CN103793291B - Distributed data copying method and device - Google Patents
Distributed data copying method and device Download PDFInfo
- Publication number
- CN103793291B CN103793291B CN201210430530.0A CN201210430530A CN103793291B CN 103793291 B CN103793291 B CN 103793291B CN 201210430530 A CN201210430530 A CN 201210430530A CN 103793291 B CN103793291 B CN 103793291B
- Authority
- CN
- China
- Prior art keywords
- source
- volume
- write
- request
- sub
- 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
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
An embodiment of the invention provides a distributed data copying method and device. The method includes that a write request is received, the write request is separated into multiple write sub-requests, each write sub-request carries part of the data carried by the write request, the write address of each write sub-request points to one source sub-volume of a source volume of a production site, the source volume is divided into multiple source sub-volumes, and each source sub-volume is located at one storage node of the production site; data is written into each source sub-volume according to the write sub-requests, and the backup information of each source sub-volume is generated; each backup information is transmitted to a disaster recovery backup site, and the backup information of the source sub-volumes jointly forms the backup information of the source volume.
Description
Technical field
The present invention relates to the communication technology, more particularly to a kind of distributed data clone method and device.
Background technology
With the development of information-intensive society, data volume is increasing, and data calamity is standby also more and more important.Now, each storage section
Point manufacturer typically realizes local and remote disaster recovery solution by remote copy technology.Disaster tolerance system generally comprises local life
Website and long-range calamity are produced for website, the data of local production website are backuped to for website by long-range calamity by remote copy technology,
Separated using website or geography separate to ensure defensive ability/resistance ability of the disaster tolerance system to disaster event, it is ensured that data are not lost or tried one's best
It is few to lose.Remote copy pattern is divided into synchronous and asynchronous two kinds.In synchronous mode, production website need wait production website and
Calamity just returns response of host after writing complete for website, therefore can ensure that zero-data loss, but exists with calamity for the logical of website
Letter time delay, affects main frame input and output（Input Output；Hereinafter referred to as：IO）Performance, therefore general synchronous mode is applied to
Distance is less than the same city disaster tolerance scene within 200km.In asynchronous mode, production website first will after receiving and writing I/O data
Its write local datastore system, and return to application system main frame and write complete confirmation message, then to transmit data to calamity standby
Website, therefore the impact suffered by the performance of locally applied system is less, transmission range is substantially unrestricted, but due to data transfer
Exist delayed, therefore calamity there may be certain loss of data for website.
In the prior art, asynchronous remote copy technology typically comprises snapshot mechanism and log mechanism, by snapshot data
Or the main frame of log buffer writes data duplication to calamity for website, Fig. 1 is that the scene of asynchronous remote copy technology in prior art is shown
It is intended to, is illustrated in figure 1 using the remote copy process of snapshot mechanism.In high-end many controlled storage systems, generally all using collection
Chinese style share framework, the multiple memory nodes in many controlled storage systems are set by the way of globally shared to a LUN
It is standby（Logic Unit Number device；Hereinafter referred to as：LUN）Shared.
However, data copy method of the prior art is needed to existing between multiple memory nodes in many controlled storage systems
Exclusive reference, increases so as to cause main frame to write IO time delays.
The content of the invention
The embodiment of the present invention provides a kind of distributed data clone method and device, it is to avoid between multiple memory nodes mutually
Scold the phenomenon of access, be reduced as far as the time delay that main frame writes IO.
The first aspect of the embodiment of the present invention is to provide a kind of distributed data clone method, including：
Receive write request, by the write request be split as it is multiple write sub- request, it is each described to write the data that sub- request carries and be
A part for the data that the write request is carried, the write address that sub- request is write described in each is pointed in the source book on production website
One source son volume, wherein, the source book is divided into multiple source volumes, and each described source volume is located at one in production website
On memory node；
Sub- request is write according to the plurality of, data is write in each source volume respectively, and generates each source respectively
The backup information of son volume；
Each backup information is sent to into calamity for website, the backup information of each source volume collectively forms source book
Backup information.
In the first possible implementation of first aspect, the calamity is divided into multiple mesh for target volume on website
Son volume is marked, each target is rolled up on a memory node in being located at the calamity for website, and methods described is further included：
Calamity receives the backup information of each source volume, and the backup information storage that each source for receiving is rolled up for website
To in target volume described at least one.
With reference to the first possible implementation of first aspect or first aspect, second in first aspect is possible
It is described to write sub- request according to the plurality of in implementation, respectively data are write in each source volume, and generate respectively each
The backup information of the source volume, including：
According to sub- request is respectively write, number is write in each source son volume pointed by the write address for respectively writing sub- request respectively
According to；
According to the write sequence of data, respectively write described sub- request and be separately recorded in the daily record of each source volume, with
The daily record of each source volume is generated respectively；Wherein, the daily record of the source volume is the backup information of the source volume.
With reference to second possible implementation of first aspect, in the third possible implementation of first aspect
In, the backup information by each source volume is sent to calamity for website, including：
The daily record of each source volume is sent to into calamity for website, so that calamity writes son for website according to what is recorded in the daily record
Request writes data in each target son volume in the target volume.
