CN105094695B - A kind of storage method and system - Google Patents
A kind of storage method and system Download PDFInfo
- Publication number
- CN105094695B CN105094695B CN201510370132.8A CN201510370132A CN105094695B CN 105094695 B CN105094695 B CN 105094695B CN 201510370132 A CN201510370132 A CN 201510370132A CN 105094695 B CN105094695 B CN 105094695B
- Authority
- CN
- China
- Prior art keywords
- file
- operated
- dbstore
- storage
- inode
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000003860 storage Methods 0.000 title claims abstract description 132
- 238000000034 method Methods 0.000 title claims abstract description 24
- 238000013507 mapping Methods 0.000 claims description 36
- 238000012545 processing Methods 0.000 claims description 17
- 230000009471 action Effects 0.000 claims description 14
- 238000012163 sequencing technique Methods 0.000 claims description 10
- 230000003139 buffering effect Effects 0.000 claims description 6
- 230000004044 response Effects 0.000 claims description 6
- 230000000875 corresponding effect Effects 0.000 description 36
- 230000014759 maintenance of location Effects 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000000151 deposition Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 230000007257 malfunction Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a kind of storage method and systems, including:Before application layer operates file, the scheduled storage class of this document is obtained by preset index node inode lists.This document is handled accordingly according to this document corresponding storage class.Scheme through the invention can avoid the problem that disk frequently addresses caused by a large amount of random small documents write-ins, degraded performance, improve write efficiency.
Description
Technical field
The present invention relates to field of distributed storage, and in particular to a kind of storage method and system.
Background technology
With the rapid development of network application, network information data amount is increasing, and the mass data storage of PB ranks becomes
It obtains more and more important.The local disk storage mode of traditional sense has been unable to meet large capacity, high reliability, the height of existing application
The requirements such as performance, extending transversely, to meet these new demands, network-based distributed memory system has obtained extensive attention.
Today's society network, mobile communication, network video and audio, e-commerce, sensor network, scientific experiment etc. are each
The data that kind of application generates, often from several KB to several MB etc., the size dimension of these data is irregular, copes with common big
File, traditional file system can obtain excellent performance, but the efficiency when handling mass small documents write-in by high concurrent
Under meeting is very low, this is because mass small documents often bring a large amount of random writing, it can be frequent because of disk when being written locally
Addressing greatly reduces performance, and the distributed storage to become more and more popular at present is since there are inborn network delay, more copies to set
Setting can be such that performance is even lower with metadata write-in.And if merely considering small documents performance, and big file process can be lost
Advantage, there are certain paradoxes for the two.
Invention content
To solve the above-mentioned problems, the present invention proposes a kind of storage method and system, can avoid a large amount of random small texts
Part write-in caused by disk frequently address, degraded performance the problem of, improve write efficiency.
In order to achieve the above object, the present invention proposes a kind of storage method, and this method includes:
Before application layer operates file, the pre- of this document is obtained by preset index node inode lists
Fixed storage class.
This document is handled accordingly according to this document corresponding storage class.
Preferably,
Storage class includes:Database purchase DBStore based on database and the file based on file
Store FileStore.
Preferably,
This method further includes:Inode lists are pre-set in the following manner:
Different files to be operated are corresponded from different inode values.
Inode values and the storage class of corresponding file are corresponded.
Correspondence between the storage class of this document, inode values and file is stored in inode lists, and
Inode lists are stored in DBStore.
Preferably,
This method further includes:The action type of different files to be operated is predefined in the following manner:
The size of this document is compared with preset threshold value.
When the size of file is less than preset threshold value, this document is determined as small documents, and the small documents are determined as
DBStore storage classes.
When the size of file is greater than or equal to preset threshold value, this document is determined as big file, and by the big file
It is determined as FileStore storage classes.
Preferably,
Carrying out corresponding processing to file according to the corresponding storage class of file includes:
Create operate in, determine the first file to be operated be DBStore storage classes after, file to be operated is straight
It connects in storage to DBStore;Wherein, the first file to be operated is metadata, and metadata is small documents.
In delete operation, the second file to be operated is determined after DBStore storage classes, to directly acquire to be operated
Location information of second file in DBStore, and the second file to be operated is deleted from DBStore;Wherein, to be operated
The second file be metadata, and metadata be small documents.
