KR101932967B1 - Method for flushing metadata in journaling file system and apparatus for writing data using the method - Google Patents

Abstract

Provides a method for storing metadata in a journaling file system. A method for storing metadata in a journaling file system according to an embodiment of the present invention includes storing metadata in a journaling file system including metadata of a file change time, The method comprising: receiving journaling information including information of a time update interval which is an interval for updating the file modification time; Calculating the file modification time included in the current metadata to be stored in the storage device based on the time update interval included in the journaling information according to the modification of the file; And comparing the file modification time of each of the current metadata with the past metadata last stored in the storage device and selectively storing the current metadata in the storage device.

Description

[0001] METHOD FOR STORING METADATA IN JARNING FILE SYSTEM AND METHOD FOR RECORDING DATA [0002] BACKGROUND OF THE INVENTION [0003]

The present invention relates to a method of storing metadata in a storage device including information on changes of data in a journaling file system and a data recording apparatus to which the method is applied.

A journaling file system is a file system that keeps track of changes made to a specified area of the file system, before committing changes to the file system. This file system is characterized by the ability to quickly recover data using changes made to the journal, in the event of a system crash or power failure, with the result that data is less likely to be corrupted.

However, such a journaling file system is disadvantageous in that an overhead occurs in an application program frequently requesting synchronous writing, such as a database management system, and performance is degraded. That is, when an application requests a synchronous write (for example, fsync () function of Linux), the update data is stored in the storage device to reflect update data on the memory equally to the storage device. At this time, the metadata including information on the change of the update data is also stored in the journal area of the storage device, resulting in journaling overhead. In particular, when metadata is stored in a journal area, the metadata is stored in units of blocks, and a header block and a commit block are additionally written, so journaling overhead can be increased.

Accordingly, there is a need for a method of storing metadata and a data recording apparatus to which the method is applied, while reducing the journaling overhead while preventing the performance degradation of the file system while maintaining the advantages of the journaling file system.

The related art is disclosed in Korean Patent Registration No. 10-1258589 entitled " Information Storage Medium in which Data is Recorded According to a Journaling File System, Method and Apparatus for Recording / Restoring Data Using a Journaling File System, May 2, 2013).

An object of the present invention is to provide a method of storing metadata that minimizes degradation of a file system by reducing journaling overhead caused by repeated synchronous write commands while maintaining the advantages of a journaling file system, and a data recording apparatus to which the method is applied.

The problems to be solved by the present invention are not limited to the above-mentioned problem (s), and another problem (s) not mentioned can be clearly understood by those skilled in the art from the following description.

According to an aspect of the present invention, there is provided a method for storing metadata in a journaling file system, the method comprising: storing metadata in a journaling file system, ) In a storage device, the method comprising the steps of: receiving journaling information including information of a time update interval which is an interval for updating the file modification time; Calculating the file modification time included in the current metadata to be stored in the storage device based on the time update interval included in the journaling information according to the modification of the file; And comparing the file modification time of each of the current metadata with the past metadata last stored in the storage device and selectively storing the current metadata in the storage device.

Preferably, the step of calculating the file modification time includes: obtaining a current time indicating a current time from a data recording apparatus on which the journaling file system is mounted; And calculating the file modification time based on the current time and the time update interval.

Preferably, the calculating the file modification time may calculate the file modification time from a result obtained by subtracting a remainder obtained by dividing the current time by the time update interval at the current time.

Preferably, storing in the storage device may store the current metadata in the storage device only if the file modification time of the historical metadata and the file modification time of the current metadata do not match each other have.

Preferably, the step of determining the file modification time and the step of storing the current metadata in the storage device are performed only for at least one journaling target file selected from among a plurality of files included in the journaling file system .

Preferably, the file modification time may be a time at which at least one of contents and attributes of the file are changed.

Preferably, the file modification time is a file content modification time (mtime), which is a time at which the contents of the file are changed, or a file content modification time (ctime), at which at least one of the contents and attributes of the file is changed A method for storing metadata in a journaling file system.

