CN103699340B - A kind of request processing method and equipment - Google Patents
A kind of request processing method and equipment Download PDFInfo
- Publication number
- CN103699340B CN103699340B CN201310687352.4A CN201310687352A CN103699340B CN 103699340 B CN103699340 B CN 103699340B CN 201310687352 A CN201310687352 A CN 201310687352A CN 103699340 B CN103699340 B CN 103699340B
- Authority
- CN
- China
- Prior art keywords
- request
- redundancy
- data
- read
- 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
Abstract
The present invention provides a kind of request processing method and equipment.One of which request processing method, including: after the first read request that Receiving Host sends, generate the second read request of RAID system;It it is the reading page of the second read request distribution;Second read request is split into the request of many height;If slow IO situation occurs in the request of any one height, and occurs that the son request of slow IO situation is asked for redundancy, then the request of redundancy is carried out RAID system degradation and read, return the data that the request of redundancy is corresponding.Therefore, when slow IO situation occurs in the request of any one height, and occur when the son request of slow IO situation is asked for redundancy, in the case of returning without waiting for the request of this redundancy, i.e. the request of redundancy can carry out RAID system degradation to read, return the data that the request of redundancy is corresponding, it is achieved that the redundancy request to there is slow IO situation quickly returns data, solve that the problem that the son request of slow IO cannot quickly return data occurs once in a while.
Description
Technical field
The present invention relates to RAID system technical field, particularly to a kind of request processing method and set
Standby.
Background technology
At RAID(Redundant Array of Independent Disks, RAID)
In system, it receives the read request of a main frame, and this read request can split into the request of many height
Executed in parallel on the multiple hard disks being issued in RAID system, the process time of host read request depends on
In the maximum process time of this little request, when these son requests split have a request to process
May result in whole host read request time-out when time is slower, and then affect whole RAID system
Performance.When the hard disk number of RAID system is more, there is slow IO(Input/Output in certain hard disk,
Input/output) probability will be linearly increasing, wherein slow IO is that IO response is relatively slow, and wherein IO rings
The response that antithetical phrase is asked should be referred to.This slow IO has become more typical in RAID system problem and has asked
Topic.
A kind of method of the slow IO of existing solution is to increase the identification to slow dish, when finding RAID system
In certain hard disk within a period of time, persistently occur responding slower situation by IO then to define this dish be slow
Dish, so can get rid of this slow by this slow dish off-line in the case of RAID system has redundancy
The dish impact on whole RAID system.Wherein RAID system have redundancy to refer to hard disk can pass through it
The RAID system degradation of his hard disk is read to return data, is from RAID system by this slow dish by slow dish off-line
System is removed.
Although the mode of the slow dish of this off-line can improve the process time of host read request, but its
Being applicable within a period of time the hard disk occurring slow IO continuously, the hard disk for accidental slow IO cannot
Process.And further, slow dish off-line directly may cause whole RAID system cannot
Read to return data by RAID system degradation, increase the risk that double plate lost efficacy.It addition, RAID system
System, after by slow dish off-line, needs to be reconstructed the data of other hard disks, can affect whole equally
The IO performance of RAID system.
Summary of the invention
The technical problem to be solved is to provide a kind of request processing method and equipment, in order to solve
Certainly the hard disk processing of accidental slow IO, the double plate that caused by slow dish off-line cannot be lost efficacy by prior art
And the problem of the IO performance affecting whole RAID system that slow dish off-line caused.Technical scheme
As follows:
First aspect, the embodiment of the present invention provides a kind of request processing method, including:
After the first read request that Receiving Host sends, generate the second reading of individual disc redundancy array system
Request, described second read request is for reading from the hard disk of described individual disc redundancy array system
Data;
For the reading page of described second read request distribution, the wherein said reading page is used for storing described second
The data that read request is corresponding;
Described second read request splits into the request of many height, and wherein split son request number depends on
In the scope of data that described second read request is asked, it is hard that each described sub-request is used for reading correspondence
Data in dish;
If there is slow IO situation in any one of sub-request, and the described son of slow IO situation occurs
Ask to ask for redundancy, then the request of described redundancy is carried out individual disc redundancy array system fall
Level is read, and returns the data that the request of described redundancy is corresponding, and wherein said redundancy is asked as passing through
Individual disc redundancy array system degradation reads to return the son request of data.
In the first possible implementation of first aspect, described method also includes: obtain each institute
State the sub-request deviant relative to described second read request, and record in each described sub-request
Respective deviant and the address of described second read request.
In conjunction with the first possible implementation of first aspect, the second in first aspect may realize
In mode, the request of described redundancy is carried out individual disc redundancy array system degradation and reads, return institute
State the data that the request of redundancy is corresponding, including:
By replicating the field information of described redundancy request, generate a new sub-request, wherein said
New son request is for reading the data in the hard disk that the request of described redundancy is corresponding;
According to the address of the second read request described in described new sub-request, find described second and read to ask
Ask, and redistribute the reading page for described second read request;
According to the redundancy relationship between hard disk and other hard disks that described new sub-request is corresponding, to described its
The data of his hard disk corresponding position carry out verification and calculate, and obtain the data that described new sub-request is corresponding,
That wherein redistributes reads the subpage frame that the page is each described sub-request distribution storage data, other
Hard disk is the son request correspondence in all sub-request that fractionation obtains in addition to described redundancy is asked
Hard disk.