In read request operation, after it is DBStore storage classes to determine third file to be operated, according to inode values
It determines the position of third file to be operated in DBStore, and reads third file to be operated from DBStore;When true
Third file to be operated is determined after FileStore storage classes, to determine that third file to be operated exists according to inode values
Position in FileStore, and third file to be operated is read from FileStore.
In write request operation, after it is DBStore storage classes to determine the 4th file to be operated, by multiple with small text
4th file existing for part form is recorded according to the sequencing of write-in, and the 4th file and inode values are corresponded,
Form mapping pair, by the 4th file to be stored in buffering area in the form of mapping pair, and will mapping to being recorded in mapping daily record,
Successful response ACK is write in return, when mapping the mapping of log recording to reaching preset amount threshold, according in operation log
The action type of record continues to execute write operation, stores the 4th file into DBStore according to the sequencing of write-in;When true
After fixed 4th file to be operated is FileStore storage classes, directly by entire with the 4th text existing for big document form
Part is sliced, and each slice is distributed in FileStore and is stored.
In order to achieve the above object, the invention also provides a kind of storage system, which includes:Acquisition module and processing
Module.
Acquisition module, for before application layer operates file, being obtained by preset index node inode lists
Obtain the scheduled storage class of this document.
Processing module, for being handled accordingly this document according to the corresponding storage class of this document.
Preferably,
Storage class includes:Database purchase DBStore based on database and the file based on file
Store FileStore.
Preferably,
The system further includes setup module;Setup module for pre-setting inode lists in the following manner:
Different files to be operated are corresponded from different inode values.
Inode values and the storage class of corresponding file are corresponded.
Correspondence between the storage class of this document, inode values and file is stored in inode lists, and
Inode lists are stored in DBStore.
Preferably,
The system further includes determination type module;Determination type module is to be operated for predefining in the following manner
The action type of different files:
The size of file is compared with preset threshold value.
When the size of file is less than preset threshold value, file is determined as small documents, and small documents are determined as
DBStore storage classes.
When the size of file is greater than or equal to preset threshold value, file is determined as big file, and big file is determined
For FileStore storage classes.
Preferably,
Processing module carries out corresponding processing according to the corresponding storage class of file to file:
Create operate in, determine the first file to be operated be DBStore storage classes after, file to be operated is straight
It connects in storage to DBStore;Wherein, the first file to be operated is metadata, and metadata is small documents.
In delete operation, the second file to be operated is determined after DBStore storage classes, to directly acquire to be operated
Location information of second file in DBStore, and the second file to be operated is deleted from DBStore;Wherein, to be operated
The second file be metadata, and metadata be small documents.
In read request operation, after it is DBStore storage classes to determine third file to be operated, according to inode values
It determines the position of third file to be operated in DBStore, and reads third file to be operated from DBStore;When true
Third file to be operated is determined after FileStore storage classes, to determine that third file to be operated exists according to inode values
Position in FileStore, and third file to be operated is read from FileStore.
In write request operation, after it is DBStore storage classes to determine the 4th file to be operated, by multiple with small text
4th file existing for part form is recorded according to the sequencing of write-in, and the 4th file and inode values are corresponded,
Form mapping pair, by the 4th file to be stored in buffering area in the form of mapping pair, and will mapping to being recorded in mapping daily record,
Successful response ACK is write in return, when mapping the mapping of log recording to reaching preset amount threshold, according in operation log
The write operation of record continues to execute write operation, stores the 4th file into DBStore according to the sequencing of write-in;Work as determination
4th file to be operated be FileStore storage classes after, directly by entire with the 4th file existing for big document form
It is sliced, each slice is distributed in FileStore and is stored.
Compared with prior art, the present invention includes:Before application layer operates file, saved by preset index
Point inode lists obtain the scheduled storage class of this document.This document is carried out according to this document corresponding storage class
Corresponding processing.Scheme through the invention can avoid disk caused by a large amount of random small documents write-ins from frequently addressing, property
The low problem of energy, improves write efficiency.
Description of the drawings
The attached drawing in the embodiment of the present invention is illustrated below, the attached drawing in embodiment be for the present invention into one
Step understands, for explaining the present invention together with specification, does not constitute limiting the scope of the invention.
Fig. 1 is the storage method flow chart of the present invention;
Fig. 2 is the storage system composition frame chart of the present invention.