Preferably, the time update interval is based on performance information on at least one of a performance of a central processing unit (CPU) of a data recording apparatus on which the journaling file system is mounted, a capacity of a memory (RAM) ≪ / RTI >

According to another aspect of the present invention, there is provided a data recording apparatus to which a method of storing metadata in a journaling file system according to an exemplary embodiment of the present invention is applied. The data recording apparatus includes metadata, including file change time information, A data recording apparatus to which a journaling file system for storing a file change time is applied, comprising: a receiving unit for receiving journaling information including information on a time update interval, which is an interval for updating the file change time; A calculating unit for calculating the file modification time included in the current metadata to be stored in the storage device based on the time update interval included in the journaling information according to the change of the file; And a storage unit for comparing the file change time of each of the current metadata with the past metadata stored last in the storage unit and selectively storing the current metadata in the storage device.

Preferably, the calculating unit may obtain a current time indicating a current time from a data recording apparatus on which the journaling file system is mounted, and calculate the file change time based on the current time and the time update interval .

Preferably, the calculating unit calculates the file change time based on the current time and the time update interval from the result of subtracting the remainder obtained by dividing the current time by the time update interval at the current time, The change time can be calculated.

The storage unit may store the current metadata in the storage device only when the file change time of the past metadata does not coincide with the file change time of the current metadata.

Preferably, the calculating unit and the storing unit may operate only on at least one journaling target file selected from among a plurality of files included in the journaling file system.

Preferably, the file modification time may be a time at which at least one of contents and attributes of the file are changed.

Preferably, the file change time may be a file content modification time (mtime), which is a time at which the content of the file is changed, or a file content modification time (ctime), at which at least one of the content and attributes of the file is changed.

The present invention has the effect of reducing the journaling overhead according to the repeated synchronous write command in the journaling file system, thereby minimizing the performance degradation of the file system.

1 is a flowchart illustrating a method for storing metadata in a journaling file system according to an embodiment of the present invention.
2 is a flowchart illustrating a method of calculating a file change time according to another embodiment of the present invention.
FIG. 3 is a view for explaining a data recording apparatus to which a metadata storage method in a journaling file system according to an embodiment of the present invention is applied.
FIG. 4 is a diagram for explaining a metadata storing method according to the related art.
5 is a diagram for explaining a metadata storing method according to an embodiment of the present invention.
6 is a view for explaining a metadata storing method according to another embodiment of the present invention.
7 is a pseudo code for explaining a file change time calculating method according to an embodiment of the present invention.
FIG. 8 is a view for explaining a result of a decrease in file write amount as the overhead decreases in the metadata storage method according to an embodiment of the present invention.
FIG. 9 is a diagram for explaining a result of improving file writing performance as the overhead decreases in the metadata storing method according to an embodiment of the present invention.
10 is a diagram for explaining a result of a decrease in file write amount of a DBMS according to a decrease in overhead in a metadata storing method according to an embodiment of the present invention.
FIG. 11 is a diagram for explaining a result of improving the file writing performance of the DBMS according to the reduction of overhead in the metadata storing method according to an embodiment of the present invention.

While the invention is susceptible to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawings and will herein be described in detail. It should be understood, however, that the invention is not intended to be limited to the particular embodiments, but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention. Like reference numerals are used for like elements in describing each drawing.

The terms first, second, A, B, etc. may be used to describe various elements, but the elements should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component. And / or < / RTI > includes any combination of a plurality of related listed items or any of a plurality of related listed items.

It is to be understood that when an element is referred to as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, . On the other hand, when an element is referred to as being "directly connected" or "directly connected" to another element, it should be understood that there are no other elements in between.

The terminology used in this application is used only to describe a specific embodiment and is not intended to limit the invention. The singular expressions include plural expressions unless the context clearly dictates otherwise. In the present application, the terms "comprises" or "having" and the like are used to specify that there is a feature, a number, a step, an operation, an element, a component or a combination thereof described in the specification, But do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or combinations thereof.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Terms such as those defined in commonly used dictionaries are to be interpreted as having a meaning consistent with the contextual meaning of the related art and are to be interpreted as either ideal or overly formal in the sense of the present application Do not.

Hereinafter, preferred embodiments according to the present invention will be described in detail with reference to the accompanying drawings.