The first in conjunction with first aspect or first aspect may implementation or the second of first aspect
Plant possible implementation, in the third possible implementation of first aspect, it is determined that process the time
There is slow IO situation in the described sub-request exceeding threshold value.
In conjunction with in the third possibility implementation of first aspect, in the 4th kind of possible reality of first aspect
In existing mode, determine that the described sub-request with Redundant representations is the request of redundancy.
The first in conjunction with first aspect or first aspect may implementation or the second of first aspect
Planting possible implementation, in the 5th kind of possible implementation of first aspect, described method is also wrapped
Include:
Discharge the request of described redundancy, and reclaim the son that the described reading page is the request distribution of described redundancy
The page.
Second aspect, the embodiment of the present invention also provides for a kind of request processing equipment, including:
Receive unit, after the first read request that Receiving Host sends, generate independent disk redundancy battle array
Second read request of row system, described second read request is for from described RAID system
The hard disk of system reads data;
Allocation unit, for the reading page for described second read request distribution, the wherein said reading page is used
In storing the data that described second read request is corresponding;
Split cells, for described second read request being split into the request of many height, wherein split
Son request number depends on the scope of data that described second read request is asked, each described sub-request
For reading the data in corresponding hard disk;
Data capture unit, if there is slow IO situation for any one of sub-request, and occurs
The described sub-request of slow IO situation is the request of redundancy, and the request of described redundancy is carried out independent disk
Redundancy array system degradation is read, and returns the data that the request of described redundancy is corresponding, wherein said redundancy
Son request is for reading to return the son request of data by individual disc redundancy array system degradation.
In the first possible implementation of second aspect, described equipment also includes: acquiring unit,
For obtaining each described sub-request deviant relative to described second read request, and in each institute
State and sub-request records respective deviant and the address of described second read request.
In conjunction with the first possible implementation of second aspect, the second in second aspect may realize
In mode, described data capture unit, including:
Request generates subelement, for there is slow IO situation when any one of sub-request, and goes out
When now the described sub-request of slow IO situation is the request of redundancy, by replicating the request of described redundancy
Field information, generates a new sub-request, and wherein said new sub-request is used for reading described redundancy
Data in the hard disk that request is corresponding;
Search subelement, for the address according to described second read request in described new sub-request, look into
Find described second read request, and redistribute the reading page for described second read request;
Data obtain subelement, between hard disk and other hard disk corresponding according to described new sub-request
Redundancy relationship, the data of other hard disk corresponding positions described are carried out verification and calculate, obtain institute
Stating the data that new son request is corresponding, the reading page wherein redistributed is each described sub-request distribution
The subpage frame of storage data, other hard disks are to split in all sub-request obtained except described redundancy
The hard disk of the son request correspondence outside request.
The first in conjunction with second aspect or second aspect may implementation or the second of second aspect
Plant possible implementation, in the third possible implementation of second aspect, described data acquisition
There is slow IO situation in the described sub-request that unit exceedes threshold value specifically for: determination processing time.
In conjunction with the third possible implementation of second aspect, in the 4th kind of possible realization of second aspect
In mode, described data capture unit further specifically for: determine that there is the described of Redundant representations
Son request is asked for redundancy.
The first in conjunction with second aspect or second aspect may implementation or the second of second aspect
Planting possible implementation, in the 5th kind of possible implementation of second aspect, described equipment also wraps
Include: recovery unit, be used for discharging the request of described redundancy, and to reclaim the described reading page be described superfluous
The subpage frame of minor request distribution.
Compared with prior art, the present invention includes advantages below:
In embodiments of the present invention, after receiving the first read request that main frame sends, generate one and be used for
From the hard disk of RAID system, read the second read request of data, and be that page is read in the second read request distribution
Face.Second read request is split into many height and asks to read the data in hard disk, when any one
There is slow IO situation in son request, and occurs when the son request of slow IO situation is asked for redundancy,
In the case of returning without waiting for the request of this redundancy, i.e. the request of redundancy can be carried out RAID system
System degradation is read, and returns the data that the request of redundancy is corresponding, it is achieved that to the redundancy slow IO situation occur
Son request quickly returns data, solves to occur once in a while that the son request of slow IO cannot quickly return data
Problem.After the son request slow IO situation occur returns, it is also possible to reclaim this sub-request, it is to avoid should
After son request long-time return data, host read request is processed and impact.
Further, there are the feelings of slow IO in sub-request in the request processing method that the embodiment of the present invention provides
Under condition, still retain the hard disk of its correspondence, so when its revert to can the hard disk of normal process slow IO time,
Can utilize the data in this hard disk that other slow dishes carry out RAID system degradation to read, reduce double plate and lose
The risk of effect.RAID system is without being reconstructed the data of hard disk simultaneously, it is ensured that whole RAID
The IO performance of system.
Accompanying drawing explanation
For the technical scheme being illustrated more clearly that in the embodiment of the present invention, embodiment will be described below
The accompanying drawing used required in is briefly described, it should be apparent that, the accompanying drawing in describing below is only
It is only some embodiments of the present invention, for those of ordinary skill in the art, is not paying wound
On the premise of the property made is laborious, it is also possible to obtain other accompanying drawing according to these accompanying drawings.