With reference to first aspect or second possible implementation of first aspect, in the 4th kind of possible implementation
In, it is described to write sub- request according to the plurality of, write data in each source volume respectively, and each source is generated respectively
The backup information of volume, including：
According to sub- request is respectively write, number is write in each source son volume pointed by the write address for respectively writing sub- request respectively
According to；
The data in the son volume of each described source after to writing data carry out snapshot process, and each source volume is generated respectively
Snapshot；Wherein, the snapshot of the source volume is the backup information of the source volume.
It is described by respectively in the 5th kind of possible implementation with reference to the 4th kind of possible implementation of first aspect
The backup information of source son volume is sent to calamity for website, including：
The snapshot of each source volume is sent to into calamity for website.
The second aspect of the embodiment of the present invention is to provide a kind of distributed data reproducing unit, including：
Receiver module, for the write request that Receiving Host sends；
Split module, for by the write request be split as it is multiple write sub- request, it is each described to write the data that sub- request is carried
It is a part for the data that the write request is carried, each write address for writing sub- request points in the source book on production website
Individual source volume, wherein, the source book is divided into the sub- volume in multiple sources, and each described source volume is located at produced in website and deposits
On storage node；
Source son volume processing module, for writing sub- request according to the plurality of, writes data in each source volume respectively,
And generate the backup information of each source volume respectively；
Sending module, for the backup information of each source volume to be sent to into calamity for website, each source volume
Backup information collectively forms the backup information of source book.
In the first possible implementation of second aspect, the source volume processing module includes：
First r/w cell, for according to sub- request is respectively write, respectively each pointed by the write address for respectively writing sub- request
Data are write in the son volume of source；
First signal generating unit, for according to the write sequence of data, by it is described respectively write sub- request be separately recorded in it is each described
In the daily record of source son volume, to generate the daily record of each source volume respectively；Wherein, the daily record of the source volume is source son volume
Backup information.
With reference to the first possible implementation of second aspect, in second possible implementation of second aspect
In, the sending module specifically for the daily record of each source volume is sent to into calamity for website so that calamity for website according to described
Writing in each target son volume of the sub- request in the target volume for recording in daily record writes data.
With reference to second aspect, in the third possible implementation of second aspect, the source volume processing module bag
Include：
Second r/w cell, for according to sub- request is respectively write, respectively each pointed by the write address for respectively writing sub- request
Data are write in the son volume of source；
Second signal generating unit, for carrying out snapshot process to the data in each described source son volume after write data, respectively
Generate the snapshot of each source volume；Wherein, the snapshot of the source volume is the backup information of the source volume.
With reference to the third possible implementation of second aspect, in the 4th kind of possible implementation, the transmission
Module by the snapshot of each source volume specifically for being sent to calamity for website.
The third aspect of the embodiment of the present invention is to provide a kind of another kind of distributed data reproducing unit, including processor and
Memory, the memory is used for store instruction；The processor is coupled with the memory, and the processor is configured to perform
Storage instruction in which memory, wherein, the processor is configurable for：The write request that Receiving Host sends, will
The write request be split as it is multiple write sub- request, write described in each sub- request write address point to production website on source book in
One source son volume, wherein, the source book is divided into multiple source volumes, and each described source volume is located at one in production website
On memory node；Sub- request is write according to the plurality of, data are write in each source volume respectively, and is generated respectively each described
The backup information of source son volume；The backup information of each source volume is sent to into calamity for website, the backup of each source volume
Information collectively forms the backup information of source book.
The embodiment of the present invention has the technical effect that：According to the write address carried in the write request of main frame, the write request is torn open
In assigning to the corresponding at least one source volume of write address, number is write in the son volume of at least one source according to the write request after fractionation
According to, rather than perform write operation in whole source book；Therefore when multiple write requests access what multiple memory nodes were shared simultaneously
During source book, different write requests perform write operation in corresponding not homologous sub- volume, are not in a write request to whole source
When volume performs write operation, it is impossible to perform the corresponding write operation of other write requests again；And when production website is carried out far to calamity for website
It is also to be replicated in units of the son volume of source when journey is replicated, the reproduction process of each source volume is independent of each other, and can copy to parallel
Calamity for website in each target son volume, that is, avoid the phenomenon of exclusive reference between multiple memory nodes, be reduced as far as
Main frame writes the time delay of IO.
Description of the drawings
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
The accompanying drawing to be used needed for having technology description is briefly described, it should be apparent that, drawings in the following description are these
Some bright embodiments, for those of ordinary skill in the art, without having to pay creative labor, can be with
Other accompanying drawings are obtained according to these accompanying drawings.