1 is a flowchart illustrating a method for storing metadata in a journaling file system according to an embodiment of the present invention.

A journaling file system is a file system that writes changes to the journal area before committing changes to the file system. By doing so, when a system crash or a power problem occurs, it can be quickly recovered using the changes recorded in the journal area, and the possibility of data corruption can be reduced.

At this time, the metadata may include information such as the changed file size, block allocation information, ownership, file access time (atime), file content modification time (mtime), and file content modification time (ctime).

The ordered mode journaling mechanism of ext4, the most widely used journaling file system on linux or android systems, allows restoration if only the metadata is fully present, so that the changed data can be immediately And stores only the metadata in the journal area.

At this time, when the synchronous write function (fsync (), fdatasync ()) or the trigger of 5 seconds interval is called, the data recording device can flush the changed data to the storage device to maintain the continuity of the data . More specifically, when a data recording apparatus executes a write command for a data block of 4 KB size, a 4 KB data block is written to a memory buffer and an inode, a timestamp, and an inode bitmap And performs a separate check (dirty mark) when the metadata of the metadata is changed. When fsync () is called, the data recording apparatus can flush the 4 KB data blocks and the metadata to the storage device. Further, the data recording apparatus can additionally record a 4-KB journaling header and a 4-KB journal reflection signal while recording the meta data.

Therefore, in order for the data recording apparatus to record the data block of 4 KB in the storage device, an overhead is required to record data of at least 16 KB in the storage device. That is, the data recording apparatus should record a data block of 4 KB, a header of 4 KB, meta data of 4 KB, and a journal reflection signal of 4 KB. This overhead becomes more noticeable when an application program (e.g., DBMS) in which synchronization writing is frequently generated is used.

In step S110, the data recording apparatus receives journaling information including information of a time update interval, which is an interval for updating the file change time.

Here, the file change time means a time at which a change is made to a certain file. That is, the file change time may refer to the time at which a change occurred in a file included in the journaling file system.

On the other hand, the time update interval means a time interval at which the file change time is updated. For example, if the time update interval is 100 ms, the file change time can be updated and increased in units of 100 ms. Thus, the file modification time may be the same as the result of performing a truncation operation in units of 100 ms in comparison with the modification time of the correct file.

More specifically, when the file change is made at time t ₀ divided by 100 ms and when the file change is made at (t ₀ + 10 ms), the result of the truncation operation is that both cases are the same file The change time t ₀ may be included in the metadata.

In another embodiment, the file modification time may be the time at which at least one of the content and attributes of the file has changed.

On the other hand, the file change time refers to the time when the file included in the journaling file system is changed, i) the contents of the file are changed, ii) the attributes (permission, owner, etc.) of the file are changed, and iii) It can be time for case.

In another embodiment, the file modification time may be a file content modification time (mtime), which is the time at which the contents of the file are changed, or a file contents modification time (ctime), at which at least one of the contents and attributes of the file is changed.

In this case, the data modification time (mtime) refers to the time when the contents of the file are actually changed after the file is created. For example, when a file is copied or moved, mtime may not be updated because the actual contents of the file are not changed.

On the other hand, the file change time (inode change time, ctime) is the time when the information of the file inode is changed. For example, if file ownership, permissions, block allocation information, and timestamp are changed, ctime can be updated. Also, ctime can occur together with update of atime and mtime which change the timestamp, so that when mtime is updated, ctime can also be updated.

In another embodiment, the time update interval may be based on performance information for at least one of performance of a central processing unit (CPU) of a data recording apparatus equipped with a journaling file system, capacity of a memory (RAM) ≪ / RTI >

For example, the performance degradation of the file system due to the overhead of the above-described journaling file system may occur more depending on the performance of the data recording apparatus. That is, as the performance of the CPU is low, the memory capacity is insufficient, and the input / output speed of the storage device is slow, the overhead of the journaling file system may affect the performance of the data recording device more.

Therefore, the time update interval can be set to be larger as the performance of the CPU is low, the memory capacity is insufficient, or the input / output speed of the storage device is slower, thereby reducing the overhead of the journaling file system. At this time, the influence of the time update interval on the overhead of the journaling file system will be described later in detail in step S130.