Specific implementation mode
For the ease of the understanding of those skilled in the art, the invention will be further described below in conjunction with the accompanying drawings, not
It can be used for limiting the scope of the invention.
The present invention is directed to develop a kind of novel storage method to solve the big small documents write-in that current distributed storage is faced
Problem, the application scenarios for user's complexity provide general solution.
In order to achieve the above object, the present invention proposes a kind of storage method, as shown in Figure 1, this method includes:
S101, before application layer operates file, this document is obtained by preset index node inode lists
Scheduled storage class.
Preferably,
Storage class includes:Database purchase DBStore based on database and the file based on file
Store FileStore.
For the ease of user's operation, rear end storage is divided into two types in advance in the present invention program, i.e., (such as with database
LevelDB the storage DBStore and storage FileStore based on file based on), in an embodiment of the present invention,
The former is suitable for small documents, and the latter is suitable for big file.Summarized according to actual test, the present invention in using 128KB as boundary, be less than or
Data equal to the value are stored using DBStore, and the data for being more than the value are stored using FileStore.
Preferably,
This method further includes:Inode lists are pre-set in the following manner:
S201, different files to be operated are corresponded from different inode values.
S202, inode values and the storage class of corresponding file are corresponded.
S203, the correspondence between the storage class of this document, inode values and file is stored in inode lists
In, and inode lists are stored in DBStore.
In embodiments of the present invention, through the above scheme so that operator adds application layer file in application layer
Also it can be quickly found out the action type corresponding to this document when operation, carry out fast operating, improve operating efficiency.
In addition, in order in user's operation rapid pin DBStore is carried out according to its file size to the operation file of user
Storage or FileStore storages need the classification that the corresponding operation file of various operations is carried out to storage class in advance, will be different
Operation file correspond in advance in its corresponding action type, when user carries out the operation, can directly invoke operation text
The corresponding action type of part is stored, and disk caused by a large amount of random small documents write-ins can be avoided frequently to address, performance
Low problem, improves write efficiency.
Preferably,
This method further includes:The action type of different files to be operated is predefined in the following manner:
S301, the size of this document is compared with preset threshold value.
In embodiments of the present invention, which could be provided as above-mentioned 128KB, and certainly, the threshold value is simultaneously unlimited
In this, it can also be set as other numerical value as the case may be.
S302, when the size of file be less than preset threshold value when, this document is determined as small documents, and the small documents are true
It is set to DBStore storage classes.
When the size of file is greater than or equal to preset threshold value, this document is determined as big file, and by the big file
It is determined as FileStore storage classes.
S102, this document is handled accordingly according to this document corresponding storage class.
It is being previously-completed the classification of rear end storage by above-mentioned scheme and the behaviour corresponding to various operations is determined
After making the corresponding action type of file, create in user, delete, the operations such as read request, write request when, so that it may directly to adjust
Stored, improved work efficiency with the corresponding action type of operation file, below we will respectively be directed to different operations
Type is described in detail.
Preferably,
Carrying out corresponding processing to file according to the corresponding storage class of file includes:
Situation one, create operate in, determine the first file to be operated be DBStore storage classes after, will be to be operated
File be directly stored in DBStore;Wherein, the first file to be operated is metadata, and metadata is small documents.
In embodiments of the present invention, it is almost 0 that establishment operation can consider its corresponding operation file, which is data volume, it
The metadata information of file is provided only, general metadata is regarded as small documents data, therefore, directly will in creating operation
In metadata storage to DBStore.
Situation two, in delete operation, determine the second file to be operated be DBStore storage classes after, directly acquire
The location information of the second file to be operated in DBStore, and the second file to be operated is deleted from DBStore;Its
In, the second file to be operated is metadata, and metadata is small documents.
In embodiments of the present invention, delete operation generally requires the metadata information for reading and deleting file, finds this yuan of number
According to storage location, and it is deleted, it is stated that general metadata is regarded as small documents number in above-mentioned establishment operation
According to, therefore, the operation file corresponding to delete operation is also centainly stored in DBStore, therefore, in delete operation, we
Directly to the file metadata being stored in DBStore is read in DBStore, its storage location is obtained, then by corresponding data
It deletes one by one.
