CN106503051B - A kind of greediness based on meta data category prefetches type data recovery system and restoration methods - Google Patents
A kind of greediness based on meta data category prefetches type data recovery system and restoration methods Download PDFInfo
- Publication number
- CN106503051B CN106503051B CN201610847382.0A CN201610847382A CN106503051B CN 106503051 B CN106503051 B CN 106503051B CN 201610847382 A CN201610847382 A CN 201610847382A CN 106503051 B CN106503051 B CN 106503051B
- Authority
- CN
- China
- Prior art keywords
- block
- data
- metadata
- file
- backup
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/27—Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a kind of greediness based on meta data category to prefetch type data recovery system and restoration methods, the data recovery system includes the reading meta data block being successively linked in sequence, block data search module and recovery data module, wherein, meta data block is read for concurrently reading backup metadata information；Block data search module is used to provide the looking up function of the block of the LRU cache in based on memory, and is responsible for management and is related to restoring the dsc data of backup；Restore data needed for data module is used to rebuild in local disk.The present invention is by backup metadata organization at file metadata and block metadata, it is segmented after file metadata is prefetched to memory completely, then Greedy is carried out to block metadata to prefetch, realize that high efficiency data are restored under the premise of not sacrificing duplicate removal rate and making full use of hardware resource, it does not need to be written over data block operation when to Backup Data duplicate removal simultaneously, thereby may be ensured that the duplicate removal rate of data.
Description
Technical field
The present invention relates to data in duplicate removal standby system to restore relevant technical field, and in particular to one kind is based on metadata
The greediness of classification prefetches type data recovery system and restoration methods.
Background technique
With the arriving of big data era, the swift and violent growth of data volume brings huge to the limited storage space of data center
Big challenge.Bandwidth needed for disk expense and network transmission needed for data de-duplication technology greatly reduces storing data, gradually
Gradually become a critical data reduction technology of current standby system.The purpose of duplicate removal standby system storing data is in order to after
Can timely restore, if enterprise disk break down or database-relative file damage etc. disasters occur when its cannot and
When restore data, the loss born will be inestimable, and data backup also becomes to have little significance.Due to the continuous data of logic
By physical dispersion in different Disk Localities, the fragment that traditional De-weight method generates has seriously affected the restorability of system.
Existing some optimization methods are all an attempt to improve the restorability in later period using algorithm is rewritten in backup before, however weight
Write algorithm itself there are the drawbacks of be exactly that must obtain better data recovery performance to sacrifice the cost of duplicate removal rate, finally lead
Cause the result of waste disk space.And traditional method can only generate a backup metadata after the completion of backup so as to the later period
Restore, leading to system, continually poor efficiency accesses the metadata on disk in recovery process, affects the restorability of system.
Summary of the invention
The first purpose of this invention is provided a kind of based on metadata point to solve drawbacks described above in the prior art
The greediness of class prefetches type data recovery system, which realizes under the premise of not sacrificing duplicate removal rate and making full use of hardware resource
High efficiency data are restored, while not needing to be written over data block operation when to Backup Data duplicate removal, thereby may be ensured that
The duplicate removal rate of data.
Another object of the present invention is provided a kind of based on metadata point to solve drawbacks described above in the prior art
The greediness of class prefetches type data reconstruction method.
