JP6276604B2 - Information processing apparatus and control method thereof - Google Patents

Description

  The present invention relates to an information processing apparatus for repairing a file system and a control method thereof.

  In an information processing apparatus including a nonvolatile memory, if the writing process is interrupted due to a power interruption or the like during the writing process to the nonvolatile memory, the consistency in the file system may be lost.

  The file system is one of functions provided by an operating system (OS) that operates on the information processing apparatus, and is for managing data held by the information processing apparatus. The file system holds not only data but also information (metadata) for managing data such as data storage position information on the nonvolatile memory, file name, and creation / change time.

  If the writing process to the file system is interrupted in the middle due to power failure or failure of the information processing device, and the consistency between the data storage location information described in the metadata and the actual data storage location is lost, etc. Access to the data may be impossible. In order to prevent such a problem from occurring, there is a file system (journaling file system) equipped with a mechanism called journaling.

  In the journaling file system, metadata is written in an area called a journal before actual data is written. As a result, even if data inconsistency occurs, by referencing the journal and forcing it back to the metadata before the inconsistency occurred, the file system consistency is maintained and access to the actual data is lost. This makes it possible to prevent such a situation (see, for example, Patent Document 1).

  However, even with a journaling file system, it is impossible to prevent inconsistencies caused by non-volatile memory failure or metadata rewriting due to system runaway of the device. Means are needed. For this reason, there is provided a repair method when inconsistency occurs on the file system (see, for example, Patent Document 2).

  In order to realize an equivalent function even in Linux (registered trademark), a command called fsck is prepared for correcting a portion where a file system inconsistency has occurred. fsck has an interactive mode and a batch application mode. In the interactive mode, the user is asked whether or not consistency occurs at each location where inconsistencies occur. In the batch application mode, all inconsistencies occur. Attempt to correct the location. If the file system cannot be corrected by fsck, the data may be deleted to maintain the integrity of the file system, and normal data may be destroyed. It is recommended to keep it.

JP 2005-149248 A JP-A-1-270156

  As described above, in order to return the file system to the state before the repair, it is necessary to save a backup of the file system. However, in the above conventional technique, since the entire area of the target file system has to be saved, the required backup capacity increases and the backup time also increases.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to reduce the backup capacity necessary for restoring the file system.

As a means for solving the above problems, an information processing apparatus according to the present invention has the following configuration. That is, an information processing apparatus,
A detection means for detecting the necessity of repair in the file system;
Backup means for creating a backup of a part of data constituting the file system when the necessity of repairing the file system is detected;
A repair means for repairing the file system after creating the backup;
Restart means for restarting the system of the information processing apparatus after the file system is repaired;
Restore means for restoring the file system before the restoration and after the detection from the backup when the system is not restarted normally .

  More specifically, after the file system is restored, a restart unit for restarting the system of the information processing apparatus, and the file system is restored from the backup when the system is not restarted normally. And restoring means.

  According to the present invention, it is possible to reduce the backup capacity required for file system repair.

Schematic which shows the structure of the information processing apparatus which concerns on one Embodiment of this invention. The figure which shows the hardware structural example in information processing apparatus Flow chart showing file system repair processing Diagram showing the configuration of the file system Flow chart showing file system repair processing Diagram showing an example of the file system configuration before modification Figure showing a configuration example of file system backup data

  Hereinafter, embodiments according to the present invention will be described with reference to the drawings. The following embodiments do not limit the present invention related to the scope of claims, and all the combinations of features described in the present embodiments are not necessarily essential to the solution means of the present invention. Absent.

<First Embodiment>
Apparatus Configuration FIG. 1 shows a basic functional configuration particularly related to file system repair in an information processing apparatus for file system repair according to the present invention. In the figure, reference numeral 100 denotes an information processing apparatus implementing the present invention.

  In the information processing apparatus 100, reference numeral 101 denotes an abnormality detection unit that confirms whether or not the file system 110 needs to be repaired. The abnormality detection unit 101 investigates whether a mismatch occurs in the file system 110 at the time of system startup due to power-on or the like. When inconsistency occurs, it is determined that there is a need for repair, and the fact is notified to each of the file system information backup unit 102, the directory information backup unit 103, and the file information backup unit 104.