Fig. 1 is a kind of flow chart of a kind of request processing method that the embodiment of the present invention provides;
Fig. 2 is the another kind of flow chart of a kind of request processing method that the embodiment of the present invention provides;
Fig. 3 is a kind of a kind of structural representation asking processing equipment that the embodiment of the present invention provides;
Fig. 4 is a kind of another kind of structural representation asking processing equipment that the embodiment of the present invention provides;
Fig. 5 is a kind of kernel texture schematic diagram asking processing equipment that the embodiment of the present invention provides;
Fig. 6 is a kind of yet another construction schematic diagram asking processing equipment that the embodiment of the present invention provides.
Detailed description of the invention
Below in conjunction with the accompanying drawing in the embodiment of the present invention, to the technical scheme in the embodiment of the present invention
It is clearly and completely described, it is clear that described embodiment is only that a part of the present invention is real
Execute example rather than whole embodiments.Based on the embodiment in the present invention, ordinary skill
The every other embodiment that personnel are obtained under not making creative work premise, broadly falls into this
The scope of invention protection.
Refer to Fig. 1, it illustrates the one of a kind of request processing method that the embodiment of the present invention provides
Flow chart, may comprise steps of:
Step 101: after the first read request that Receiving Host sends, generates the second reading of RAID system
Request, the second read request is for reading data from the hard disk of RAID system.
It is to be understood that the first read request represents that main frame needs are from each hard disk of RAID system
Read data.After RAID system receives this first read request, generate for from RAID system
Hard disk in read the second read request of data.
Step 102: be the reading page of the second read request distribution, wherein reads the page for storage the second reading
The data that request is corresponding.
RAID system distributes for the reading page storing the returned data of each hard disk by the second read request
Face, this reading page is the data space that the distribution of each hard disk is independent, after hard disk returns data,
RAID system is stored in the data space of correspondence.
Step 103: the second read request is split into the request of many height.
When the second read request is on the path being issued to hard disk, and split into the request of many height, Qi Zhongsuo
The son request number split depends on the scope of data that the second read request is asked.Such as certain second
The scope of data that read request is asked is 0-100, then during the second read request issues, and RAID
The RAID layer of system determines which hard disk is this scope of data be stored in, by the quantity of these hard disks
As splitting into the quantity of sub-request, the every height request after fractionation is respectively used to read corresponding hard disk
In data.
Concrete, request split process according to RAID system and LUN(Logical Unit Number,
LUN) specific object carry out, it splits identical with existing request, real to this present invention
Execute the process that example no longer illustrates that request splits.And the mapping relations of son request and hard disk follow RAID
In system each RAID definition mapping relations and computing formula, with existing sub-request and hard disk
Map identical, this present embodiment is no longer elaborated.
Step 104: if slow IO situation occurs in the request of any one height, and the son of slow IO situation occurs
Ask to ask for redundancy, then the request of redundancy is carried out RAID system degradation and read, return redundancy
The data that request is corresponding, wherein the request of redundancy is for reading to return data by RAID system degradation
Son request.
In embodiments of the present invention, RAID system can include an intervalometer, is used for recording every height
The process time of request, when the process time of group request exceedes threshold value, the RAID of RAID system
The determination processing time is exceeded this sub-request of threshold value and slow IO situation occurs by layer.
Further, group is asked during issuing, and can determine superfluous according to the rank of RAID system
The number of minor request.Such as when RAID system rank is RAID5 rank, RAID system
In all hard disks the most non-off-line time all sub-requests be redundancy request, if there being more than one hard disk
During off-line, all sub-requests are all irredundant sub-requests.When RAID system rank is RAID6 rank
Time, in RAID system the most non-off-line of all hard disks or only a hard disk off-line time, all sons please
Ask and be the request of redundancy, if all sub-requests are all irredundant sons when having two or more hard disk off-line
Request, the most irredundant sub-request refers to that the son that cannot pass through RAID system degradation reading return data please
Ask.
When RAID layer judges that slow IO situation occurs in son request, and when this sub-request is the request of redundancy,
RAID layer carries out RAID system degradation to it and reads, and wherein the reading of RAID system degradation is by reading
Other data in hard disks in addition to this redundancy corresponding hard disk of request in RAID system, then pass through
RAID algorithm calculates the data that the request of this redundancy needs to read.Wherein read in other hard disks
The hard disk asked of data and this redundancy in data there is same memory location and length
Degree, other hard disks are to split the son in addition to asking in all sub-request that obtains except described redundancy to ask
Seek corresponding hard disk.
Such as: when RAID system rank is RAID5, the hard disk that its all sub-requests are corresponding is respectively
For disk1, disk2, disk3, disk1 to be read is stored in position 1 a length of 100
Data, if disk1 corresponding son request for redundancy ask, then can first read disk2 with
It is stored in the data of position 1 a length of 100 on disk3, then the data read on the two dish are entered
Row even-odd check, it is possible to calculate the data being stored in position 1 a length of 100 on disk1.