The first purpose of this invention can be reached by adopting the following technical scheme that:
A kind of greediness based on meta data category prefetches type data recovery system, and the data recovery system includes successively suitable
Reading meta data block, block data search module and the recovery data module of sequence connection,
Wherein, the reading meta data block is for concurrently reading backup metadata information, wherein the backup metadata
Information is stored in two files, the logical message of file, referred to as file metadata in the backup of file record, another
File then has recorded the relevant information of block, referred to as block metadata；
Described piece of data search module is used to provide the looking up function of the block of the LRU cache in based on memory, and is responsible for pipe
Reason is related to restoring the dsc data of backup；
Data needed for the recovery data module is used to rebuild in local disk；
Restore the request of a certain Backup Data when the data recovery system receives, first by the reading meta data block from
The backup is opened in disk and backs up metadata information file accordingly, by the content of file metadata, can learn that recovery should
Block metadata needed for Backup Data, after obtaining the fingerprint of block and the location information where it, by described piece of data search module
The looking up function for the block that LRU cache in based on memory provides, the read request of a block number evidence is induced according to each piece of fingerprint,
By the container sequence number where block, the container comprising the block is searched in the LRU cache that the data recovery system maintains, if
Cache hit then fills block number evidence in the queue maintained in memory, subsequently into the recovery data module, if caching is not ordered
In, the data recovery system will find corresponding block number evidence from container cell, and prefetch the container comprising the block to caching,
The recovery data module is entered back into later.
Further, the file metadata includes: the serial number of backup, backs up number after included file number, backup
According to the name, the length of name of each file in the path of storage and the length of pathname and backup, for the block for being included
Number, the size of file and file ID number, described piece of metadata include: each piece of fingerprint, the container sequence number where block, block it is big
It is small.
Further, it is multiple and different to establishing under the same catalogue simultaneously that multiple threads are run in the recovery data module
File.
Further, the file ID number is by the offset in backup serial number, block metadata and the block number after file declustering
It is formed by displacement operation, the offset in described piece of metadata refers to the reading block metadata operation of each induction in block metadata
Start the offset of read operation in file.
Further, the reading meta data block is complete file metadata first after reading backup metadata information and starting
Portion is read in advance in the LRU cache of memory, then carries out greedy reading to block metadata according to the information in it.
Further, the file metadata in the LRU cache of memory is read in advance and is divided into multiple segments, is wrapped in each segment
Backup metadata information containing several files, and each segment can induce the operation of a reading block metadata.
Further, parallel reading block metadata operation is induced by the file ID number.
Further, described piece of data search module opens up a global LRU cache in memory, and the data are restored
The mechanism of the shared caching of internal system, meanwhile, when shared buffer memory is written to LRU cache in peripheral operation unit, LRU cache
It is locked, other modules cannot modify to LRU cache operation, it is necessary to just can be carried out relevant operation after waiting caching unlock.
Further, the reading meta data block, described piece of data search module and the recovery data module are parallel
Work.
Another object of the present invention can be reached by adopting the following technical scheme that:
A kind of greediness based on meta data category prefetches type data reconstruction method, and the data reconstruction method includes:
S1, restore the request of a certain Backup Data when data recovery system receives, read meta data block and opened from disk
The backup metadata information file all in advance reads file metadata in backup metadata information file the LRU cache of memory
In, by the content of file metadata in meta data file, learns block metadata needed for restoring the backup, obtain the finger of block
The size of container sequence number, block where line, block；
S2, greedy reading is carried out to block metadata, reads the file metadata in memory in advance and be divided into multiple segments, each
The backup metadata information of multiple files is contained in segment, and each segment can induce the operation of a reading block metadata；
S3, by block data search module based on memory in LRU cache provide block looking up function, according to each
The fingerprint of block induces the read request of a block number evidence, and by the container sequence number where block, packet is searched in the caching that system maintains
Container containing the block fills the block data queue maintained in memory if cache hit, subsequently into recovery data module；
If cache miss, corresponding block number evidence is found from container cell, and prefetches the container comprising the block to caching, Zhi Houzai
Into recovery data module；
S4, restore specified the restoration path of data module, all write data units are built under specified restoration path
It founds respective catalogue and file carries out rebuilding required data.
The present invention has the following advantages and effects with respect to the prior art:
(1), the present invention prefetches file metadata by backup metadata organization at file metadata and block metadata completely
It is segmented after to memory, Greedy then is carried out to block metadata and is prefetched, the restorability of data reconstruction method is improved.
(2), the present invention avoids the phenomenon that sacrificing duplicate removal rate generation, to save for classical rewriting algorithm
About disk storage overhead.
(3), the present invention reduces the extensive of system using parallel mechanism for traditional data reconstruction method
The multiple time.