In step S120, the data recording apparatus calculates a file modification time included in the current metadata to be stored in the storage device based on the time update interval included in the journaling information, in accordance with the change of the file.

For example, when the time update interval is 100 ms, the data recording apparatus can calculate the file change time included in the current meta data by subtracting the file change time in units of 100 ms.

A specific method of calculating the file change time will be described later in detail with reference to FIG.

Finally, in step S130, the data recording device compares the file change time of each of the current metadata with the past metadata last stored in the storage device, and selectively stores the current metadata in the storage device.

That is, the data recording apparatus compares the file change time of the past metadata with the file change time of the current metadata, and may selectively store the current metadata in the storage device only when the comparison result satisfies a predetermined condition .

For example, if the data recording apparatus has generated two or more metadata including a file change time that is within one time update interval (e.g., 100 ms), only one of the two or more metadata having the same file change time Only the meta data of the user can be stored in the storage device. In other words, the data recording apparatus may not record the remaining metadata except for the first metadata in the generation order and the generation time in the storage device. By doing so, the data recording apparatus can obtain the effect of reducing the overhead due to the repeated writing of metadata in the journaling file system while maintaining the journaling function.

Referring to FIGS. 4 to 6, a process of reducing the overhead will be described.

4 is a view for explaining a method of storing metadata according to the related art. The metadata 401 includes data 401, 402, 403, and 404 of a memory located in a memory and metadata 411 and 412 413 and 414 are all stored in the storage in accordance with the write command 421, 422, 423 and 424 of the data recording apparatus and the synchronous write command 431, 432, 433 and 434 such as fsync Lt; / RTI >

5 according to an embodiment of the present invention includes a change 501 of data located in a memory and a corresponding metadata 511 512 513 and 514 When data is written to a storage according to synchronous write commands 531, 532, 533 and 534 such as write commands 521, 522, 523 and 524 of the data recording apparatus and fsync 512, 514) may be recorded in the storage device.

The reason is that only 511 of metadata 511 and 512 having the same file change time t ₁ are recorded in the storage device and only 513 of metadata 513 and 514 having the same file change time t ₃ are stored Is recorded in the device. This is because the time update interval has increased in FIG. 5 as compared with FIG.

Finally, FIG. 6 shows another embodiment of the present invention in which the timer interrupt interval itself of the data recording apparatus is adjusted. In this case, only 611 among metadata 611 and 612 having the same file modification time t ₁ are recorded in the storage device, and 613 and metadata 613 having the same file modification time t ₃ , The same effect as in Fig. 5 can be obtained. However, the timer interrupt of the system generally affects other functions of the system, such as other scheduling, and thus may cause other problems.

The effect of reducing the overhead will be described with reference to FIGS. 8 to 11. FIG.

8 is a diagram for explaining an effect of reducing a write amount of a metadata storing method according to an embodiment of the present invention. At this time, sequential write (SW), random write (RW), Base (default time update interval = timer interrupt interval), CG (100 ms time update interval), Ideal (infinity time update interval) And (b) desktop PC, we can confirm that the amount of writing decreases in order of CG and Ide compared with Base. Specifically, when using CG on a mobile device, the amount of metadata to be written is reduced by up to 87%, and on a desktop PC environment, it is seen that the write amount is much larger than a mobile device. On the other hand, the more effective in the desktop PC environment is that the basic timer interrupt interval differs between PC (4 ms) and mobile (10 ms).

Referring to FIG. 9, it can be seen that (a) in the mobile deivce and (b) in the desktop PC, the performance is the best in Ideal, the next is CG, and the next is Base. Also, CG is very close to Ideal's performance.

10 is a diagram for explaining an effect of reducing a write amount in a DBMS by a metadata storing method according to an embodiment of the present invention. In this case, PERSIST and WAL modes are DBMS SQLite mode and OLTP DBMS MySQL mode. Similar to FIG. 8, it can be seen that the amount of writing decreases in the order of Base, CG, Ideal.

Referring to FIG. 11, it can be seen that (a) in the case of mobile deivce and (b) in the case of desktop PC, the performance is the best in Ideal, the next is CG, and the next is Base. Also, CG is very close to Ideal's performance.