Fig. 1 is the schematic diagram of a scenario of asynchronous remote copy technology in prior art；
Fig. 2 is the flow chart of distributed data clone method embodiment one of the present invention；
Fig. 3 is the flow chart of distributed data clone method embodiment two of the present invention；
Fig. 4 is that LUN divides schematic diagram in distributed data clone method embodiment two of the present invention；
Fig. 5 is the flow chart of distributed data clone method embodiment three of the present invention；
Fig. 6 is that LUN divides schematic diagram in distributed data clone method embodiment three of the present invention；
Fig. 7 is the structural representation of distributed data reproducing unit embodiment one of the present invention；
Fig. 8 is the structural representation of distributed data reproducing unit embodiment two of the present invention；
Fig. 9 is the structural representation of distributed data reproducing unit embodiment three of the present invention.
Specific embodiment
To make purpose, technical scheme and the advantage of the embodiment of the present invention clearer, below in conjunction with the embodiment of the present invention
In accompanying drawing, the technical scheme in the embodiment of the present invention is clearly and completely described, it is clear that described embodiment is
The a part of embodiment of the present invention, rather than the embodiment of whole.Based on the embodiment in the present invention, those of ordinary skill in the art
The every other embodiment obtained under the premise of creative work is not made, belongs to the scope of protection of the invention.
Fig. 2 is the flow chart of distributed data clone method embodiment one of the present invention, as shown in Fig. 2 the present embodiment is provided
A kind of distributed data clone method, can specifically include following steps：
Step 201, Receiving Host send write request, by the write request be split as it is multiple write sub- request, respectively write sub- request
The data of carrying are a parts for the data that the write request is carried, and each write address for writing sub- request points to the source on production website
Source son volume in volume.Wherein, the source book is divided into multiple source volumes, and each source son volume is located at first in production website
On memory node.
In the present embodiment, website and calamity are produced for each self-contained multiple memory nodes in website, production website and source book
Corresponding, calamity is corresponding with target volume for website, can will produce the corresponding source book of website and be divided into multiple source volumes, each source
Son volume is located on a memory node in production website.After production website receives the write request of main frame transmission, acquisition is write
The write address carried in request, is split to write request in the corresponding at least one source volume of write address according to the write address.This
The volume at place can be specially LUN.Wherein, the write address in write request is integrally corresponding to one or more sources son volume, writing herein
Address can be specifically identified the position that data will be write in source book, and due to producing the corresponding source book of website multiple source are divided into
Volume, the write address in write request can be pointing directly to source book, can be a source son volume, or multiple source volumes, that is, write
The data of request are wanted to write at the position of one or more source son volumes.Therefore, the present embodiment can be according to the write address to writing
Request is split, and write request is split in corresponding one or more the source son volumes of write address, i.e. the several sources of write address correspondence
Son volume, is just split as write request several, and respectively each source volume corresponding with write address is corresponding for each write request after fractionation.
Step 202, according to multiple sub- request is write, and writes data in the son volume of each source respectively, and generates each source volume respectively
Backup information.
In the present embodiment, after write request is split to according to write address in the son volume of at least one source, after fractionation
Write request respectively at least one source son volume in write data, a whole write request is split to one or more by the present embodiment
In the son volume of source, respectively the write operation of the write request after splitting is performed in the son volume of each source, rather than performed in a whole source book
Write operation.Meanwhile, not only source book divided in the present embodiment, also the backup of source book is also carried out into corresponding division, herein
The backup of source book can be specially the daily record of source book or the snapshot of source book.While writing data in a source son volume, will write
Write request after the data for entering or fractionation generates the backup of source son volume, for example, the write request after fractionation is recorded corresponding
In the daily record of source son volume, or by the snapshot of the corresponding source volume of data Ji Lu in the son volume of write source.
Step 203, calamity is sent to for website by the backup information of each source son volume, the common structure of backup information of each source volume
Into the backup information of source book.
In the present embodiment, after the data write operation for completing source book, need the data duplication in production website to calamity
In standby website, so as to realize data backup, this step be specially by the backup copy of source book to calamity for the target volume in website in,
So that calamity is backed up in the data duplication in the source book to the target volume for website according to described.In the present embodiment, calamity
The standby corresponding target volume of website is also carried out corresponding division, the target volume after division and calamity for website in each node it is relative
Should.When data backup is carried out, can by source book the backup of active sub- volume copy to calamity for the corresponding target volume of website
In, then calamity can be realized in the data duplication in source book to target volume for website according to backup.
A kind of distributed data clone method is present embodiments provided, according to the write address carried in the write request of main frame,
The write request is split in the corresponding at least one source volume of write address, according to the write request after fractionation at least one source
Data are write in volume, rather than write operation is performed in whole source book；Therefore when simultaneously multiple write requests access multiple storage sections
Point share source book when, different write requests performs write operation in corresponding not homologous sub- volume, be not in one write ask
When asking to whole source book execution write operation, it is impossible to perform the corresponding write operation of other write requests again；And when production website is standby to calamity
It is also to be replicated in units of the son volume of source when website carries out remote copy, the reproduction process of each source son volume is independent of each other, can be with
Each target son volume during calamity is copied to parallel for website, that is, avoid the phenomenon of exclusive reference between multiple memory nodes, to the greatest extent may be used
Reduce the time delay that main frame writes IO energy.
