JP2003263351A - Increase suppressing method of file name retrieval time and computer system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To suppress increase of a processing time for file open, file creation or the like caused by increase of the number of files. <P>SOLUTION: When the number of files on a directory (hereafter referred to as DIR) is small, a file is created directly on the DIR. When creating a file with files over the number of n, subdirectories (hereafter referred to as SDIR) to the number of m having respectively a number as the name of the SDIR are created on the DIR, and file characteristic numbers created sequentially with file names are allocated to the files, and an SDIR wherein the numerical value of a part of the file characteristic number agrees with the numerical value of a part of the number of the SDIR name is selected from the SDIR's to the number of m, and the file is created on the selected SDIR. When the number of files on the SDIR exceeds n again, SDIR's to the number of m are created on the SDIR, and the files are allocated to the SDIR's according to the file characteristic numbers, to thereby prevent the number of files on one DIR and one SDIR from exceeding n. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for suppressing performance degradation due to file concentration in the same directory in a computer system having a hierarchical file system.

[0002]

2. Description of the Related Art A UNIX (registered trademark) file system generally used in UNIX (registered trademark) manages a group of a plurality of files called a directory by a hierarchical structure. Due to the hierarchical structure, the files are appropriately distributed, and the number of files in one directory does not have to increase. As shown in FIG. 7, each directory has entry information of a file name length 700, a file name 701, and a file unique number 702 called an inode number for each file. If the number of files in one directory is large, this entry information will be large, and in the process of the operating system that accesses the file,
The search process for obtaining the file unique number from the file name takes time. When accessing a file, the operating system uses the file unique number instead of the file name. As shown in FIG. 8, in the file name search process for obtaining the file unique number, first, the file entries of the directory for file search are sequentially checked (step 800),
If it matches the file name to be searched (step 801),
The file unique number in the entry is obtained and the process ends (step 802). If the file names do not match (step 801), if there is no next entry (step 8)
03) If the processing ends in error, if there is any, the processing is moved to the next entry (step 804) and this is repeated until the file name is found. In this method, if the number of files in the directory is a and the time required to check one entry is t, the time required to search the file name is a * t / 2 (* means multiplication) on average, and The larger the number a, the longer the processing time in proportion to the number of files.

[0003]

In the above conventional file system, the more the number of files on the directory increases, the slower the process for obtaining the file unique number from the file name associated with the opening and creation of the file, that is, the file name search. Become. The time required for this file name search generally increases in proportion to the number of files in a single directory. An object of the present invention is to prevent an increase in processing time such as file opening and file creation due to the increase in the number of files. Another object of the present invention is to change the processing method inside the operating system, and to improve the performance without changing the basic structure of the existing file system. Further, another object of the present invention is to allow a user who uses an operating system to obtain the effect of performance improvement without being aware of the change of the internal processing.

[0004]

In order to achieve the above object, the present invention provides a file name in a computer system having a processor device and an external storage device in which a file system having a hierarchical structure as a target of data input / output exists. This is a method of suppressing increase in search time, and when a file is stored in a directory in the file system, a file unique number linked to the file name of the file is created,
The created file unique number is assigned to the file, and when the number of file entries in the directory does not reach a predetermined value, the file is saved in the directory and the number of file entries in the directory is predetermined. A predetermined number of subdirectories are created on the directory and the file is stored in the subdirectory having a number that matches at least a part of the file unique number,
When the number of file entries in the directory exceeds a predetermined value, the file is saved in an already created subdirectory having a number matching at least a part of the file unique number.

[0005]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment for carrying out the present invention will be specifically described below with reference to the drawings. FIG. 2 is a diagram showing the system configuration of an embodiment of the present invention. In FIG. 2, reference numeral 1 is a computer system, which is a central processing unit 2,
It is composed of a main storage device 3 and an external storage device 6. A user program 4 and an operating system 5 are stored in the main storage device 3. A file system 7 is configured in the external storage device 6, and files 8 and the like are placed therein. Upon receiving a file creation request from the user program 4, the operating system 5 creates a file in the file system 7.