In another embodiment, when the data recording device stores the current metadata in its storage device, only when the file change time of the past metadata and the file change time of the current metadata do not coincide with each other, Lt; / RTI >

For example, if the data recording device matches the file change time of the past metadata with the file change time of the current metadata, the current metadata may not be stored in the storage device. In contrast, when the file change time of the past metadata does not match the file change time of the current metadata, the current metadata can be stored in the storage device.

In another embodiment, steps S120 and S130 may be performed only for at least one journaling target file.

At this time, the file to be journalled may be a file in which a write command is frequently executed. For example, files of SQLite, which is a DBMS commonly used in an Android system, and files of MySQL, which is a DBMS commonly used in a PC and a server system, can be selected as a journaling target file, and when the data recording apparatus changes its journaling target file Steps S120 and S130 may be performed.

That is, by performing the journaling operation only on the selected journaling target file, the data recording apparatus improves the write performance of most files that do not require the journaling function, while the journaling target file that requires the journaling function is journaling It can have the advantage of maintaining the function.

As described above, the metadata storage method in the journaling file system according to an embodiment of the present invention can reduce the journaling overhead according to the repeated synchronization write command, thereby minimizing the write performance degradation of the file system.

2 is a diagram for explaining a method of calculating a file change time according to another embodiment of the present invention.

In step S210, the data recording apparatus acquires the current time indicating the current time from the data recording apparatus on which the journaling file system is mounted.

For example, the data recording apparatus can acquire information on the current time from the operating system (OS) included therein, which represents the current time. At this time, the current time may be a relative time based on January 1, 1970, which is the reference time of UNIX or LINUX.

In step S220, the data recording apparatus calculates the file change time based on the current time and the time update interval.

For example, the data recording apparatus may calculate a value obtained by subtracting the remainder obtained by dividing the current time by the unit of time update interval by a predetermined criterion from the current time as a file change time.

In another embodiment, when the data recording apparatus calculates the file change time, the file change time can be calculated from the result obtained by subtracting the remainder obtained by dividing the current time from the current time by the time update interval.

That is, referring to FIG. 7, in the third line, the file change time (now) can be calculated from a value obtained by subtracting the remainder obtained by dividing the current time (fs_time) by the time update interval (interval) at the current time (fs_time).

Also, in the fifth to seventh lines, a separate check (dirty is indicated by 1) is performed only when the file change time now differs from the file change time (mtime) last updated.

Also, in the eighth through tenth lines, only when the check is set to 1, the latest updated file change time (mtime) is updated to the newly calculated file change time (now).

FIG. 3 is a view for explaining a data recording apparatus to which a metadata storage method in a journaling file system according to an embodiment of the present invention is applied.

3, a data recording apparatus 300 to which a metadata storage method in a journaling file system according to an embodiment of the present invention is applied includes a receiving unit 310, a calculating unit 320, and a storing unit 330 do.

Meanwhile, the data recording apparatus 300 may be mounted on a smart phone, a wearable device, a tablet PC, a notebook PC, or the like.

The receiving unit 310 receives journaling information including information of a time update interval, which is an interval for updating the file change time.

In another embodiment, the file modification time may be the time at which at least one of the content and attributes of the file has changed.

In another embodiment, the file modification time may be a file content modification time (mtime), which is the time at which the contents of the file are changed, or a file contents modification time (ctime), at which at least one of the contents and attributes of the file is changed.

The calculating unit 320 calculates the file change time included in the current metadata to be stored in the storage device based on the time update interval included in the journaling information, according to the change of the file.

In another embodiment, the calculating unit 320 can obtain the current time indicating the current time from the data recording apparatus on which the journaling file system is mounted, and calculate the file change time based on the current time and the time update interval have.

In yet another embodiment, the calculating unit 320 calculates the file change time based on the current time and the time update interval, subtracting the remainder obtained by dividing the current time by the time update interval at the present time The file change time can be calculated.

The storage unit 330 compares the file change time of each of the current metadata and the past metadata stored last in the storage device, and selectively stores the current metadata in the storage device.