Fig. 3 is the flow chart of distributed data clone method embodiment two of the present invention, as shown in figure 3, the present embodiment is provided
A kind of distributed data clone method, can specifically include following steps：
Step 301, Receiving Host send write request, by the write request be split as it is multiple write sub- request, respectively write sub- request
The data of carrying are a parts for the data that the write request is carried, and each write address for writing sub- request points to the source on production website
Source son volume in volume, wherein, the source book is divided into multiple source volumes, and each source son volume is located at one in production website
On memory node.
In the present embodiment, it is many controlled storage systems to produce website, and multiple memory nodes are included in production website, and is produced
Website is corresponding with source book.Storage computing resource distribution situation of the present embodiment in production website, by production website correspondence
Source book be divided into multiple sources son volume.Storage computing resource distribution situation herein can specially produce each storage section in website
The distribution situation of the storage computing resource such as hard disk, CPU, internal memory of point, the source book in the present embodiment can be specially source LUN, will
Source LUN is divided into the sub- LUN in multiple sources.It is illustrated in figure 4 LUN in distributed data clone method embodiment two of the present invention and divides and shows
Be intended to, production website include N number of memory node, i.e. Node0, Node1 ..., NodeN, N number of memory node is common in prior art
A source LUN is enjoyed, and a source LUN is divided into the sub- LUN in multiple sources by the present embodiment, each source after division is sub- LUN and one
Memory node is corresponding, will source LUN be divided into the sub- LUN in N number of source, lun0, lun1 in such as Fig. 4 ..., lunN, wherein lun0
Corresponding with Node0, lun1 is corresponding with Node1 ..., and lunN is corresponding with NodeN.
In addition, in the present embodiment, after production website receives the write request of main frame, what is carried in acquisition write request writes
Address, is split to write request in the corresponding at least one source volume of write address according to the write address.Wherein, writing in write request
Integrally corresponding to one or more source son volumes, write address herein can be specifically identified the position that data will be write in source book for address
Put, the present embodiment can split according to the write address to write request, by write request be split to write address it is corresponding one or
In the son volume of multiple sources, i.e. several sources volumes of write address correspondence are just split as write request several, and after fractionation sub- request point is respectively write
Each source volume not corresponding with write address is corresponding.
Step 302, according to sub- request is respectively write, writes respectively in the son volume of each source pointed by the write address of sub- request is respectively write
Data.
After write request is split to according to write address in the son volume of at least one source, according to fractionation after write sub- request difference
Data are write in the son volume of at least one source, the present embodiment is split to a whole write request in one or more source son volumes, point
The write operation for writing sub- request after splitting is not performed in the son volume of each source, rather than write operation is performed in a whole source book.
Step 303, according to the write sequence of data, will respectively write sub- request and be separately recorded in the daily record of each source volume, with
The daily record of each source volume is generated respectively.
Wherein, the daily record of the source son volume is the backup information of the source son volume.
As shown in figure 4, in the present embodiment, not only source LUN is divided into into the sub- LUN in N number of source, also by source LUN corresponding days
Will is also carried out corresponding division, will the corresponding log of source LUN be also divided into N number of sub- daily record, i.e., log0, log1 in figure ...,
LogN, log0 and lun0 are corresponding, and log1 is corresponding with lun1 ..., and logN is corresponding with lunN.The present embodiment writing after according to fractionation
Son request write data at least one source son volume in after, also by fractionation after write sub- request be separately recorded at least one source son
In the daily record of volume, so as to generate the daily record of each source volume.
Step 304, the daily record of each source son volume is sent to into calamity for point, so that calamity writes son for website according to what is recorded in the daily record
Request writes data in each target son volume in the target volume；The backup information of each source volume collectively forms the backup letter of source book
Breath.
In the present embodiment, calamity is used to carry out disaster-tolerant backup to producing website for website, and calamity is also to control one for website more
Storage system, for multiple memory nodes are included in website, and calamity is corresponding with target volume for website for calamity.The present embodiment is standby according to calamity
Storage computing resource distribution situation in website, multiple target volumes are divided into by calamity for the corresponding target volume of website.Herein
Storage computing resource distribution situation can be specially calamity and calculate money for the storage such as the hard disk of each memory node, CPU, internal memory in website
The distribution situation in source, the target volume in the present embodiment can be specially target LUN, will target LUN be divided into multiple targets
LUN.Above-mentioned Fig. 4 can also represent dividing condition of the calamity for corresponding target LUN of website.
In addition, after the data write operation for completing source book, need by the data duplication in production website to calamity for website in,
So as to realize data backup, this step is specially the daily record reproduction of source book to calamity in the target volume in website.Entering line number
According to during backup, can by the LUN of source the daily record of active sub- LUN copy to calamity in corresponding target LUN of website.
Step 305, calamity receives the backup information of each source volume, and the backup information that each source for receiving is rolled up for website
Store at least one target son volume.
Wherein, calamity is divided into multiple target volumes for target volume on website, and each target son volume is located at the calamity for website
In a memory node on.