FIG. 1 shows an arrangement example of directories and files. File A exists in a file system with a hierarchical structure, and the absolute path name is "/ usr / local
/ fileA ”. The "/ usr / local" directory internally creates a multi-stage subdirectory, and files are also created in that subdirectory. That is, the files are distributed and arranged in subdirectories so that a large number of files are not collected in one directory internally. However, the user can see the file as "/ usr / local / fileA", and the subdirectory cannot be seen.

FIG. 3 shows the internal structure of a subdirectory directory. When the number of files is small,
Like the "AAA" file and the "BBB" file, the file is directly placed on the file entry of the directory as in the prior art. If the number of files increases and exceeds n (eg 100 or 1000) entries,
M (primary) subdirectories, ". $ 00" in the case of FIG.
100 files from "0" to ". $ 099" are created, and files are placed in the subdirectory. Further, when the number of files on each sub-directory exceeds n, similarly (secondary) m sub-directories, 100 in the case of FIG. 3, “. $ 000” to “. $ 099”, There is a structure in which this is repeated. Distributing files on these subdirectories prevents a large number of files from being concentrated on one directory.

FIG. 4 is a schematic diagram for explaining an example of a method for obtaining a file unique number linked to a file name from the file name. In the conventional file system, the file unique number is randomly assigned regardless of the file name. In the present invention, a subdirectory is selected using a part of the file unique number.
Therefore, it is necessary to know a part of the subdirectory, that is, the file unique number, from the file name. FIG. 4 shows an example of a method for creating a file unique number. Add the first 2 characters of the file name and divide by 100 to make the remainder number the last 2 digits of the file unique number, then add the next 2 characters and divide by 100 to make the remainder 4 digits of the last 4 digits the upper 2 The file unique number is determined by repeating this, such as changing to a digit. The last two digits specify the primary subdirectory, and the last four digits of the upper two digits specify the (secondary) subdirectory created under the (first) subdirectory. specify. Although the number of characters is 2 in this example, the number of characters may be another number, and the divisor is 100, but other divisors may be used. In FIG. 4, when the file name is ABCDE, the ASCII code is: A: 41 (hexadecimal number) = 65 (decimal number) B: 42 (hexadecimal number) = 66 (decimal number) C: 43 (hexadecimal number) = 67 (Decimal number) D: 44 (hexadecimal number) = 68 (decimal number) E: 45 (hexadecimal number) = 69 (decimal number), and the number obtained by adding the two characters A and B is 131, and 1
When 31 is divided by 100, the remainder is 31 and C, D
The number obtained by adding the two characters is 135, and the remainder when 135 is divided by 100 is 35. Assuming that the part not linked with the file name is 122, the file unique number is 1
It becomes 223531. However, since there is a case where the unique file unique number is already used, any number may be added to the higher digits. When a file whose file name is ABCDE shown in FIG. 4 is created and the number of file entries in the directory exceeds n, the directory (primary)
M subdirectories are created. In the case of FIG. 3, m =
100, and subdirectories ". $ 000" to ". $ 099"
At this time, since the file having the file name ABCDE has the file unique number 1223531, the file entry is made in the subdirectory “. $ 031” according to the numerical value “31” of the lower one to two digits.
The number of file entries in the subdirectory ". $ 031" is n
If the number exceeds the limit, m (secondary) subdirectories are further created under the subdirectory “. $ 031”. In the case of FIG. 3, m = 100, sub directories “. $ 000” to “. $ 099” are created, and the file name is A.
A file that is BCDE has a file unique number of 122.
Since it is 3531, a file entry is created in the subdirectory ". $ 035" according to the numerical value "35" of the lower 3-4 digits.

FIG. 5 is a flowchart showing an outline of the operation when the operating system creates a file. First, if a file with the same file name already exists (step 500), if access is permitted, then it is determined whether there is access permission (step 502). All the allocated blocks are released (step 503), the file is provided as a new file of size 0, and the file creation process ends. If there is no access permission (step 502), an error will occur. Steps 502, 503
Is the same as before.