  Reference numeral 102 denotes a file system information backup unit that stores system information of the file system 110 in a secure memory. Reference numeral 103 denotes a directory information backup unit that saves directory information existing in the file system 110 on a secure memory. A file information backup unit 104 saves file information existing in the file system 110 on a secure memory. Each of the backup units 102, 103, and 104 operates when the abnormality detection unit 101 is notified of the necessity of repair in the file system 110, and obtains target information from the file system 110 to generate backup data 109. To do. Then, after the backup data 109 is created, the restoration unit 105 is notified that the backup process has been completed.

  Reference numeral 105 denotes a repair unit that attempts to repair an inconsistency occurring in the file system 110. After the repair process is completed, the repair unit 105 notifies the restart unit 106 that the repair process has ended. Reference numeral 106 denotes a restart unit that restarts the system in the information processing apparatus 100. After the system is restarted, the restart result detection unit 107 is notified that the restart process has been completed. Reference numeral 107 denotes an activation result detection unit that confirms whether the system is normally activated and operating. When a system activation failure is detected, the activation result detection unit 107 notifies the backup restoration unit 108 of that fact. A backup restoration unit 108 restores the file system 110 to a state before the restoration process by the restoration unit 105 using the backup data 109.

  The above is the basic configuration related to the file system restoration in the information processing apparatus of this embodiment. Hereinafter, regarding the system configuration and the file system restoration processing procedure in this embodiment, it is assumed that the system is Linux (registered trademark), which is a UNIX (registered trademark) operating system (OS), and the file system is ext3 format. Will be described in detail.

  FIG. 2 is a block diagram illustrating a part of the hardware configuration of the information processing apparatus according to the present embodiment. In the figure, a CPU 201 controls the entire information processing apparatus 200 by executing an OS and a program. Note that the file system restoration processing in this embodiment is provided by the CPU 201 executing a program. The volatile memory 202 is a random access memory (RAM) that temporarily stores programs and data, and is also referred to as a system memory. The non-volatile memory 203 is a (non-volatile) storage medium such as a hard disk drive (HDD) in which data and programs can be recorded even after power supply is cut off. The file system 110 and backup data 109 of this embodiment are held in the nonvolatile memory 203.

Outline of File System Repair Processing Hereinafter, the repair processing of the file system 110 at the time of system startup (or restart) will be described using the flowchart of FIG.

  First, when the system is activated in S300, the abnormality detection unit 101 confirms whether or not the file system 110 needs to be repaired in S301. Specifically, it is confirmed whether or not inconsistency occurs in the metadata (file management information) of the file system 110, and if there is no inconsistency, the processing is terminated as it is. If inconsistency occurs, the file system information backup unit 102, the directory information backup unit 103, and the file information backup unit 104 perform backup processing of the file system 110 in S302. When an inconsistency in the file system 110 is detected, conventionally, a complete copy of the file system 110 is stored as a backup. However, in the present embodiment, only the portion to be repaired is used as the backup data 109. It is characterized by storing. Details of this backup processing will be described later.

  After the creation of the backup is completed, in S303, the restoration unit 105 uses the fsck command to repair the inconsistency of the file system 110. Details of the repair process using the fsck command will be described later.

  After the file system 110 is repaired, the restart unit 106 restarts the system in S304. In S305, the activation result detection unit 107 confirms whether or not the system has been activated normally. Specifically, a watchdog timer is set at the start of system startup or restart, and the watchdog timer is canceled when normal startup of the system is confirmed. Therefore, when the system cannot be started normally, the watchdog timer is not canceled and the time is up after a predetermined time has elapsed. If the system can be started normally, the processing in FIG. 3 is terminated as it is. If the system has not been started normally, the process proceeds to S306, and the backup restoration processing of the file system 110 is performed.