It is total in the present embodiment, be to write sub- request after fractionation due to what is recorded in the daily record of source book, then calamity is being received for website
To production website send daily record after, therefrom obtain split after write sub- request, directly according to fractionation after write sub- request, in mesh
Data are write in mark volume, so as to the disaster-tolerant backup of data in realizing to producing the corresponding source book of website.In the present embodiment, calamity is standby
The corresponding target volume of website has been also carried out the similar division of source book corresponding with production website, then when write operation is performed, be also
Write address according to writing in sub- request is assigned in each target volume to write data.
A kind of distributed data clone method is present embodiments provided, according to the write address carried in the write request of main frame,
The write request is split in the corresponding at least one source volume of write address, according to the write request after fractionation at least one source
Data are write in volume, rather than write operation is performed in whole source book；Therefore when simultaneously multiple write requests access multiple storage sections
Point share source book when, different write requests performs write operation in corresponding not homologous sub- volume, be not in one write ask
When asking to whole source book execution write operation, it is impossible to perform the corresponding write operation of other write requests again；And when production website is standby to calamity
It is also to be replicated in units of the son volume of source when website carries out remote copy, the reproduction process of each source son volume is independent of each other, can be with
Each target son volume during calamity is copied to parallel for website, that is, avoid the phenomenon of exclusive reference between multiple memory nodes, to the greatest extent may be used
Reduce the time delay that main frame writes IO energy.
Fig. 5 is the flow chart of distributed data clone method embodiment three of the present invention, as shown in figure 5, the present embodiment is provided
A kind of distributed data clone method, can specifically include following steps：
Step 501, Receiving Host send write request, by the write request be split as it is multiple write sub- request, respectively write sub- request
The data of carrying are a parts for the data that the write request is carried, and each write address for writing sub- request points to the source on production website
Source son volume in volume, wherein, the source book is divided into multiple source volumes, and each source son volume is located at one in production website
On memory node.
Step 502, according to sub- request is respectively write, writes number in the son volume of each source pointed by the address of sub- request is respectively write respectively
According to.
Data in each source son volume after write data are carried out snapshot process by step 503, and each source volume is generated respectively
Snapshot.
Wherein, the snapshot of the source son volume is the backup information of the source son volume.
Step 504, by each source son volume snapshot be sent to calamity for website so that calamity for website according in the snapshot record write
Son request writes data in each target son volume in the target volume；The backup information of each source volume collectively forms the backup of source book
Information.
Step 505, calamity receives the backup information of each source volume, and the backup information that each source for receiving is rolled up for website
Store in target volume described at least one.
Above-mentioned steps 501-505 can be similar with step 301-305 in above-described embodiment two respectively, herein no longer one by one
Repeat.
The present embodiment is with the difference of above-described embodiment two, and the backup of source book or target volume is snapshot, rather than daily record.
Fig. 6 is that LUN divides schematic diagram in distributed data clone method embodiment three of the present invention, as shown in fig. 6, in the present embodiment,
Not only source LUN is divided into into the sub- LUN in N number of source, also the snapshot of source LUN is also carried out into corresponding division, will source LUN it is corresponding
Snap is also divided into N number of sub- daily record, i.e., snap0, snap1 in figure ..., snapN, snap0 is corresponding with lun0, snap1 with
Lun1 correspondences ..., snapN is corresponding with lunN.Write sub- request of the present embodiment after according to fractionation writes data at least one
Source son volume in after, also to write operation after at least one source son volume in data carry out snapshot process, by after snapshot data protect
In there is corresponding snapshot, so as to generate the snapshot of at least one source son volume.
It is data of the source book in certain historical juncture due to what is recorded in the snapshot of source book, then calamity is receiving life for website
After producing the snapshot that website sends, in each target son volume in the data duplication to target volume that directly will be recorded in each snapshot, so as to
The disaster-tolerant backup of data in realizing to producing the corresponding source book of website.In the present embodiment, calamity is for the corresponding target volume of website
The similar division of source book corresponding with production website is carried out, then when source book and target volume adopt identical partition strategy, often
Individual source son volume one target son volume of correspondence, the snapshot that now directly can be rolled up source to its one-to-one mesh
In mark volume；And when source book and target volume are provided without identical partition strategy, it is each after each source son volume after division and division
There is the corresponding relation on address between target son volume, in data duplication, can be answered data according to the corresponding relation of the two
In making corresponding target volume.
A kind of distributed data clone method is present embodiments provided, according to the write address carried in the write request of main frame,
The write request is split in the corresponding at least one source volume of write address, according to the write request after fractionation at least one source
Data are write in volume, rather than write operation is performed in whole source book；Therefore when simultaneously multiple write requests access multiple storage sections
Point share source book when, different write requests performs write operation in corresponding not homologous sub- volume, be not in one write ask
When asking to whole source book execution write operation, it is impossible to perform the corresponding write operation of other write requests again；And when production website is standby to calamity
It is also to be replicated in units of the son volume of source when website carries out remote copy, the reproduction process of each source son volume is independent of each other, can be with
Each target son volume during calamity is copied to parallel for website, that is, avoid the phenomenon of exclusive reference between multiple memory nodes, to the greatest extent may be used
Reduce the time delay that main frame writes IO energy.