Detailed description of the invention
Fig. 1 is the structural representation that greediness disclosed in the present invention based on meta data category prefetches type data recovery system
Figure；
Fig. 2 is the organization chart that backup metadata information is classified in the present invention.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art
Every other embodiment obtained without creative efforts, shall fall within the protection scope of the present invention.
Embodiment one
The present embodiment discloses a kind of greediness based on meta data category and prefetches type data recovery system.
The purpose of data backup is exactly to restore.Due to the continuous data of logic by physical dispersion in different disk positions
It sets, the fragment that traditional De-weight method generates has seriously affected the restorability of system.Some optimization methods existing for before are all
Be an attempt to improve the restorability in later period using algorithm is rewritten in backup, however rewrite algorithm itself there are the drawbacks of be exactly
Better data recovery performance must be obtained to sacrifice the cost of duplicate removal rate, eventually lead to the result of waste disk space.And
And traditional method can only generate a backup metadata after the completion of backup so as to later period recovery, cause system in recovery process
In the continually metadata on inefficient access disk.
It prefetches type data recovery system as shown in Figure 1, the present embodiment discloses a kind of greediness based on meta data category and includes:
It reads meta data block, block data search module, restore data module.Reading meta data block is that data are restored to carry out the preparation of early period
Backup metadata information is such as concurrently read in work.Block data search module based on memory in LRU cache provide seeking for block
Function is looked for, is responsible for administrative section and is related to restoring the dsc data of backup.Restore data module to be responsible for needed for local disk is rebuild
Data.
These three modules are concurrently to complete respective work between each other in recovery process.Each module is respectively independent complete
At needing to complete task, and there is also information exchanges between module.
Restore the request of a certain Backup Data when system receives, reading meta data block is initially entered, immediately from disk
The corresponding meta data file of the backup is opened, by the content of meta data file, can learn and restore which the backup specifically needs
A little stored block messages after the information such as the fingerprint for obtaining block and the position where it, immediately enter block data search module, often
A fingerprint induces the read request of a block number evidence, and by the fingerprint of block, the appearance comprising the block is searched in the caching that system maintains
Device fills block number evidence in the queue maintained in memory if cache hit, subsequently into recovery data module.If caching is not
Hit, system will find corresponding block number evidence from container cell, and prefetch the container comprising the block to caching, enter back into later
Restore data module.
Wherein, restore to run multiple threads in data module simultaneously to establishing multiple and different files under the same catalogue,
Per thread is all to be independently completed respective task, such recovery operation there is no need to etc. each data restore after again again
Splice data, avoids increasing additional calculating and time overhead.
System is stored backup metadata information hereof, and backup metadata information is divided into two, and is distributed in
In two files, the logical message of file, i.e. file metadata, another file are then had recorded in the backup of file record
The relevant information of block, referred to as block metadata.File metadata houses the serial number of backup, the file number which is included,
The length in data are stored after backup path and pathname.Due to containing multiple files in once backing up, each file by
Divide it is blocking, so file metadata also successively house backup in each file name, the length of name, the block for being included
Number, the size of file and file ID number.Container sequence number, block where block metadata record each piece of fingerprint, block
Size.
The relationship of file metadata and block metadata be it is closely coupled, the role of guide is played in recovery process,
Once record sequence entanglement, Backup Data will be unable to restore in the sequence entanglement in it, especially block metadata.
The size of file metadata is less than the size of block metadata, because the number of file is far smaller than each in backup
The summation of file block number.