  In S306, the backup restoration unit 108 uses the backup data 109 saved in S302 to return the file system 110 to the state before the restoration processing in S303. This restoration was realized by completely replacing the file system that had been repaired in the past with a file system saved as a backup. In this embodiment, the file system 110 is restored only by returning the metadata stored as a backup to an appropriate location on the file system 110. By this backup restoration process, even if the file system 110 has been changed in some way by the repair process in S302, it can be reliably returned to the state before the repair.

  Each detection process in S301 and S305 is performed before the OS is started, for example, after the boot loader is started. However, these detection processes can be incorporated into the OS.

  Further, after the file system is restored in S306, a detailed investigation of the content of the failure that has occurred in the file system may be performed, or other repair processing may be attempted. Further, the system may be started up with the file system abnormality intact.

Ext3 File System Here, the structure of the ext3 file system held and used as the file system 110 in the nonvolatile memory 203 in this embodiment will be described with reference to FIG. In the figure, reference numeral 400 denotes a boot block, which holds data necessary for starting the system. 401 and 402 are a plurality of block groups constituting the ext3 file system.

  Next, a detailed configuration of the Nth block group 402 (hereinafter simply referred to as a block group) will be described. The super block 500 exists only in one block in the block group, and represents various information regarding the configuration of the file system. One file system holds one super block, that is, all block groups hold super blocks having the same value.

  The group descriptor 501 holds block group management information, and there are as many block groups as exist in the file system. The data block bitmap 502 manages the free space on the data block, and only one block exists in the block group. The i-node bitmap 503 is a bitmap for managing the vacancy of the i-node, and there is only one block in the block group. An i-node table 504 is a table that holds i-nodes. A plurality of i-node tables 504 exist in a block group, and 32 i-nodes are stored for each table.

  The above-described parts from the super block 500 to the i-node table 504 represent file system metadata. The metadata is file management information, and includes file system information indicating information of the file system itself and file information indicating information for each file. Data such as files and directories actually managed by the file system is stored in the data block 505. In ext3, a directory is also handled as one type of file, and a directory entry for managing a table of files included in the directory is recorded in the data portion of the file.

  Here, i-nodes stored in the i-node table 504 will be described. The i-node handles file configuration data. Specifically, file type (regular file, directory, symbolic link), file size, access time, creation time, modification time, deletion time, block size, pointer to data block 505, reference count indicating the number of hard links Hold, etc.

  Next, the directory entry will be described. The directory entry is used to manage the directory information in the file system as a table, and is created on the data block 505. Specifically, the i-node number, record length, file type, file name length, file name, etc. are held for each entry.

● File system repair processing (S302)
Hereinafter, the restoration process of the ext3 file system by fsck executed by the restoration unit 105 in S302 will be described with reference to the flowchart of FIG.

  First, in step S600, the i-node table is checked to check whether there is any inconsistency.

  In step S601, the directory entry is inspected to confirm whether or not inconsistency occurs in the value held in the directory entry. Next, in S602, the connectivity (connectivity) between the directory entries is confirmed, and it is confirmed whether any inconsistency has occurred.

  In step S603, the reference count indicating the number of hard links held by the i-node is inspected to confirm whether inconsistency has occurred. In S604, it is confirmed whether or not there is a mismatch between the block on the disk and the i-node bitmap. Of course, according to fsck, if any inconsistency occurs after each of the above confirmations, the repair processing of the occurrence location is immediately performed.

  According to FIG. 5, it can be seen that fsck is trying to repair only metadata and directory entries. Therefore, in this embodiment, only metadata and directory entries to be repaired are stored as backup targets.

Backup processing (S301) / Backup restoration processing (S305)
Hereinafter, the backup process in S301 and the backup restoration process in S305 will be described.

  FIG. 6 shows a configuration example of the ext3 file system 110 before backup, that is, the backup target. 700 represents the entire ext3 file system 110, and 701, 702, 703, and 704 represent metadata for each block group. Reference numerals 705, 706, 707, 708, and 709 denote directory entries existing in the file system 700.