One of ordinary skill in the art will appreciate that：Realizing all or part of step of above-mentioned each method embodiment can lead to
Cross the related hardware of programmed instruction to complete.Aforesaid program can be stored in a computer read/write memory medium.The journey
Sequence upon execution, performs the step of including above-mentioned each method embodiment；And aforesaid storage medium includes：ROM, RAM, magnetic disc or
Person's CD etc. is various can be with the medium of store program codes.
Fig. 7 is the structural representation of distributed data reproducing unit embodiment one of the present invention, as shown in fig. 7, the present embodiment
There is provided a kind of distributed data reproducing unit, each step in said method embodiment one can be specifically performed, herein not
Repeat again.The distributed data reproducing unit that the present embodiment is provided can specifically include receiver module 701, split module 702, source
Son volume processing module 703 and sending module 704.Wherein, receiver module 701 is used for the write request that Receiving Host sends.Split mould
Block 702 be used for by the write request be split as it is multiple write sub- request, respectively write the number that the data that sub- request carries are that the write request is carried
According to a part, each write sub- request write address point to production website on source book in a source son volume, wherein, the source book
Multiple source volumes are divided into, each source son volume is located on a memory node in production website.Source son volume processing module
703 are used to write sub- request according to the plurality of, write data in the respectively source son volume respectively, and generate the standby of the sub- volume in the respectively source respectively
Part information.Sending module 704 is used to for the backup information of the respectively source son volume to be sent to calamity for website, the backup letter that respectively source is rolled up
Breath collectively forms the backup information of source book.
The distributed data reproducing unit of the present embodiment can perform the technical scheme of embodiment of the method shown in Fig. 1, in fact
Existing principle is similar, and here is omitted.
A kind of distributed data reproducing unit is present embodiments provided, according to the write address carried in the write request of main frame,
The write request is split in the corresponding at least one source volume of write address, according to the write request after fractionation at least one source
Data are write in volume, rather than write operation is performed in whole source book；Therefore when simultaneously multiple write requests access multiple storage sections
Point share source book when, different write requests performs write operation in corresponding not homologous sub- volume, be not in one write ask
When asking to whole source book execution write operation, it is impossible to perform the corresponding write operation of other write requests again；And when production website is standby to calamity
It is also to be replicated in units of the son volume of source when website carries out remote copy, the reproduction process of each source son volume is independent of each other, can be with
Each target son volume during calamity is copied to parallel for website, that is, avoid the phenomenon of exclusive reference between multiple memory nodes, to the greatest extent may be used
Reduce the time delay that main frame writes IO energy.
Fig. 8 is the structural representation of distributed data reproducing unit embodiment two of the present invention, above-mentioned as shown in fig. 7, originally
Embodiment provides a kind of distributed data reproducing unit, can specifically perform each step in said method embodiment two,
Here is omitted.The source son volume processing module 703 in the distributed data reproducing unit that the present embodiment is provided includes：First
The signal generating unit 7032 of r/w cell 7031 and first.Wherein, the first r/w cell 7031 is used for according to sub- request is respectively write, each at this respectively
Write in the son volume of each source pointed by the write address of sub- request and write data；First signal generating unit 7032 is used for according to the write of data
Sequentially, this is respectively write sub- request to be separately recorded in the respectively daily record of the source son volume, to generate the daily record of the respectively source son volume respectively.Its
In, the daily record of the source son volume is the backup information of the source son volume.
Preferably, sending module 704 specifically for by each source son volume daily record be sent to calamity for website so that calamity is for website root
Data are write according to writing in each target son volume of the sub- request in the target volume for record in the daily record.
Fig. 9 is the structural representation of distributed data reproducing unit embodiment three of the present invention, above-mentioned as shown in fig. 7, originally
Embodiment provides a kind of distributed data reproducing unit, can specifically perform each step in said method embodiment three,
Here is omitted.The source son volume processing module 703 in the distributed data reproducing unit that the present embodiment is provided includes：Second
The signal generating unit 7034 of r/w cell 7033 and second.Wherein, the second r/w cell 7033 is used for according to sub- request is respectively write, each at this respectively
Write in the son volume of each source pointed by the write address of sub- request and write data；After second signal generating unit 7034 is used for writing data
Respectively the data in the source son volume carry out snapshot process, and the snapshot of the respectively source son volume is generated respectively.Wherein, the snapshot of the source son volume is
The backup information of the source son volume.
Preferably, sending module 704 by the snapshot of the respectively source son volume specifically for being sent to calamity for website.
The present embodiment also provides a kind of distributed data reproducing unit, can specifically include processor and memory, and this is deposited
Reservoir is used for store instruction；The processor is coupled with the memory, and the processor is configured to execution and is stored in the memory
Instruction, wherein, the processor is configurable for：The write request that Receiving Host sends, by the write request multiple writing is split as
Son request, each write address for writing sub- request points to the source son volume in the source book on production website, wherein, the source book quilt
Multiple source volumes are divided into, each source son volume is located on a memory node in production website；Please according to the plurality of son of writing
Ask, write data in the respectively source son volume respectively, and generate the backup information of the respectively source son volume respectively；By the standby of the respectively source son volume
Part information is sent to calamity for website, and the backup that respectively source is rolled up collectively forms the backup information of source book.