Application technique scheme, after receiving the first read request that main frame sends, generates one and is used for
From the hard disk of RAID system, read the second read request of data, and be that page is read in the second read request distribution
Face.Second read request is split into many height and asks to read the data in hard disk, when any one
There is slow IO situation in son request, and occurs when the son request of slow IO situation is asked for redundancy,
In the case of returning without waiting for the request of this redundancy, i.e. the request of redundancy can be carried out RAID system
System degradation is read, and returns the data that the request of redundancy is corresponding, it is achieved that to the redundancy slow IO situation occur
Son request quickly returns data, solves to occur once in a while that the son request of slow IO cannot quickly return data
Problem.After the son request slow IO situation occur returns, it is also possible to reclaim this sub-request, it is to avoid should
After son request long-time return data, host read request is processed and impact.
Further, there are the feelings of slow IO in sub-request in the request processing method that the embodiment of the present invention provides
Under condition, still retain the hard disk of its correspondence, so when its revert to can the hard disk of normal process slow IO time,
Can utilize the data in this hard disk that other slow dishes carry out RAID system degradation to read, reduce double plate and lose
The risk of effect.RAID system is without being reconstructed the data of hard disk simultaneously, it is ensured that whole RAID
The IO performance of system.
Refer to Fig. 2, it illustrates another of a kind of request processing method that the embodiment of the present invention provides
Plant flow chart, elaborate, how by the process of RAID system degradation reading return data, can include
Following steps:
Step 201: after the first read request that Receiving Host sends, generates the second reading of RAID system
Request, the second read request is for reading data from the hard disk of RAID system.
Step 202: be the reading page of the second read request distribution, wherein reads the page for storage the second reading
The data that request is corresponding.
Step 203: the second read request is split into the request of many height.
Step 201 is to step 203: identical with step 101 to step 103, implements this present invention
Example is no longer illustrated by.
Step 204: obtain every height request deviant relative to the second read request, and at every height
Request records respective deviant and the address of the second read request.
The initial address of the such as second read request is lba=0, a length of lenth=100, and this second reading please
Ask and be split into two sons request REQ1 and REQ2, wherein REQ1lba=0lenth=20, REQ2
Lba=20lenth=100, then in the request of the two, REQ1 is relative to the second read request REQ0
Skew be 0, REQ2 is 20 relative to the skew of the second read request REQ0.
The each layer split at RAID system due to request all can be carried out, so a hard disk is corresponding
Son request is probably the second read request and generates after n times split, so needing when splitting every time
Record the address of the second read request, so can also be by ground second reading in the case of sub-request timed out
Second read request of its correspondence of address search of request, improves the search procedure of the second read request.
Step 205: if slow IO situation occurs in the request of any one height, and the son of slow IO situation occurs
Asking to ask for redundancy, by replicating the field information of redundancy request, generating a new son please
Asking, wherein new son request is for reading the data in the hard disk that the request of redundancy is corresponding.
In embodiments of the present invention, RAID system includes an intervalometer, is used for recording the request of every height
The process time, this intervalometer can be added on block device layer.The process time of group request exceedes
During threshold value, the RAID layer determination processing time of RAID system exceedes the described sub-request of threshold value to be occurred
Slow IO situation.When group request has Redundant representations further, determine that described sub-request is redundancy
Request.The determination of redundancy request refers to the explanation in step 104, is not illustrating this.
In addition block device layer is after receiving sub-request, if be detected that the son request received has superfluous
Remaining labelling, then generate a new sub-request, and this new sub-request is by replicating all of redundancy request
Field generates.This new sub-request is labeled with " quick back page replacement " labelling with failure further
Result is directly returned to RADI layer, and wherein quickly back page replacement is used for prompting for the second read request
Redistribute the reading page.
Step 206: according to the address of the second read request in new son request, find the second read request,
And be that the second read request redistributes the reading page.
Step 207: according to the redundancy relationship between hard disk and other hard disks that new son request is corresponding, right
The data of other hard disk corresponding positions carry out verification and calculate, and obtain the data that new son request is corresponding,
That wherein redistributes reads the subpage frame that the page is every height request distribution storage data, other hard disks
For splitting the hard disk of the son request correspondence in addition to asking in all sub-request that obtains except redundancy.
After RAID layer receives the request that failure result returns, it is judged that whether this request has " quickly returns
The page is replaced " labelling, if had, need to carry out page replacement according in request: with according to sub-request
Directly find " original read request " and apply for the address of the second new read request, finding
Two read requests, and be that the second read request redistributes the reading page, the reading page that this is redistributed is every
Height request distributes data space again, this data space can with distribute before
Data space is different, the data of distribution before using the data space redistributed to replace it
Memory space, distributes for it before to avoid the request of redundancy to take after long-time return data
Data space so which data main frame cannot be it is determined that reads, to main machine-readable data
Impact, cause the waste of memory space simultaneously.
Further, RAID layer closes according to the redundancy between hard disk and other hard disks of new son request correspondence
The data of other hard disk corresponding positions are carried out verifying calculating by system, obtain new son request correspondence
Data, complete RADI system degradation and read, thus without waiting for the return of time-out IO.