As shown in Fig. 2, classification metadata institutional framework proposed by the present invention, complete file metadata when restoring to start
Portion is read in memory in advance, then carries out greedy reading to block metadata according to the information in it.Prefetching file metadata completely can
To reduce system to the access times of disk.In order to increase the greedy dynamics for reading data, the file member number in memory is read in advance
According to multiple segments are divided into, the metadata information of multiple files is contained in a segment, and each segment can induce one
Read the operation of block metadata.In order to make the reading block metadata operation of induction parallel, this index value of file ID number has been introduced.Pass through
This index value is analyzed, the reading block metadata operation of each induction, which is known that since the corresponding position of block meta data file, to be read
The relevant metadata information of block.File ID number be by backup serial number, block metadata offset and file in include block number
It is formed by displacement operation, the offset in block metadata refers to the reading block metadata operation of each induction in block meta data file
The middle offset for starting read operation.It can be seen that, file metadata shows File in figure1、Filem、FilenThree ID numbers, pass through
Each ID number is analyzed, can learn and restore file File1It needs to read k1A block metadata information, FilemIt needs to read k2A, Filen
It needs to read k3It is a, it is assumed that three is S respectively1、Sm、SnThree different segment start file ID numbers, due to S1、Sm、SnRespectively lure
The reading block metadata operation of hair corresponds respectively to different filec descriptor and file structure, reads between metadata operation mutually not
Interference, in addition the information deviated in ID number, so that it may which the different location in block meta data file finds required block metadata
Information.Prefetching to the greediness of block metadata is realized premised on file ID number, if each reading block is first without this index value
The information that data manipulation is read will be rambling.
The greediness of block metadata prefetches, and when restoring to start, file metadata is all read in advance the LRU cache of memory
In, greedy reading is then carried out to block metadata according to the information in it, in order to increase the greedy dynamics for reading data, is read in advance
File metadata in memory is divided into multiple segments, the metadata information of multiple files is contained in a segment, and every
A segment can induce the operation of a reading block metadata.
The data recovery system realizes that caching is shared, block data search module pass through open up in memory one it is global
LRU cache, the mechanism of the shared caching of all operation modules in system, so can be fully using between data
Locality.
When shared buffer memory is written to LRU cache in peripheral operation unit, LRU cache is locked, and other modules cannot be right
LRU cache is modified operation, it is necessary to just can be carried out relevant operation after waiting caching unlock.
The data recovery system parallel recovery, in recovery process, read the metadata of block, block data search, restore data it
Between be parallel, for same file recovery, their operation order cannot overturn.This system is regarded parallel one
One assembly line.When recovery process starts, file metadata is prefetched in memory by system completely, file metadata is divided into more
After a segment, each segment gives a pipeline processes, means that an assembly line can complete the recovery of multiple files.In this way
It between operation is parallel inside assembly line as soon as system, which possesses multiple assembly lines, to be corresponded to corresponding segment and restored, and
And the operation between assembly line and assembly line is also parallel.
Embodiment two
Present embodiment discloses a kind of greediness based on meta data category to prefetch type data reconstruction method, the data recovery side
Method specifically includes the following steps:
S1, restore the request of a certain Backup Data when data recovery system receives, read meta data block and opened from disk
The backup metadata information file all in advance reads file metadata in backup metadata information file the LRU cache of memory
In, by the content of file metadata in meta data file, it can learn block metadata needed for restoring the backup, obtain block
The size of container sequence number, block where fingerprint, block.
S2, greedy reading is carried out to block metadata, reads the file metadata in memory in advance and be divided into multiple segments, each
The backup metadata information of multiple files is contained in segment, and each segment can induce the operation of a reading block metadata；
S3, by block data search module based on memory in LRU cache provide block looking up function, according to each
The fingerprint of block induces the read request of a block number evidence, and by the container sequence number where block, packet is searched in the caching that system maintains
Container containing the block fills the block data queue maintained in memory if cache hit, subsequently into recovery data module；
If cache miss, corresponding block number evidence is found from container cell, and prefetches the container comprising the block to caching, Zhi Houzai
Into recovery data module；
S4, restore specified the restoration path of data module, all write data units are built under specified restoration path
It founds respective catalogue and file carries out rebuilding required data.
A global LRU cache is opened in block data search module, the shared caching of all operating units
Mechanism, system can be fully using the localities between data, and the container if active cell is caching addition can be another
The data of unit offer relevant block.