Situation three, in read-write requests operation, generally first can operate or write according to the read request that is preserved in inode lists and ask
The operation file for asking operation corresponding determines the storage class corresponding to this document, is then sent again into corresponding storage and reads to ask
Operation or write request is asked to operate.Below safety pin to read request operate and write request operation illustrate.
In read request operation, after it is DBStore storage classes to determine third file to be operated, according to inode values
It determines the position of third file to be operated in DBStore, and reads third file to be operated from DBStore;When true
Third file to be operated is determined after FileStore storage classes, to determine that third file to be operated exists according to inode values
Position in FileStore, and third file to be operated is read from FileStore.
In embodiments of the present invention, it is operated for read request, the file searched in caching corresponding to read request can be arrived first, such as
Exist in fruit caching, then directly return the result, if do not had in caching, just goes to get reading according to inode values in DBStore and ask
The position of corresponding file is asked, is then gone according to said program corresponding in DBStore storages or FileStore storages
Position acquisition this document object.
In write request operation, after it is DBStore storage classes to determine the 4th file to be operated, by multiple with small text
4th file existing for part form is recorded according to the sequencing of write-in, and the 4th file and inode values are corresponded,
Form mapping pair, by the 4th file to be stored in buffering area in the form of mapping pair, and will mapping to being recorded in mapping daily record,
Successful response ACK is write in return, later when mapping the mapping of log recording to reaching preset amount threshold, according to operation day
The action type recorded in will continues to execute write operation, stores the 4th file into DBStore according to the sequencing of write-in,
Ensure the write efficiency of application layer with this;It, directly will be whole after it is FileStore storage classes to determine the 4th file to be operated
A is sliced with the 4th file existing for big document form, and each slice is distributed in FileStore and is stored, is obtained
Obtained higher aggregate bandwidth.
In embodiments of the present invention, it is operated for write request, can ensure that size file data can through the above scheme
Efficiently rule.
In order to achieve the above object, the invention also provides a kind of storage systems 01, as shown in Fig. 2, the system includes:It obtains
Modulus block 02 and processing module 03.
Acquisition module 02, for before application layer operates file, passing through preset index node inode lists
Obtain the scheduled storage class of this document.
Preferably,
Storage class includes:Database purchase DBStore based on database and the file based on file
Store FileStore.
Processing module 03, for being handled accordingly this document according to the corresponding storage class of this document.
Preferably,
The system further includes setup module 04;Setup module 04 for pre-setting inode lists in the following manner:
Different files to be operated are corresponded from different inode values.
Inode values and the storage class of corresponding file are corresponded.
Correspondence between the storage class of this document, inode values and file is stored in inode lists, and
Inode lists are stored in DBStore.
Preferably,
The system further includes determination type module 05;Determination type module 05 waits grasping for predefining in the following manner
The action type for the different files made:
The size of file is compared with preset threshold value.
When the size of file is less than preset threshold value, file is determined as small documents, and small documents are determined as
DBStore storage classes.
When the size of file is greater than or equal to preset threshold value, file is determined as big file, and big file is determined
For FileStore storage classes.
Preferably,
Processing module 03 carries out corresponding processing according to the corresponding storage class of file to file:
Create operate in, determine the first file to be operated be DBStore storage classes after, file to be operated is straight
It connects in storage to DBStore;Wherein, the first file to be operated is metadata, and metadata is small documents.
In delete operation, the second file to be operated is determined after DBStore storage classes, to directly acquire to be operated
Location information of second file in DBStore, and the second file to be operated is deleted from DBStore;Wherein, to be operated
The second file be metadata, and metadata be small documents.
In read request operation, after it is DBStore storage classes to determine third file to be operated, according to inode values
It determines the position of third file to be operated in DBStore, and reads third file to be operated from DBStore;When true
Third file to be operated is determined after FileStore storage classes, to determine that third file to be operated exists according to inode values
Position in FileStore, and third file to be operated is read from FileStore.
In write request operation, after it is DBStore storage classes to determine the 4th file to be operated, by multiple with small text
4th file existing for part form is recorded according to the sequencing of write-in, and the 4th file and inode values are corresponded,
Form mapping pair, by the 4th file to be stored in buffering area in the form of mapping pair, and will mapping to being recorded in mapping daily record,
Successful response ACK is write in return, when mapping the mapping of log recording to reaching preset amount threshold, according in operation log
The write operation of record continues to execute write operation, stores the 4th file into DBStore according to the sequencing of write-in;Work as determination
4th file to be operated be FileStore storage classes after, directly by entire with the 4th file existing for big document form
It is sliced, each slice is distributed in FileStore and is stored.