It should be noted is that: every height request may read the data of certain position in hard disk,
Then read to be from other hard disks, read the data at same position carrying out RADI system degradation.Entering
When row verification calculates, RAID layer is previously stored with the redundancy relationship between hard disk, is therefore receiving
After new sub-request, can be according to the hard disk of new son request correspondence, from the redundancy relationship prestored
Find other hard disks with it with redundancy relationship.
Additionally, block device layer is after detecting that the son request received has redundant marks, add " after
Platform reclaims " during labelling asks to redundancy, and issue abort order to SCSI layer and terminate this redundancy
Son request.There is the redundancy request of " backstage recovery " labelling when being back to RAID layer, RAID
Judge whether labelling " backstage recovery " labelling, if had, directly this redundancy request, and reclaim
Read the subpage frame that the page is the request distribution of redundancy, it is to avoid host read request is processed and impacts.
Corresponding with said method embodiment, the embodiment of the present invention also provides for a kind of request processing equipment,
Its structural representation is as it is shown on figure 3, may include that reception unit 11, allocation unit 12, split
Unit 13 and data capture unit 14, wherein:
Receive unit 11, after the first read request that Receiving Host sends, generate RAID system
The second read request, the second read request for from the hard disk of RAID system read data.
It is to be understood that the first read request represents that main frame needs are from each hard disk of RAID system
Read data.After RAID system receives this first read request, generate for from RAID system
Hard disk in read the second read request of data.
Allocation unit 12, for being the reading page of the second read request distribution, wherein reads the page and is used for depositing
Store up the data that the second read request is corresponding.This reading page i.e. is the data storage that the distribution of each hard disk is independent
Space, after hard disk returns data, RAID system is stored in the data space of correspondence.
Split cells 13, for splitting into the request of many height by the second read request.
When the second read request is on the path being issued to hard disk, and split into the request of many height, Qi Zhongsuo
The son request number split depends on the scope of data that the second read request is asked.Such as certain second
The scope of data that read request is asked is 0-100, then during the second read request issues, and RAID
The RAID layer of system determines which hard disk is this scope of data be stored in, by the quantity of these hard disks
As splitting into the quantity of sub-request, the every height request after fractionation is respectively used to read corresponding hard disk
In data.
Concrete, request split process is carried out according to the specific object of RAID system and LUN, its
Split identical with existing request, this embodiment of the present invention is no longer illustrated the process that request splits.And
And son is asked and the mapping relations of hard disk follow the mapping of each RAID definition in RAID system
Relation and computing formula, identical with existing sub-request and hard disk mapping, the most detailed to this present embodiment
Thin elaboration.
, there is slow IO situation for the request of any one height, and slow IO occurs in data capture unit 14
The son request of situation is asked for redundancy, then the request of redundancy is carried out RAID system degradation and read, return
Returning the data that the request of redundancy is corresponding, wherein the request of redundancy is for returning by RAID system degradation reading
Return the son request of data.
Wherein, data capture unit 14 exceedes the son request of threshold value specifically for: determination processing time
Slow IO situation occurs.When group request has Redundant representations further, data capture unit 14 is true
Stator request is asked for redundancy.
RAID system degradation reading is by reading in RAID system except the corresponding hard disk of this redundancy request
Outside other data in hard disks, then by RAID algorithm calculate this redundancy request need read
The data taken.Wherein read the hard disk that the data in other hard disks are asked with this redundancy
In data there is same memory location and length.
Such as: when RAID system rank is RAID5, the hard disk that all sub-requests are corresponding is respectively
Disk1, disk2, disk3, disk1 to be read is stored in the number of position 1 a length of 100
According to, if disk1 corresponding son request for redundancy request, then can first read disk2 and disk3
On be stored in the data of position 1 a length of 100, then the data read on the two dish are carried out very
Even parity check, it is possible to calculate the data being stored in position 1 a length of 100 on disk1.
, when slow IO situation occurs in the request of any one height, and there is slow IO in application technique scheme
When the son request of situation is asked for redundancy, in the case of returning without waiting for the request of this redundancy,
I.e. the request of redundancy can carry out RAID system degradation to read, return the number that the request of redundancy is corresponding
According to, it is achieved that the redundancy request to there is slow IO situation quickly returns data, and solution occurs once in a while
The problem that the son request of slow IO cannot quickly return data.When the son request slow IO situation occur is returned
Hui Hou, it is also possible to reclaim this sub-request, it is to avoid main frame is read after returning data for a long time by this sub-request
Request processes and impacts.
Further, in the case of slow IO occurs in sub-request, the hard disk of its correspondence is still retained, so
When its revert to can the hard disk of normal process slow IO time, it is possible to use the data in this hard disk are to other
Slow dish carries out RAID system degradation and reads, and reduces the risk that double plate lost efficacy.Simultaneously RAID system without
The data of hard disk are reconstructed, it is ensured that the IO performance of whole RAID system.
On the basis of Fig. 3, the request processing equipment that the embodiment of the present invention provides can also be wrapped further
Include: acquiring unit 15, as shown in Figure 4.Wherein acquiring unit 15, are used for obtaining the request of every height
Relative to the deviant of the second read request, and in every height is asked, record respective deviant and the
The address of two read requests.