In conclusion the present embodiment, which proposes a kind of greediness based on meta data category, prefetches type data reconstruction method, one
Backup meta data category is organized into file metadata and block metadata, the storage of reasonable arrangement data by aspect；On the other hand
Under the premise of not sacrificial system duplicate removal rate is with hardware resource is made full use of, pass through pre- having fetched to the more radical formula of metadata
Effect improves restorability and handling capacity, can effectively optimize the recovery of required file based on the data reconstruction method that greediness prefetches
Time meets the requirement that modernization stores data.
The above embodiment is a preferred embodiment of the present invention, but embodiments of the present invention are not by above-described embodiment
Limitation, other any changes, modifications, substitutions, combinations, simplifications made without departing from the spirit and principles of the present invention,
It should be equivalent substitute mode, be included within the scope of the present invention.
Claims (10)
1. a kind of greediness based on meta data category prefetches type data recovery system, which is characterized in that the data recovery system
Reading meta data block, block data search module and recovery data module including sequential connection,
Wherein, the reading meta data block is for concurrently reading backup metadata information, wherein the backup metadata information
It is stored in two files, the logical message of file, referred to as file metadata, another file in the backup of file record
Then have recorded the relevant information of block, referred to as block metadata；
Described piece of data search module is used to provide the looking up function of the block of the LRU cache in based on memory, and is responsible for management and relates to
And restore the dsc data of backup；
Data needed for the recovery data module is used to rebuild in local disk；
Restore the request of a certain Backup Data when the data recovery system receives, first by the reading meta data block from disk
Metadata information file is backed up in middle opening backup accordingly, by the content of file metadata, can learn and restore the backup
Block metadata needed for data is based on after obtaining the fingerprint of block and the location information where it by described piece of data search module
The looking up function for the block that LRU cache in memory provides, the read request of a block number evidence is induced according to each piece of fingerprint, is passed through
Container sequence number where block searches the container comprising the block in the LRU cache that the data recovery system maintains, if caching
Hit, then fill block number evidence in the queue maintained in memory, subsequently into the recovery data module, if cache miss, institute
Corresponding block number evidence will be found from container cell by stating data recovery system, and prefetch the container comprising the block to caching, later
Enter back into the recovery data module.
2. a kind of greediness based on meta data category according to claim 1 prefetches type data recovery system, feature exists
Data are stored after, the file metadata includes: the included file number of the serial number of backup, backup, backup path and
In the length and backup of pathname the name, the length of name, the number for the block for being included, the size of file of each file and
File ID number, described piece of metadata include: the size of each piece of fingerprint, the container sequence number where block and block.
3. a kind of greediness based on meta data category according to claim 1 prefetches type data recovery system, feature exists
In running multiple threads in the recovery data module simultaneously to establishing multiple and different files under the same catalogue.
4. a kind of greediness based on meta data category according to claim 2 prefetches type data recovery system, feature exists
In the file ID number passes through displacement operation shape by the offset in backup serial number, block metadata and the block number after file declustering
At the offset in described piece of metadata refers to that the reading block metadata operation of each induction starts to read behaviour in block meta data file
The offset of work.
5. a kind of greediness based on meta data category according to claim 1 prefetches type data recovery system, feature exists
In file metadata is all read memory first after reading backup metadata information and starting by the reading meta data block in advance
In LRU cache, greedy reading is then carried out to block metadata according to the information in it.
6. a kind of greediness based on meta data category according to claim 1 prefetches type data recovery system, feature exists
In reading the file metadata in the LRU cache of memory in advance and be divided into multiple segments, each segment includes the standby of several files
Part metadata information, and each segment can induce the operation of a reading block metadata.
7. a kind of greediness based on meta data category according to claim 4 prefetches type data recovery system, feature exists
In inducing parallel reading block metadata operation by the file ID number.
8. a kind of greediness based on meta data category according to claim 1 prefetches type data recovery system, feature exists
In,
Described piece of data search module opens up a global LRU cache, the data recovery system intra-sharing in memory
The mechanism of one caching, meanwhile, when shared buffer memory is written to LRU cache in peripheral operation unit, LRU cache is locked, other
Module cannot modify to LRU cache operation, it is necessary to just can be carried out relevant operation after waiting caching unlock.