Finally it should be noted that：Various embodiments above only to illustrate technical scheme, rather than a limitation；To the greatest extent
Pipe has been described in detail with reference to foregoing embodiments to the present invention, it will be understood by those within the art that：Its according to
So the technical scheme described in foregoing embodiments can be modified, either which part or all technical characteristic are entered
Row equivalent；And these modifications or replacement, do not make the essence disengaging various embodiments of the present invention technology of appropriate technical solution
The scope of scheme.
Claims (5)
1. a kind of distributed data clone method, it is characterised in that include：
Receive write request, by the write request be split as it is multiple write sub- request, it is each that described to write the data that sub- request carries be described
A part for the data that write request is carried, writes one that the write address of sub- request is pointed in the source book on production website described in each
Source son volume, wherein, the source book is divided into multiple source volumes, and each described source volume is located at a storage in production website
On node；
Sub- request is write according to the plurality of, data are write in each source volume respectively, and generates each source volume respectively
Backup information；
Each backup information is sent to into calamity for website, the backup information of each source volume collectively forms the backup of source book
Information；
The calamity is divided into multiple target volumes for target volume on website, and each target son volume is located at the calamity in website
On one memory node, methods described is further included：
Calamity receives the backup information of each source volume for website, and by the backup information storage of each source for receiving son volume to extremely
In a few target volume；
It is described to write sub- request according to the plurality of, respectively data are write in each source volume, and each source is generated respectively
The backup information of son volume, including：
According to sub- request is respectively write, data are write in each source son volume pointed by the write address for respectively writing sub- request respectively；
According to the write sequence of data, respectively write described sub- request and be separately recorded in the daily record of each source volume, with respectively
Generate the daily record of each source volume；Wherein, the daily record of the source volume is the backup information of the source volume.
2. method according to claim 1, it is characterised in that it is standby that the backup information by each source volume is sent to calamity
Website, including：
The daily record of each source volume is sent to into calamity for website, so that calamity writes sub- request for website according to what is recorded in the daily record
Data are write in each target son volume in the target volume.
3. a kind of distributed data reproducing unit, it is characterised in that include：
Receiver module, for the write request that Receiving Host sends；
Split module, for by the write request be split as it is multiple write sub- request, it is each that described to write the data that sub- request carries be institute
A part for the data of write request carrying is stated, each write address for writing sub- request points to a source in the source book on production website
Son volume, wherein, the source book is divided into multiple source volumes, and each described source volume is located at a storage section in production website
Point on；
Source son volume processing module, for writing sub- request according to the plurality of, writes data in each source volume respectively, and divides
Do not generate the backup information of each source volume；
Sending module, for the backup information of each source volume to be sent to into calamity for website, the backup of each source volume
Information collectively forms the backup information of source book；
The source volume processing module includes：
First r/w cell, for according to sub- request is respectively write, respectively in each source pointed by the write address for respectively writing sub- request
Data are write in volume；
First signal generating unit, for according to the write sequence of data, respectively writes described sub- request and is separately recorded in each source
In the daily record of volume, to generate the daily record of each source volume respectively；Wherein, the daily record of the source volume is the standby of the source volume
Part information.
4. device according to claim 3, it is characterised in that the sending module is specifically for by each source volume
Daily record is sent to calamity for website, so that calamity writes each target of the sub- request in target volume for website according to what is recorded in the daily record
Data are write in volume.