The each layer split at RAID system due to request all can be carried out, so a hard disk is corresponding
Son request is probably the second read request and generates after n times split, so needing when splitting every time
The address of the second read request is recorded in sub-request by acquiring unit 15, so at sub-request timed out
In the case of can also improve the by second read request of its correspondence of address search of ground two read requests
The search procedure of two read requests.
In the request processing equipment shown in above-mentioned Fig. 3 and Fig. 4, the structure of data capture unit 14
Schematic diagram can be refering to shown in Fig. 5, and the request that may include that generates subelement 141, searches subelement
142 and data obtain subelement 143.Wherein,
Request generates subelement 141, for there is slow IO situation when the request of any one height, and goes out
The now son request of slow IO situation is asked for redundancy, by replicating the field information of redundancy request,
Generating a new sub-request, wherein new son request is for reading in the hard disk that the request of redundancy is corresponding
Data.
In embodiments of the present invention, request generates the process that subelement 141 can be asked by judgement
Time determines whether sub-request slow IO situation occurs, when the process time of group request exceedes threshold value,
The request generation subelement 141 determination processing time exceedes the son request of threshold value and slow IO situation occurs.Enter
When one step group request has Redundant representations, request generates subelement 141 and determines that sub-request is redundancy
Son request.
Search subelement 142, for the address according to the second read request in new son request, find
Second read request, and be that the second read request redistributes the reading page.The reading page that this is redistributed is
Every height request distributes data space again, this data space can with distribute before
Data space different, with avoid the request of redundancy through long-time return data after take it
The front data space for its distribution so that main frame cannot be right it is determined that read which data
Main machine-readable data impacts, and causes the waste of memory space simultaneously.
Data obtain subelement 143, between hard disk and other hard disk corresponding according to new son request
Redundancy relationship, the data of other hard disk corresponding positions are carried out verification and calculate, obtaining new son please
Seek corresponding data, wherein read the subpage frame that the page is every height request distribution storage data, other
Hard disk is the son request correspondence in all sub-request that fractionation obtains in addition to described redundancy is asked
Hard disk.
It should be noted is that: every height request may read the data of certain position in hard disk,
Then read to be from other hard disks, read the data at same position carrying out RADI system degradation.Example
As: the data in hard disk disk1 position 2 are read in sub-request 1, if fruit request 1 is for there is slow IO
The redundancy request of situation, then disk2 and disk3 with it with corresponding relation can read each self-alignment
Put the data in 2 and carry out even-odd check, calculate the data being stored on disk1 in position 2, complete
Become the process that RAID system degradation is read.
Additionally, the request processing equipment that the embodiment of the present invention provides can further include: reclaim single
Unit, is used for discharging the request of redundancy, and reclaims the subpage frame that the reading page is the request distribution of redundancy,
To avoid redundancy to ask when returning to data to subpage frame, cause main frame cannot determine subpage
In face, the data of storage which read request corresponding, impacts host read request.
The embodiment of the present invention also provides for another kind of request processing equipment, and its structural representation can be refering to figure
Shown in 6, may include that in data-interface 21(figure and be not drawn into), memorizer 22 and controller 23.
Wherein, data-interface 21 is connected with the data-interface of main frame 24 by data connection device, as
Data-interface 21 constitutes communication link by netting twine and is connected with the data-interface of main frame 24, to receive
The first read request that main frame 24 sends, and the first read request is sent to controller by inner passage
23。
Controller 23, after receiving the first read request, generates the second read request of RAID system,
And during the second read request issues, split into the request of many height.When any one of sub-request
Slow IO situation occurs, and occurs when the described sub-request of slow IO situation is the request of redundancy, to institute
State the request of redundancy and carry out RAID system degradation reading, return the data that the request of described redundancy is corresponding,
The request of redundancy refers to read to return by RAID system degradation the request of data..
In embodiments of the present invention, wherein the second read request is used for reading from the hard disk of RAID system
Data, the son request number split by the second read request depends on what described second read request was asked
Scope of data, each described sub-request is for reading the data in corresponding hard disk.
After controller 23 generates the second read request, divide for described second read request in memorizer 22
The reading page joined, wherein reads the page and is used for the data that storage the second read request is corresponding, i.e. split into
The data that every height request returns all are stored in the reading page.
It should be noted that each embodiment in this specification all uses the mode gone forward one by one to describe,
What each embodiment stressed is the difference with other embodiments, between each embodiment
Identical similar part sees mutually.For device class embodiment, due to itself and method
Embodiment basic simlarity, so describe is fairly simple, relevant part sees the portion of embodiment of the method
Defend oneself bright.
Finally, in addition it is also necessary to explanation, in this article, the relation of such as first and second or the like
Term is used merely to separate an entity or operation with another entity or operating space, and not
Necessarily require or imply and there is the relation of any this reality or suitable between these entities or operation
Sequence.And, term " includes ", " comprising " or its any other variant are intended to non-exclusive
Comprising, so that include the process of a series of key element, method, article or equipment not only of property
Including those key elements, but also include other key elements being not expressly set out, or also include for
The key element that this process, method, article or equipment are intrinsic.In the situation not having more restriction
Under, statement " including ... " key element limited, it is not excluded that including described key element
Process, method, article or equipment there is also other identical element.