Compared with prior art, the present invention includes:Before application layer operates file, saved by preset index
Point inode lists obtain the scheduled storage class of this document.This document is carried out according to this document corresponding storage class
Corresponding processing.Scheme through the invention can avoid disk caused by a large amount of random small documents write-ins from frequently addressing, property
The low problem of energy, improves write efficiency.
The classification of storage class in the present invention program completes size text from the common cooperation of different action types
The read-write of number of packages evidence, compared to traditional storage method, storage method of the invention mainly has following characteristics:
1) efficient metadata management.It is stored since the metadata of file has been put into as small block data in DBStore,
So good performance can be obtained in random read-write, and depositing for file parts data object is managed using database
Position is put, efficiently, fast, and database has born transactional operating feature, that is, the continuity operations of affairs, it is in office
When what link malfunctions, it can all stop the operation of the things, this admirably ensure that the consistency of data.
2) aggregate bandwidth.Can (be e.g., first to read by action type in the present invention program when handling the write-in of small documents data
Request operation or write request operation) and the data of operation file be recorded in the journal file of disk, and safeguarded in memory
Operating list in make a label, wait the operation amount in journal files to reach after predetermined value and seek unity of action, realize suitable again
The effect of sequence write-in had both avoided disk caused by a large amount of random small documents write-ins and had frequently addressed, and improved write efficiency in this way,
It can ensure that system can carry out operation playback based on journal file in unexpected power-off again, it is ensured that data consistency.In bulk
When data are written, and the effect of the high concurrent feature of each operation and various respective behaviors acquisition aggregate bandwidth can be relied on.
It should be noted that embodiment described above be merely for convenience of it will be understood by those skilled in the art that, and
It is not used in and limits the scope of the invention, under the premise of not departing from the inventive concept of the present invention, those skilled in the art couple
Any obvious replacement and improvement that the present invention is made etc. are within protection scope of the present invention.
Claims (6)
1. a kind of storage method, which is characterized in that the method includes:
Before application layer operates file, the predetermined of the file is obtained by preset index node inode lists
Storage class;
The file is handled accordingly according to the file corresponding storage class;
Wherein, the storage class includes:Database purchase DBStore based on database and based on file
File stores FileStore;
Wherein, the method further includes:The inode lists are pre-set in the following manner:
Different files to be operated are corresponded from different inode values;
The inode values and the storage class of corresponding file are corresponded;
Correspondence between the storage class of the file, the inode values and the file is stored in the inode
In list, and the inode lists are stored in the DBStore;
Wherein, in write request operation, after it is DBStore storage classes to determine the 4th file to be operated, by multiple with small
4th file existing for document form is recorded according to the sequencing of write-in, and by the 4th file with it is described
Inode values correspond, and form mapping pair, by the 4th file to be stored in buffering area in the form of mapping pair, and by institute
Mapping is stated to being recorded in mapping daily record, successful response ACK is write in return, when the mapping of the mapping log recording is to reaching
When to preset amount threshold, write operation is continued to execute according to the action type recorded in operation log, according to the priority of write-in
Sequence will be in the 4th file storage to the DBStore;When determine the 4th file to be operated be FileStore store class
It after type, is directly sliced, each slice is distributed to described with the 4th file existing for big document form by entire
It is stored in FileStore.
2. storage method as described in claim 1, which is characterized in that the method further includes:In the following manner in advance really
The action type of fixed different files to be operated:
The size of the file is compared with preset threshold value;
When the size of the file is less than the preset threshold value, the file is determined as small documents, and by the small text
Part is determined as the DBStore storage classes;
When the size of the file is greater than or equal to the preset threshold value, the file is determined as big file, and by institute
It states big file and is determined as the FileStore storage classes.