9. a kind of greediness based on meta data category according to claim 1 prefetches type data recovery system, feature exists
In the reading meta data block, described piece of data search module and the recovery data module are concurrent working.
10. a kind of greediness based on meta data category prefetches type data reconstruction method, which is characterized in that the data reconstruction method
Include:
S1, restore the request of a certain Backup Data when data recovery system receives, reading meta data block, that this is opened from disk is standby
Part metadata information file is all in advance read file metadata in backup metadata information file in the LRU cache of memory, is led to
The content for crossing file metadata in meta data file learns block metadata needed for restoring the backup, obtains fingerprint, the block institute of block
Container sequence number, block size；
S2, greedy reading is carried out to block metadata, reads the file metadata in memory in advance and is divided into multiple segments, each segment
In contain the backup metadata information of multiple files, and each segment can induce the operation of a reading block metadata；
S3, by block data search module based on memory in LRU cache provide block looking up function, according to each piece
Fingerprint induces the read request of a block number evidence, by the container sequence number where block, searches in the caching that system maintains comprising being somebody's turn to do
The container of block fills the block data queue maintained in memory if cache hit, subsequently into recovery data module；If
Cache miss finds corresponding block number evidence from container cell, and prefetches the container comprising the block to caching, enters back into later
Restore data module；
S4, restore specified the restoration path of data module, all write data units are established respectively under specified restoration path
From catalogue and file carry out rebuilding required data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610847382.0A CN106503051B (en) | 2016-09-23 | 2016-09-23 | A kind of greediness based on meta data category prefetches type data recovery system and restoration methods |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610847382.0A CN106503051B (en) | 2016-09-23 | 2016-09-23 | A kind of greediness based on meta data category prefetches type data recovery system and restoration methods |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106503051A CN106503051A (en) | 2017-03-15 |
| CN106503051B true CN106503051B (en) | 2019-05-14 |
Family
ID=58291007
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610847382.0A Active CN106503051B (en) | 2016-09-23 | 2016-09-23 | A kind of greediness based on meta data category prefetches type data recovery system and restoration methods |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106503051B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108733361B (en) * | 2017-04-20 | 2022-03-04 | 北京京东尚科信息技术有限公司 | Method and device for realizing concurrent containers |
| CN109508254B (en) | 2017-09-14 | 2020-09-08 | 华为技术有限公司 | Data recovery method and device |
| EP3671402B1 (en) * | 2017-09-14 | 2022-05-04 | Huawei Technologies Co., Ltd. | Data recovery method and apparatus |
| CN107885463B (en) * | 2017-11-10 | 2021-08-31 | 下一代互联网重大应用技术（北京）工程研究中心有限公司 | Target file processing method and device |
| CN110018990B (en) * | 2017-11-28 | 2023-07-18 | 阿里巴巴集团控股有限公司 | Method and device for caching snapshot and method and device for reading snapshot |
| CN108984338B (en) * | 2018-06-01 | 2021-11-12 | 暨南大学 | Offline optimal cache replacement device and method for data recovery of duplicate removal backup system |
| CN109710180B (en) * | 2018-12-07 | 2022-04-01 | 国云科技股份有限公司 | Distributed storage block equipment recovery method |
| CN109992708B (en) * | 2019-04-12 | 2020-09-04 | 苏州浪潮智能科技有限公司 | Method, device, equipment and storage medium for metadata query |
| CN113626381B (en) * | 2021-06-30 | 2023-12-22 | 济南浪潮数据技术有限公司 | Optimization method and device based on interleaving read-ahead of distributed file system |
| CN113625961B (en) * | 2021-07-07 | 2024-05-17 | 暨南大学 | Self-adaptive threshold value repeated data deleting method based on greedy selection |