If there is no file with the same file name in step 500, a file unique number linked to the file name is assigned (step 501) as shown in FIG. If the file unique number cannot be obtained due to lack of numbers, a free file unique number is assigned regardless of the file name (step 505). In this case, the file is created not in the sub-directory but in the normal directory (step 509), and the file creation process ends. If even a file unique number not linked with the file name is not assigned in step 507, an error occurs.

When a file unique number associated with the file name is assigned in step 504, the number of file entries in the directory is checked, and when the number of file entries is n (eg, 100 or 1000) or more, (Step 506), it is determined whether or not the subdirectory exists (step 508), and if it exists (for example, in the case of the example in FIG. 3, the subdirectory ". $").
000 ”to“. $ 099 ”exist), the process proceeds to step 511.
If it does not exist, create a subdirectory (for example, in the case of the example in Fig. 3, the subdirectory ". $ 00
"0" to ". $ 099" are created) (step 510) and the process proceeds to step 511. In step 511, a subdirectory is selected based on the file unique number and moved to that subdirectory. For example, in the case of the example in FIG. 3, the subdirectory is “. $ 000” to “. $ 099”, the file name is ABCDE in FIG. 4, and the file unique number assigned to this file name is 1223531. Therefore, select the subdirectory ". $ 031" according to the numerical value "31" of the last 1 to 2 digits, and select this subdirectory ". $ 031".
Move to. Next, the process returns to the process of checking the number of files (step 506) and the processes of steps 508, 510 and 511 are repeated. Finally, when a subdirectory with less than n files is reached in the directory (Step 50
6) Create a file in that subdirectory (step 509).

[0012] For example, the subdirectory ". $ 03
If the number of file entries on the subdirectory “. $ 031” exceeds n when moving to “1”, the (secondary) subdirectory already exists under the subdirectory “. $ 031”. If the number (m = 100 in the case of FIG. 3) exists, the file unique number is 1223531, so the file is moved to the subdirectory “. $ 035” according to the numerical value “35” of the lower 3 to 4 digits. If the (secondary) subdirectory does not exist, m = 10 in the case of FIG.
0, the subdirectories “. $ 000” to “. $ 099” are created, and the file unique number is 1223531. Therefore, the subdirectory “. $ 035” is moved according to the numerical value “35” of the lower 3 to 4 digits.

FIG. 6 is a flow chart showing the outline of the operation when the operating system opens a file. First, in order to move to the directory containing the file to be opened, the directory part of the specified pathname is analyzed and moved to the target directory (step 600). The file entry of the moved directory is searched, and the entry is examined until it matches the desired file name (step 601). It is judged whether or not the corresponding file name is found (step 602), and if the corresponding file name is not found even after checking all entries, it is checked whether or not there is a subdirectory (step 603). If an error occurs, a subdirectory is selected based on the file name if it exists, the subdirectory is moved to (step 606), and the operation for checking the entry (step 601) is repeated from this step. When the corresponding file name is found in the entry in step 602, the file unique number is acquired from the entry (step 604). It is determined whether or not the file has access permission (step 605). If there is no access permission, an error occurs. If there is access permission, the file information obtained from the file unique number is acquired (step 60).
7), the process ends.

Next, the time required for the file name search will be described. It is assumed that the total number of files in a certain directory is a, the time required to check one entry is t, the maximum number of files in one directory is n, and the number of subdirectories in one directory is m. In this case, in the conventional file name search, the average processing time required for the file name search is a
* T / 2. On the other hand, in this embodiment, if only the time for examining the directory entry is considered and the process of moving to the subdirectory is ignored, the average processing time required for the file name search is about log _ma * (n + m).
/ 2) t + n / 2t. Here, log _ma is the number of subdirectory hierarchies, (n + m / 2) t is the average time to search the subdirectory, and n / 2t is the average search time in the last subdirectory. For example, a = 1
1,000,000, t = 0.001 seconds, n = 100, m = 100
Then, while the conventional processing time is 500 seconds,
In the present invention, it is shortened to 0.5 seconds, which is 1 / 10,000.