For convenience of description, it is divided into various unit to be respectively described with function when describing apparatus above.Certainly,
The function of each unit can be realized in same or multiple softwares and/or hardware when implementing the present invention.
As seen through the above description of the embodiments, those skilled in the art can be clearly
Solve the present invention to add the mode of required general hardware platform by software and realize.Based on so
Understanding, the part that prior art is contributed by technical scheme the most in other words can
Embodying with the form with software product, this computer software product can be stored in storage medium
In, such as ROM/RAM, magnetic disc, CD etc., instruct with so that a computer sets including some
Standby (can be personal computer, server, or the network equipment etc.) performs the present invention, and each is real
Execute the method described in some part of example or embodiment.
Above a kind of request processing method provided by the present invention and equipment are described in detail, this
Literary composition applies specific case principle and the embodiment of the present invention are set forth, above enforcement
The explanation of example is only intended to help to understand method and the core concept thereof of the present invention;Simultaneously for this
The those skilled in the art in field, according to the thought of the present invention, in detailed description of the invention and range of application
On all will change, in sum, this specification content should not be construed as the limit to the present invention
System.
Claims (12)
1. a request processing method, it is characterised in that including:
After the first read request that Receiving Host sends, generate the second reading of individual disc redundancy array system
Request, described second read request is for reading from the hard disk of described individual disc redundancy array system
Data;
For the reading page of described second read request distribution, the wherein said reading page is used for storing described second
The data that read request is corresponding;
Described second read request splits into the request of many height, and wherein split son request number depends on
In the scope of data that described second read request is asked, it is hard that each described sub-request is used for reading correspondence
Data in dish;
If there is slow IO situation in any one of sub-request, and the described son of slow IO situation occurs
Ask to ask for redundancy, then the request of described redundancy is carried out individual disc redundancy array system fall
Level is read, and returns the data that the request of described redundancy is corresponding, and wherein said redundancy is asked as passing through
Individual disc redundancy array system degradation reads to return the son request of data;
Wherein, slow IO refers to slow input/output.
Method the most according to claim 1, it is characterised in that described method also includes: obtain
Take each described sub-request deviant relative to described second read request, and please at each described son
Ask the middle address recording respective deviant and described second read request.
Method the most according to claim 2, it is characterised in that described redundancy is asked into
Row individual disc redundancy array system degradation is read, and returns the data that the request of described redundancy is corresponding, bag
Include:
By replicating the field information of described redundancy request, generate a new sub-request, wherein said
New son request is for reading the data in the hard disk that the request of described redundancy is corresponding;
According to the address of the second read request described in described new sub-request, find described second and read to ask
Ask, and redistribute the reading page for described second read request;
According to the redundancy relationship between hard disk and other hard disks that described new sub-request is corresponding, to described its
The data of his hard disk corresponding position carry out verification and calculate, and obtain the data that described new sub-request is corresponding,
That wherein redistributes reads the subpage frame that the page is each described sub-request distribution storage data, other
Hard disk is the son request correspondence in all sub-request that fractionation obtains in addition to described redundancy is asked
Hard disk.
4. according to the method described in claims 1 to 3 any one, it is characterised in that at judgement
The reason time exceedes the described sub-request of threshold value and slow IO situation occurs.
Method the most according to claim 4, it is characterised in that determine that there is Redundant representations
Described sub-request is the request of redundancy.
6. according to the method described in claims 1 to 3 any one, it is characterised in that described side
Method also includes:
Discharge the request of described redundancy, and reclaim the son that the described reading page is the request distribution of described redundancy
The page.
7. a request processing equipment, it is characterised in that including:
Receive unit, after the first read request that Receiving Host sends, generate independent disk redundancy battle array
Second read request of row system, described second read request is for from described RAID system
The hard disk of system reads data;
Allocation unit, for the reading page for described second read request distribution, the wherein said reading page is used
In storing the data that described second read request is corresponding;
Split cells, for described second read request being split into the request of many height, wherein split
Son request number depends on the scope of data that described second read request is asked, each described sub-request
For reading the data in corresponding hard disk;
Data capture unit, if there is slow IO situation for any one of sub-request, and occurs
The described sub-request of slow IO situation is the request of redundancy, and the request of described redundancy is carried out independent disk
Redundancy array system degradation is read, and returns the data that the request of described redundancy is corresponding, wherein said redundancy
Son request is for reading to return the son request of data by individual disc redundancy array system degradation;
Wherein, slow IO refers to slow input/output.
Equipment the most according to claim 7, it is characterised in that described equipment also includes: obtain
Take unit, for obtaining each described sub-request deviant relative to described second read request, and
Respective deviant and the address of described second read request is recorded in each described sub-request.
Equipment the most according to claim 8, it is characterised in that described data capture unit,
Including:
Request generates subelement, for there is slow IO situation when any one of sub-request, and goes out
When now the described sub-request of slow IO situation is the request of redundancy, by replicating the request of described redundancy
Field information, generates a new sub-request, and wherein said new sub-request is used for reading described redundancy
Data in the hard disk that request is corresponding;
Search subelement, for the address according to described second read request in described new sub-request, look into
Find described second read request, and redistribute the reading page for described second read request;
Data obtain subelement, between hard disk and other hard disk corresponding according to described new sub-request
Redundancy relationship, the data of other hard disk corresponding positions described are carried out verification and calculate, obtain institute
Stating the data that new son request is corresponding, the reading page wherein redistributed is each described sub-request distribution
The subpage frame of storage data, other hard disks are to split in all sub-request obtained except described redundancy
The hard disk of the son request correspondence outside request.