  FIG. 7 shows a configuration example of the backup data 109 of the file system 700 shown in FIG. Reference numeral 800 denotes the entire backup data 109. Reference numeral 801 denotes storage destination data indicating the storage destination of the metadata 701 on the file system 700, and indicates where the metadata 701 exists on the file system 700. When the backup restoration process is executed in S305, the metadata 701 can be returned to the correct position on the file system 700 by using the storage destination data 801.

  According to FIG. 7, it can be seen that the backup data 800 holds the metadata 701, 702, 703, and 704 shown in FIG. Similarly, the directory entries 705, 706, 707, 708, and 709 are stored with storage destination data 805, 806, 807, 808, and 809 indicating storage destinations on the file system 700. With these storage destination data, each metadata and directory entry to be restored by fsck is restored to the correct position on the file system 700. The metadata 701, 702, 703, and 704 and the storage destination data 801, 802, 803, and 804 are created by the file system information backup unit 102 and the file information backup unit 104. Directory entries 705, 706, 707, 708, and 709 and their storage destination data 805, 806, 807, 808, and 809 are created by the directory information backup unit 103.

  As described above, according to the present embodiment, only the metadata and directory entry information to be repaired by fsck are backed up, and a repair failure at the time of system restart is detected to bring the file system to the state before repair. Restore. As a result, it is possible to maintain the file system by preventing the loss of data due to the repair failure while reducing the backup capacity required for the safe repair of the file system. Therefore, the file system can be safely repaired by fsck even on an embedded device having a strong memory constraint.

<Other embodiments>
The present invention can also be realized by executing the following processing. That is, software (computer-readable program) that implements the functions of the above-described embodiments is supplied to a system or apparatus via a network or various storage media. The system or apparatus computer (or CPU, MPU, or the like) reads out and executes the program.

Claims (13)

An information processing apparatus,
A detection means for detecting the necessity of repair in the file system;
Backup means for creating a backup of a part of data constituting the file system when the necessity of repairing the file system is detected;
A repair means for repairing the file system after creating the backup;
Restart means for restarting the system of the information processing apparatus after the file system is repaired;
Restoration means for restoring a file system before the restoration and after the detection from the backup when the system is not restarted normally. apparatus.   2. The information processing apparatus according to claim 1, wherein a part of the data constituting the file system is a part to be repaired by the repair unit. It said backup means, the system information of the file system, directory information, and file information, the information processing apparatus according to claim 1 or 2, characterized in that to create the backup. Said part of the data constituting the file system, according to claim 1 to 3, wherein the metadata indicating the system information and file information in the file system, and a directory entry showing information of files contained in the directory The information processing apparatus according to any one of the above. The information processing apparatus according to claim 4 , wherein the repair unit sets the metadata and the directory entry as repair targets. The information processing apparatus according to claim 5 , wherein the repair unit repairs the file system using an fsck command. The information processing apparatus according to claim 6 , wherein the file system is ext3 format. The information processing apparatus further comprises start result detection means for detecting whether or not the restart of the system has been normally performed based on a time required for restarting the system of the information processing apparatus by the restart means. Item 4. The information processing apparatus according to Item 1 . It said detection means, the information processing apparatus according to any one of claims 1 to 8, characterized in that to detect the inconsistency of the metadata in the file system. It said detection means, the information processing apparatus according to any one of claims 1 to 9, characterized in that said sensing at system startup of the information processing apparatus. A control method for an information processing apparatus having detection means, backup means, repair means , restart means, and restore means ,
The detection means detects the necessity of repair in the file system;
When the backup means detects that the file system needs to be repaired, it creates a backup of a part of the data constituting the file system,
Said repair means, after the creation end the backup, have line repair of the file system,
The restarting means restarts the system of the information processing apparatus after repairing the file system,
The information processing apparatus control method , wherein the restoration unit restores the file system before the restoration and after the detection from the backup when the system is not restarted normally. . A program for causing a computer device to function as each unit of the information processing device according to any one of claims 1 to 10 , when the computer device is executed. A computer-readable storage medium storing the program according to claim 12 .

Images