[0015]

As described above, according to the present invention, the increase in the processing time such as file opening and file creation due to the increase in the number of files can be significantly suppressed, and the stability independent of the scale of the number of files in one directory can be achieved. The above performance can be realized, and the availability of the entire computer system can be improved. In addition, since the basic structure of the conventional file system is not changed, there is no need for the time required to transfer to the file system of the present invention, and existing utilities and recovery programs can be used as they are even in the event of a failure. Moreover, since the appearance of the file to the user is not changed by the present invention, the user can enjoy the performance improvement by the present invention without being aware of it. Moreover, the processing time required for the file name search can be greatly shortened.

[Brief description of drawings]

FIG. 1 is a diagram showing an arrangement example of directories and files.

FIG. 2 is a diagram showing a system configuration of an embodiment of the present invention.

FIG. 3 is a diagram showing an internal structure of a directory that is a subdirectory.

FIG. 4 is a schematic diagram for explaining a method of obtaining a file unique number associated with a file name from the file name.

FIG. 5 is a flowchart showing an operation outline when the operating system creates a file.

FIG. 6 is a flowchart showing an outline of operations performed when the operating system opens a file.

FIG. 7 is a diagram showing the contents of a conventional directory.

FIG. 8 is a diagram showing a flowchart of a conventional file name search process for obtaining a file unique number in a directory.

[Explanation of symbols]

1 calculator 2 Central processing unit 3 main memory 4 User program 5 Operating system 6 External storage device 7 file system 8 files

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Kei Imai             6-81 Onoe-cho, Naka-ku, Yokohama-shi, Kanagawa             Hitachi Software Engineering Stock Association             In-house (72) Inventor Kenzo Moriyama             5030 Totsuka Town, Totsuka Ward, Yokohama City, Kanagawa Prefecture             Ceremony company Hitachi Ltd. software division F-term (reference) 5B082 EA01

Claims (4)

[Claims] 1. A method for suppressing an increase in file name search time in a computer system having a processor device and an external storage device in which a file system having a hierarchical structure as a target of data input / output exists, the file system comprising: When a file is stored in a directory inside, a file unique number linked to the file name of the file is created, the created file unique number is assigned to the file, and the number of file entries in the directory reaches a predetermined value. If not, save the file in the directory, and if the number of file entries in the directory reaches a predetermined value, create a predetermined number of subdirectories in the directory, The file has a number that matches at least part of the file's unique number. When the number of file entries in the directory exceeds a predetermined value, the file is already created in a subdirectory having a number that matches at least a part of the file unique number. A method for suppressing an increase in the file name search time, which is characterized by storing in a file. 2. The method for suppressing an increase in file name search time according to claim 1, wherein the file unique number extracts at least one character set consisting of one or more characters from a plurality of characters forming the file name. Then, the code of each extracted character is converted into a decimal number, the decimal numbers of the respective characters that make up the set in each set are added, and the remaining numerical value obtained by dividing the addition result by a predetermined divisor is obtained. A method of suppressing an increase in file name search time, which is generated by combining the numerical value and a numerical value that does not work with a file name. 3. A computer system having a processor device and an external storage device in which a file system having a hierarchical structure as a target of data input / output exists, when a file is stored in a directory in the file system. A means for creating a file unique number linked to the file name of the file and assigning the created file unique number to the file; and, when the number of file entries in the directory does not reach a predetermined value, Means for storing on the directory; means for creating a predetermined number of subdirectories on the directory when the number of file entries in the directory reaches a predetermined value; and a file unique number for the file Of the subdata having a number matching at least a part of Means for storing in a directory, and when the number of file entries in the directory exceeds a predetermined value, the file is stored in an already created subdirectory having a number matching at least a part of the file unique number. A computer system having means for storing. 4. When a file is stored in a directory in the file system, a file unique number associated with the file name of the file is created, and the created file unique number is assigned to the file. When it is determined that the number of file entries in the file does not reach the predetermined value, the procedure of storing the file in the directory and the number of file entries in the directory are determined to reach the predetermined value. In that case, a procedure of creating a predetermined number of subdirectories on the directory, a procedure of storing the file in the subdirectory having a number that matches at least a part of the file unique number, and a file in the directory Judged that the number of entries exceeds a predetermined value When it is a program for executing the steps of storing in a subdirectory that has already been created with a number that matches the at least a portion of the file unique number of the file to the computer.