In another embodiment, the storage unit 330 may store the current metadata in the storage device only when the file change time of the past metadata does not coincide with the file change time of the current metadata.

In yet another embodiment, the calculating unit 320 and the storing unit 330 may operate only for at least one journaling target file selected from a plurality of files included in the journaling file system.

The above-described embodiments of the present invention can be embodied in a general-purpose digital computer that can be embodied as a program that can be executed by a computer and operates the program using a computer-readable recording medium.

The computer readable recording medium includes a magnetic storage medium (e.g., ROM, floppy disk, hard disk, etc.), optical reading medium (e.g., CD ROM, DVD, etc.).

The present invention has been described with reference to the preferred embodiments. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims. Therefore, the disclosed embodiments should be considered in an illustrative rather than a restrictive sense. The scope of the present invention is defined by the appended claims rather than by the foregoing description, and all differences within the scope of equivalents thereof should be construed as being included in the present invention.

Claims (15)

A method for storing metadata in a storage device, the metadata including information on a file change time, which is a time at which a file is changed in a journaling file system,
Receiving journaling information including information of a time update interval which is an interval for updating the file change time;
Calculating the file modification time included in the current metadata to be stored in the storage device based on the time update interval included in the journaling information according to the modification of the file; And
Comparing the file change time of each of the past metadata stored last in the storage with the file change time of each of the current metadata and if the file change time of the past metadata does not coincide with the file change time of the current metadata And storing the current metadata in the storage device. ≪ Desc / Clms Page number 20 > The method according to claim 1,
The step of calculating the file change time
Obtaining a current time indicating a current time from a data recording apparatus on which the journaling file system is mounted; And
Calculating the file modification time based on the current time and the time update interval
And storing the meta data in the journaling file system. 3. The method of claim 2,
The step of calculating the file change time
Wherein the file change time is calculated from a result obtained by subtracting a remainder obtained by dividing the current time by the time update interval at the current time. delete The method according to claim 1,
Determining the file modification time and storing the current metadata in the storage device
Wherein the at least one journaling target file is selected from a plurality of files included in the journaling file system. The method according to claim 1,
The file modification time
Wherein at least one of contents and attributes of the file is changed. The method according to claim 6,
The file modification time
Wherein the file content modification time (ctime) is a time at which at least one of contents and attributes of the file are changed, the file content modification time (ctime) being a time at which the content of the file is changed . The method according to claim 1,
The time update interval
The performance of at least one of a performance of a central processing unit (CPU) of a data recording apparatus on which the journaling file system is mounted, a capacity of a memory (RAM), and an input / output speed of the storage device, A method for storing metadata in a system. A data recording apparatus to which a journaling file system is applied that stores metadata including information of a file change time, which is a time at which a file is changed, in a storage device,
A receiving unit for receiving journaling information including information of a time update interval which is an interval for updating the file change time;
A calculating unit for calculating the file modification time included in the current metadata to be stored in the storage device based on the time update interval included in the journaling information according to the change of the file; And
Comparing the past metadata stored last in the storage device with the file modification time of each of the current metadata,
And storing the current metadata in the storage device when the file change time of the past metadata does not match the file change time of the current metadata,
Wherein the journaling file system is a data recording apparatus to which a journaling file system is applied. 10. The method of claim 9,
The calculating unit
Acquiring a current time indicating a current time from a data recording apparatus equipped with the journaling file system,
Wherein the file change time is calculated based on the current time and the time update interval. 11. The method of claim 10,
The calculating unit
When calculating the file change time based on the current time and the time update interval,
Wherein the file change time is calculated from a result obtained by subtracting a remainder obtained by dividing the current time by the time update interval at the current time. delete 10. The method of claim 9,
The calculation unit and the storage unit
Wherein the journaling file system is operated only for at least one journaling target file selected from a plurality of files included in the journaling file system. 10. The method of claim 9,
The file modification time
Wherein at least one of contents and attributes of the file is changed. 15. The method of claim 14,
The file modification time
Wherein the file content modification time (ctime) is a time at which at least one of contents and attributes of the file is changed, the file content modification time (mtime) being a time at which the contents of the file are changed.

Images