5. a kind of distributed data reproducing unit, it is characterised in that including processor and memory, the memory is used to store
Instruction；The processor is coupled with the memory, and the processor is configured to perform storage instruction in which memory,
Wherein, the processor is configurable for：The method that perform claim requires any claim in 1-2.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210430530.0A CN103793291B (en) | 2012-11-01 | 2012-11-01 | Distributed data copying method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210430530.0A CN103793291B (en) | 2012-11-01 | 2012-11-01 | Distributed data copying method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103793291A CN103793291A (en) | 2014-05-14 |
| CN103793291B true CN103793291B (en) | 2017-04-19 |
Family
ID=50668999
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210430530.0A Active CN103793291B (en) | 2012-11-01 | 2012-11-01 | Distributed data copying method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103793291B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104484136B (en) * | 2014-12-25 | 2017-09-29 | 深圳联友科技有限公司 | A kind of method of sustainable high concurrent internal storage data |
| CN104571956A (en) * | 2014-12-29 | 2015-04-29 | 成都致云科技有限公司 | Data writing method and splitting device |
| CN104571957B (en) * | 2014-12-29 | 2018-03-27 | 成都极驰科技有限公司 | A kind of method for reading data and assembling device |
| CN106407040B (en) | 2016-09-05 | 2019-05-24 | 华为技术有限公司 | A kind of duplicating remote data method and system |
| CN109426588B (en) * | 2017-08-28 | 2023-03-14 | 中兴通讯股份有限公司 | Method and system for realizing data protection and data distribution processing module |
| CN108268218B (en) * | 2018-01-10 | 2021-02-02 | 北京思特奇信息技术股份有限公司 | Writing method and reading method of variable length data segment based on network coding |
| CN108762984B (en) * | 2018-05-23 | 2021-05-25 | 杭州宏杉科技股份有限公司 | Method and device for continuous data backup |
| CN110737716B (en) * | 2018-07-02 | 2022-09-23 | 华为技术有限公司 | Data writing method and device |
| CN110544136A (en) * | 2019-09-10 | 2019-12-06 | 恩亿科(北京)数据科技有限公司 | Method, device, equipment and storage medium for calculating synchronous putting probability |
| CN113553007A (en) * | 2021-07-20 | 2021-10-26 | 中国工商银行股份有限公司 | Host disk data snapshot copy method, device, equipment and storage medium |
| CN113721849B (en) * | 2021-08-23 | 2024-04-12 | 深圳市杉岩数据技术有限公司 | Data copying and unloading method based on distributed storage and terminal equipment |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101540777A (en) * | 2008-03-18 | 2009-09-23 | 环达电脑(上海)有限公司 | System and method for backing up distributed data for network memory device |
| CN101630303A (en) * | 2009-08-24 | 2010-01-20 | 成都市华为赛门铁克科技有限公司 | Request message processing method, device and multi-processor memory system |
| CN102158540A (en) * | 2011-02-18 | 2011-08-17 | 广州从兴电子开发有限公司 | System and method for realizing distributed database |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5423037A (en) * | 1992-03-17 | 1995-06-06 | Teleserve Transaction Technology As | Continuously available database server having multiple groups of nodes, each group maintaining a database copy with fragments stored on multiple nodes |
| CN101741911B (en) * | 2009-12-18 | 2014-04-30 | 中兴通讯股份有限公司 | Multi-copy collaboration-based write operation method, system and node |
| CN101819543A (en) * | 2010-02-01 | 2010-09-01 | 成都市华为赛门铁克科技有限公司 | Snapshot data processing method, system and storage device |
| EP2643771B1 (en) * | 2010-11-26 | 2017-11-08 | Telefonaktiebolaget LM Ericsson (publ) | Real time database system |
| CN102624866B (en) * | 2012-01-13 | 2014-08-20 | 北京大学深圳研究生院 | Data storage method, data storage device and distributed network storage system |
-
2012
- 2012-11-01 CN CN201210430530.0A patent/CN103793291B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101540777A (en) * | 2008-03-18 | 2009-09-23 | 环达电脑(上海)有限公司 | System and method for backing up distributed data for network memory device |
| CN101630303A (en) * | 2009-08-24 | 2010-01-20 | 成都市华为赛门铁克科技有限公司 | Request message processing method, device and multi-processor memory system |
| CN102158540A (en) * | 2011-02-18 | 2011-08-17 | 广州从兴电子开发有限公司 | System and method for realizing distributed database |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103793291A (en) | 2014-05-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103793291B (en) | Distributed data copying method and device | |
| CN103777897B (en) | Method and system for copying data between primary and secondary storage locations | |
| CN102467408B (en) | Method and device for accessing data of virtual machine | |
| US10664177B2 (en) | Replicating tracks from a first storage site to a second and third storage sites | |
| CN102929748B (en) | Data back up method and device | |
| US20050071393A1 (en) | Data storage subsystem | |
| CN101755257B (en) | Managing the copying of writes from primary storages to secondary storages across different networks | |
| CN100470490C (en) | Method, system, and article of manufacture for consistent copying of storage volumes | |
| CN101808137B (en) | Data transmission method, device and system | |
| US20070174673A1 (en) | Storage system and data restoration method thereof | |
| CN102096557B (en) | Capacity expansion method, device and system for independent redundant array of inexpensive disc (RAID) | |
| CN102129434B (en) | Method and system for reading and writing separation database | |
| CN103049220B (en) | Storage controlling method, memory control device and solid-state memory system | |
| CN103814360A (en) | Data replication method and storage system | |
| CN103780638A (en) | Data synchronization method and system | |
| CN103229171A (en) | Snapshot based replication | |
| CN105740248B (en) | A kind of method of data synchronization, apparatus and system | |
| CN106933747B (en) | Data-storage system and date storage method based on multithread | |
| US20100217857A1 (en) | Consolidating session information for a cluster of sessions in a coupled session environment | |
| CN104520802B (en) | Data sending method, data receiving method and storage device | |
| CN105988901A (en) | Data copying method and storage system | |
| CN106528338A (en) | Remote data replication method, storage equipment and storage system | |
| TW201502777A (en) | Data flush of group table | |
| CN106873902B (en) | File storage system, data scheduling method and data node | |
| CN101937378B (en) | Method for carrying out back-up protection on data of storage equipment and computer system |
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 |