| CN114328030B (en) * | 2022-03-03 | 2022-05-20 | 成都云祺科技有限公司 | File data backup method, system and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102541751A (en) * | 2010-11-18 | 2012-07-04 | 微软公司 | Scalable chunk store for data deduplication |
| CN102736961A (en) * | 2011-03-11 | 2012-10-17 | 微软公司 | Backup and restore strategies for data deduplication |
| CN103345449A (en) * | 2013-06-19 | 2013-10-09 | 暨南大学 | Method and system for prefetching fingerprints oriented to data de-duplication technology |
| CN103473150A (en) * | 2013-08-28 | 2013-12-25 | 华中科技大学 | Fragment rewriting method for data repetition removing system |
| CN104050103A (en) * | 2014-06-06 | 2014-09-17 | 华中科技大学 | Cache replacement method and system for data recovery |
-
2016
- 2016-09-23 CN CN201610847382.0A patent/CN106503051B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102541751A (en) * | 2010-11-18 | 2012-07-04 | 微软公司 | Scalable chunk store for data deduplication |
| CN102736961A (en) * | 2011-03-11 | 2012-10-17 | 微软公司 | Backup and restore strategies for data deduplication |
| CN103345449A (en) * | 2013-06-19 | 2013-10-09 | 暨南大学 | Method and system for prefetching fingerprints oriented to data de-duplication technology |
| CN103473150A (en) * | 2013-08-28 | 2013-12-25 | 华中科技大学 | Fragment rewriting method for data repetition removing system |
| CN104050103A (en) * | 2014-06-06 | 2014-09-17 | 华中科技大学 | Cache replacement method and system for data recovery |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106503051A (en) | 2017-03-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106503051B (en) | A kind of greediness based on meta data category prefetches type data recovery system and restoration methods | |
| CN104765575B (en) | information storage processing method | |
| CN104350477B (en) | For solid-state drive device（SSD）Optimization context remove | |
| US9513817B2 (en) | Free space collection in log structured storage systems | |
| US10078598B1 (en) | Maintaining a separate LRU linked list for each thread for multi-threaded access | |
| Giles et al. | Softwrap: A lightweight framework for transactional support of storage class memory | |
| US9836419B2 (en) | Efficient data movement within file system volumes | |
| US10956324B1 (en) | System and method for persisting hardware transactional memory transactions to persistent memory | |
| US7356651B2 (en) | Data-aware cache state machine | |
| CN109416681A (en) | The data de-duplication of workload optimization is carried out using ghost fingerprint | |
| US11756618B1 (en) | System and method for atomic persistence in storage class memory | |
| US10810123B1 (en) | Flush strategy for using DRAM as cache media system and method | |
| Subedi et al. | Stacker: an autonomic data movement engine for extreme-scale data staging-based in-situ workflows | |
| CN106662981A (en) | Storage device, program, and information processing method | |
| CN103493024B (en) | For the method and system of cache hierarchy | |
| CN110162528A (en) | Magnanimity big data search method and system | |
| CN103856567A (en) | Small file storage method based on Hadoop distributed file system | |
| US20100332446A1 (en) | Storage pool scrubbing with concurrent snapshots | |
| US20090037495A1 (en) | Method and system for state maintenance of a large object | |
| WO2013174305A1 (en) | Ssd-based key-value type local storage method and system | |
| US10061775B1 (en) | Scalable and persistent L2 adaptive replacement cache | |
| US20140215127A1 (en) | Apparatus, system, and method for adaptive intent logging | |
| CN104765574A (en) | Data cloud storage method | |
| CN106469123A (en) | A kind of write buffer distribution based on NVDIMM, method for releasing and its device | |
| CN109313605A (en) | The storage priority-based and access of compressed memory lines in memory in processor-based system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20190730 Address after: 511400 Room 3102, 82 Hanxing East Road, Zhongcun Street, Panyu District, Guangzhou, Guangdong Province Patentee after: Guangzhou Mo wing Mdt InfoTech Ltd Address before: 510632 Whampoa West Road, Guangdong, Guangzhou, No. 601, No. Patentee before: Jinan University |