3. storage method as claimed in claim 2, which is characterized in that described according to the corresponding storage class pair of the file
The file carries out corresponding processing:
Create operate in, determine first file to be operated be the DBStore storage classes after, will be described to be operated
File be directly stored in the DBStore;Wherein, first file to be operated is metadata, and first number
According to for small documents;
In delete operation, determine that second file to be operated is described after the DBStore storage classes, to directly acquire
The location information of the second file to be operated in the DBStore, and by second file to be operated from described
It is deleted in DBStore;Wherein, second file to be operated is metadata, and the metadata is small documents;
In read request operation, after it is the DBStore storage classes to determine the third file to be operated, according to described
Inode values determine position of the third file to be operated in the DBStore, and institute is read from the DBStore
State third file to be operated;After it is the FileStore storage classes to determine the third file to be operated, according to institute
It states inode values and determines position of the third file to be operated in the FileStore, and from the FileStore
Read the third file to be operated.
4. a kind of storage system, which is characterized in that the system comprises:Acquisition module and processing module;
The acquisition module, for before application layer operates file, being obtained by preset index node inode lists
Obtain the scheduled storage class of the file;
The processing module, for being handled accordingly the file according to the corresponding storage class of the file;
Wherein, the storage class includes:Database purchase DBStore based on database and based on file
File stores FileStore;
Wherein, the system also includes setup modules;The setup module for pre-setting the inode in the following manner
List:
Different files to be operated are corresponded from different inode values;
The inode values and the storage class of corresponding file are corresponded;
Correspondence between the storage class of the file, the inode values and the file is stored in the inode
In list, and the inode lists are stored in the DBStore;
Wherein, in write request operation, after it is DBStore storage classes to determine the 4th file to be operated, by multiple with small
4th file existing for document form is recorded according to the sequencing of write-in, and by the 4th file with it is described
Inode values correspond, and form mapping pair, by the 4th file to be stored in buffering area in the form of mapping pair, and by institute
Mapping is stated to being recorded in mapping daily record, successful response ACK is write in return, when the mapping of the mapping log recording is to reaching
When to preset amount threshold, write operation is continued to execute according to the action type recorded in operation log, according to the priority of write-in
Sequence will be in the 4th file storage to the DBStore;When determine the 4th file to be operated be FileStore store class
It after type, is directly sliced, each slice is distributed to described with the 4th file existing for big document form by entire
It is stored in FileStore.
5. storage system as claimed in claim 4, which is characterized in that the system also includes determination type modules;The class
Type determining module is used to predefine the action type of different files to be operated in the following manner:
The size of the file is compared with preset threshold value;
When the size of the file is less than the preset threshold value, the file is determined as small documents, and by the small text
Part is determined as the DBStore storage classes;
When the size of the file is greater than or equal to the preset threshold value, the file is determined as big file, and by institute
It states big file and is determined as the FileStore storage classes.
6. storage system as claimed in claim 5, which is characterized in that the processing module is deposited according to the file is corresponding
Storage type carries out corresponding processing to the file:
Create operate in, determine first file to be operated be the DBStore storage classes after, will be described to be operated
File be directly stored in the DBStore;Wherein, first file to be operated is metadata, and first number
According to for small documents;
In delete operation, determine that second file to be operated is described after the DBStore storage classes, to directly acquire
The location information of the second file to be operated in the DBStore, and by second file to be operated from described
It is deleted in DBStore;Wherein, second file to be operated is metadata, and the metadata is small documents;
In read request operation, after it is the DBStore storage classes to determine the third file to be operated, according to described
Inode values determine position of the third file to be operated in the DBStore, and institute is read from the DBStore
State third file to be operated;After it is the FileStore storage classes to determine the third file to be operated, according to institute