10. according to the equipment described in claim 7 to 9 any one, it is characterised in that described number
Specifically for the described sub-request that: determination processing time exceedes threshold value, slow IO feelings occur according to acquiring unit
Condition.
11. equipment according to claim 10, it is characterised in that described data capture unit
Further specifically for: determine that having the described sub-request of Redundant representations is the request of redundancy.
12. according to the equipment described in claim 7 to 9 any one, it is characterised in that described in set
Standby also include: recovery unit, be used for discharging the request of described redundancy, and reclaim the described reading page and be
The subpage frame of described redundancy request distribution.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310687352.4A CN103699340B (en) | 2013-12-16 | 2013-12-16 | A kind of request processing method and equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310687352.4A CN103699340B (en) | 2013-12-16 | 2013-12-16 | A kind of request processing method and equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103699340A CN103699340A (en) | 2014-04-02 |
CN103699340B true CN103699340B (en) | 2016-12-07 |
Family
ID=50360881
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310687352.4A Active CN103699340B (en) | 2013-12-16 | 2013-12-16 | A kind of request processing method and equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103699340B (en) |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106534242B (en) * | 2015-09-14 | 2019-11-29 | 阿里巴巴集团控股有限公司 | The processing method and device requested in a kind of distributed system |
CN106933495B (en) * | 2015-12-30 | 2019-07-23 | 华为技术有限公司 | A kind of method for reading data, RAID controller and storage equipment |
CN110727401B (en) * | 2019-09-09 | 2021-03-02 | 无锡江南计算技术研究所 | Memory access system |
CN110896415B (en) * | 2019-11-22 | 2022-05-24 | 浪潮电子信息产业股份有限公司 | Data readdir method, system, equipment and computer medium |
CN116661708B (en) * | 2023-07-31 | 2023-11-03 | 苏州浪潮智能科技有限公司 | Processing method of read-write task based on hard disk array and electronic equipment |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101055557A (en) * | 2006-04-13 | 2007-10-17 | 国际商业机器公司 | Data processing system and method |
CN101657802A (en) * | 2006-12-06 | 2010-02-24 | 弗森多系统公司(dba弗森-艾奥) | The device, the system and method that are used for long-range direct memory access (DMA) solid storage device |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9171044B2 (en) * | 2010-02-16 | 2015-10-27 | Oracle International Corporation | Method and system for parallelizing database requests |
US9430291B2 (en) * | 2010-12-30 | 2016-08-30 | International Business Machines Corporation | Distributed topology enabler for identity manager |
-
2013
- 2013-12-16 CN CN201310687352.4A patent/CN103699340B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101055557A (en) * | 2006-04-13 | 2007-10-17 | 国际商业机器公司 | Data processing system and method |
CN101657802A (en) * | 2006-12-06 | 2010-02-24 | 弗森多系统公司(dba弗森-艾奥) | The device, the system and method that are used for long-range direct memory access (DMA) solid storage device |
Also Published As
Publication number | Publication date |
---|---|
CN103699340A (en) | 2014-04-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107250975B (en) | Data storage system and data storage method | |
US8239584B1 (en) | Techniques for automated storage management | |
CN103699340B (en) | A kind of request processing method and equipment | |
CN109542342B (en) | Metadata management and data reconstruction method, equipment and storage medium | |
WO2017092480A1 (en) | Data migration method and device | |
EP2180407A2 (en) | Fast data recovery from HDD failure | |
US8495295B2 (en) | Mass storage system and method of operating thereof | |
WO2020236353A1 (en) | Memory disaggregation for compute nodes | |
CN103577115B (en) | Arrangement processing method, device and the server of data | |
US9405583B2 (en) | Resource provisioning based on logical profiles and piecewise objective functions | |
CN103763383A (en) | Integrated cloud storage system and storage method thereof | |
US20130013850A1 (en) | Relative heat index based hot data determination for block based storage tiering | |
US20150269000A1 (en) | Resource provisioning based on logical profiles and objective functions | |
US10324794B2 (en) | Method for storage management and storage device | |
CN104267913A (en) | Storage method and system allowing dynamic asynchronous RAID level adjustment | |
CN107729536A (en) | A kind of date storage method and device | |
CN108733326B (en) | Disk processing method and device | |
CN102226892A (en) | Disk fault tolerance processing method and device thereof | |
CN111897486A (en) | Intelligent unified storage system based on software definition | |
CN114442937B (en) | File caching method and device, computer equipment and storage medium | |
US11461250B2 (en) | Tuning data storage equipment based on comparing observed I/O statistics with expected I/O statistics which are defined by operating settings that control operation | |
Thomasian et al. | Mirrored disk organization reliability analysis | |
US11379326B2 (en) | Data access method, apparatus and computer program product | |
US11150991B2 (en) | Dynamically adjusting redundancy levels of storage stripes | |
US20130019122A1 (en) | Storage device and alternative storage medium selection method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |