10d that executes these steps S402 to S405 corresponds to the search unit and search means described above.

  In step S406, the CPU 10d determines, based on the i-node number (ID) in the detection record and the file name in the file path, a file in which information related to the directory (specified in the directory path inquired in step S402 is stored). ) To instruct the namespace management module 22 to generate a directory entry for the file. Note that the CPU 10d executing step S406 corresponds to the above-described restoration unit and restoration unit.

  In the next step S407, the CPU 10d updates the number of external free blocks recorded in the free block number storage area 24f by reducing the number of disk blocks storing files designated for restoration. .

In the next step S408, the CPU 10d deletes the detected record from the queue table 24e in FIG. 6, and moves up the record that exists after that record. Then the CPU
10d ends the restoration process according to FIG.

  When the application 11 requests the restoration of the file by the restoration program 24d, it is confirmed whether the directory path in the file path immediately before the deletion exists (step S403), and whether the file is registered in the queue table 24e. Then, the file name is generated in the name space (step S406), the number of external free blocks is reduced by the amount of the file (step S407), and the file queue table 24e is entered. Is deleted (step S408).

<Action>
FIG. 12 shows a directory tree based on the contents of the directory shown in FIG. For example, if an instruction to delete a file whose file path is “/A/C/sample1.exe” is given through the application 11, the deletion acceptance program 24a registers the file in the queue table 24e as shown in FIG. As shown in FIG. 13, the directory entry of the file is deleted, and the directory tree becomes as shown in FIG. However, as shown in FIG. 13, i-nodes are still assigned to the file.

Thereafter, the continuation of the record in the queue table 24e is triggered by the deletion acceptance program 24a or the free block reservation program 24b when deletion or writing of any file is successful, or when a certain period of time elapses. It will be reviewed. When a certain period of time has passed since the block release control process 24c issued a deletion instruction, or when the free capacity of the file data area 26 decreases, the records in the queue table 24e are deleted in order from the oldest one. . For example, when a record related to a file whose file path is “/A/C/sample1.exe” is deleted from the queue table 24e, the i-node assignment to the file is canceled as shown in FIG. The disk block in which the data of the file is stored can be erased by overwriting, and the deletion of the file is completed.

In addition, when a restoration is instructed through the application 11 for a file whose file path immediately before the deletion instruction is “/A/C/sample1.exe” before the record in the queue table 24e is deleted, the restoration process 24d The presence / absence of the directory path “/ A / C /” is confirmed, and it is confirmed whether or not a record including “/A/C/sample1.exe” exists in the queue table 24e. Thereafter, a directory entry is generated so as to change from the state of FIG. 13 to the state of FIG. 4, and the directory tree is restored from the state of FIG. 14 to the state of FIG.

Configuration diagram of a computer according to the present embodiment Memory development diagram of file system program Explanatory drawing showing the configuration and data layout of the local file system Diagram showing detailed data layout Configuration diagram of deleted file management module The figure which shows an example of the table structure of a queue table Diagram showing the flow of delete acceptance processing Diagram showing the flow of securing free blocks Diagram showing the flow of block release control processing Diagram showing the flow of block release control processing Diagram showing the flow of restoration processing Figure showing an example of a directory tree A diagram showing the relationship between an i-node table and a file Figure showing an example of a directory tree A diagram showing the relationship between an i-node table and a file

Explanation of symbols

10 Computer 10c Storage Unit 10e Memory Unit 11 Application 12a Virtual File System 12b Local File System 12c Interface 12d Metadata 13 Device Driver 21 Access Control Module 22 Namespace Management Module 23 Disk Block Management Module 24 Delete File Management Module 24a Delete Acceptance Program 24b Free block reservation program 24c Block release control program 24d Restore program 24e Queue table 24f Free block count storage area 25 Metadata area 25a Inode table 26 File data area

Claims (5)

Manages the name space, which is a hierarchical classification system of files, manages the correspondence between file names and disk blocks in the data area where the data is stored, and controls file access from applications A file management device for
A first accepting unit for accepting deletion of any file from the application;
A first deletion unit that deletes the file name from the name space when the first reception unit receives the deletion of the file;
When the first accepting unit accepts the deletion of the file, a registration unit that registers a record in which the file identification information for specifying the file, the file path, and the date and time of deletion are associated with each other in the queue table;
When a predetermined event occurs, an extraction unit that extracts, from the queue table, a record whose retention period from the deletion date and time to a point in time exceeds a predetermined upper limit value;
When the extraction unit extracts a record whose retention period exceeds a predetermined upper limit, the association between the file specified by the file identification information in the record and the disk block in the data area is canceled. , The release part that releases the disk block that stores the data of that file,
A second deletion unit for deleting the record extracted by the extraction unit from the queue table;
A second accepting unit for accepting restoration of a deleted file and a file path from an operator;
A search unit that searches the queue table using a directory path in the file path as a search condition when the second reception unit receives a restoration of a deleted file and a file path from an operator; and
A file management device comprising: a restoration unit that generates a file specified by file identification information in a record in a directory specified by the directory path when the search unit detects a record . When the free space of the data area is below a predetermined lower limit value, the extraction unit further determines that the total data amount is within the data area from among records whose storage period does not exceed the predetermined upper limit value. The file management apparatus according to claim 1, wherein records of one or more files exceeding a difference between a free space and the predetermined lower limit value are extracted. The file management apparatus according to claim 1, wherein the event is a file write from an application. The file management apparatus according to claim 1, wherein the event is a lapse of a predetermined time for periodic execution. Manages the name space, which is a hierarchical classification system of files, manages the correspondence between file names and disk blocks in the data area where the data is stored, and controls file access from applications File management program for
Computer
First accepting means for accepting deletion of any file from the application;
When the first accepting unit accepts the deletion of the file, a first deleting unit that deletes the file name from the name space;
When the first accepting unit accepts the deletion of the file, a registering unit that registers a record in which the file identification information for specifying the file, the file path, and the deletion date are associated with each other in the queue table;
When a predetermined event occurs, an extraction unit that extracts from the queue table a record whose retention period from the deletion date and time to the time exceeds a predetermined upper limit value;
When the extraction unit extracts a record whose retention period exceeds a predetermined upper limit, the association between the file specified by the file identification information in the record and the disk block in the data area is canceled. A release means for releasing the disk block in which the file data is stored,
Second deletion means for deleting the record extracted by the extraction means from the queue table;
A second accepting means for accepting restoration of a deleted file and a file path from an operator;
When the second accepting unit accepts restoration of a deleted file and a file path from an operator, a search unit that searches the queue table using a directory path in the file path as a search condition; and
When the search means detects a record, it functions as a restoration means for generating a file specified by the file identification information in the record in a directory specified by the directory path. program.

Images