It states inode values and determines position of the third file to be operated in the FileStore, and from the FileStore
Read the third file to be operated.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510370132.8A CN105094695B (en) | 2015-06-29 | 2015-06-29 | A kind of storage method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510370132.8A CN105094695B (en) | 2015-06-29 | 2015-06-29 | A kind of storage method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105094695A CN105094695A (en) | 2015-11-25 |
CN105094695B true CN105094695B (en) | 2018-09-04 |
Family
ID=54575255
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510370132.8A Active CN105094695B (en) | 2015-06-29 | 2015-06-29 | A kind of storage method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105094695B (en) |
Families Citing this family (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105487824A (en) * | 2015-12-07 | 2016-04-13 | 联想(北京)有限公司 | Information processing method, storage device and electronic device |
CN105404691A (en) * | 2015-12-14 | 2016-03-16 | 曙光信息产业股份有限公司 | File storage method and apparatus |
CN105787012B (en) * | 2016-02-23 | 2019-08-27 | 深圳市瑞驰信息技术有限公司 | A kind of method and storage system improving storage system processing small documents |
CN107315533A (en) * | 2016-04-26 | 2017-11-03 | 杭州海康威视数字技术股份有限公司 | A kind of date storage method and device |
CN107580066A (en) * | 2017-09-20 | 2018-01-12 | 郑州云海信息技术有限公司 | The method, apparatus and system of file access in a kind of distributed NAS storage system |
CN111488122A (en) | 2017-10-25 | 2020-08-04 | 华为技术有限公司 | Data writing method and storage device |
CN108897698B (en) * | 2018-06-29 | 2021-10-29 | 郑州云海信息技术有限公司 | File data block addressing method, system and equipment and storage medium |
CN109241011B (en) * | 2018-09-21 | 2023-01-06 | 联想(北京)有限公司 | Virtual machine file processing method and device |
CN112822084B (en) * | 2019-11-18 | 2022-08-30 | 瑞昱半导体股份有限公司 | Gateway control chip and network packet processing method thereof |
CN111324581A (en) * | 2020-02-14 | 2020-06-23 | 苏州浪潮智能科技有限公司 | File writing method, device, equipment and storage medium |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103246700A (en) * | 2013-04-01 | 2013-08-14 | 厦门市美亚柏科信息股份有限公司 | Mass small file low latency storage method based on HBase |
CN104111804A (en) * | 2014-06-27 | 2014-10-22 | 暨南大学 | Distributed file system |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8768980B2 (en) * | 2009-11-02 | 2014-07-01 | Stg Interactive S.A. | Process for optimizing file storage systems |
-
2015
- 2015-06-29 CN CN201510370132.8A patent/CN105094695B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103246700A (en) * | 2013-04-01 | 2013-08-14 | 厦门市美亚柏科信息股份有限公司 | Mass small file low latency storage method based on HBase |
CN104111804A (en) * | 2014-06-27 | 2014-10-22 | 暨南大学 | Distributed file system |
Non-Patent Citations (1)
Title |
---|
基于NoSQL的PDM图文档分布式存储与并行查询;徐四鹏;《中国优秀硕士学位论文全文数据库 信息科技辑》;20150615;正文第13-18、20-26页 * |
Also Published As
Publication number | Publication date |
---|---|
CN105094695A (en) | 2015-11-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105094695B (en) | A kind of storage method and system | |
US8782343B2 (en) | System and method for optimizing data remanence over hybrid disk clusters using various storage technologies | |
US9317213B1 (en) | Efficient storage of variably-sized data objects in a data store | |
EP2879040B1 (en) | Data storage method, data storage apparatus, and storage device | |
US10725666B2 (en) | Memory-based on-demand data page generation | |
CN110647497A (en) | HDFS-based high-performance file storage and management system | |
CN103399823B (en) | The storage means of business datum, equipment and system | |
CN107832423B (en) | File reading and writing method for distributed file system | |
WO2011121746A1 (en) | File server apparatus, method of controlling storage system, and program | |
CN104111804A (en) | Distributed file system | |
CN103037004A (en) | Implement method and device of cloud storage system operation | |
US7624091B2 (en) | Data prefetch in storage device | |
CN108089825B (en) | Storage system based on distributed cluster | |
CN113626431A (en) | LSM tree-based key value separation storage method and system for delaying garbage recovery | |
WO2021184996A1 (en) | Data storage method and apparatus for database | |
CN103034592A (en) | Data processing method and device | |
CN111143290A (en) | Memory management method, system, equipment and computer readable storage medium | |
CN107423425B (en) | Method for quickly storing and inquiring data in K/V format | |
CN103595720A (en) | Offloaded data transferring method, device and client | |
US20060143313A1 (en) | Method for accessing a storage device | |
US11474981B2 (en) | Database dual-core storage system based on optical disk and method using the system | |
CN114374687B (en) | File transmission method and device between thermomagnetic storage and blue light storage | |
CN103491124A (en) | Method for processing multimedia message data and distributed cache system | |
CN111796767B (en) | Distributed file system and data management method | |
CN101996237A (en) | Write command processing method, equipment and system of storage area network equipment |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |