CN108334278A - A kind of storage system balanced management method and apparatus - Google Patents
A kind of storage system balanced management method and apparatus Download PDFInfo
- Publication number
- CN108334278A CN108334278A CN201711350651.3A CN201711350651A CN108334278A CN 108334278 A CN108334278 A CN 108334278A CN 201711350651 A CN201711350651 A CN 201711350651A CN 108334278 A CN108334278 A CN 108334278A
- Authority
- CN
- China
- Prior art keywords
- ssd
- main force
- writing
- write
- storage system
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0602—Interfaces specially adapted for storage systems specifically adapted to achieve a particular effect
- G06F3/0614—Improving the reliability of storage systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0628—Interfaces specially adapted for storage systems making use of a particular technique
- G06F3/0646—Horizontal data movement in storage systems, i.e. moving data in between storage devices or systems
- G06F3/0647—Migration mechanisms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0668—Interfaces specially adapted for storage systems adopting a particular infrastructure
- G06F3/0671—In-line storage system
- G06F3/0673—Single storage device
- G06F3/0679—Non-volatile semiconductor memory device, e.g. flash memory, one time programmable memory [OTP]
Abstract
The invention discloses a kind of storage system balanced management method and apparatus, this method includes：When system is in ageing state, a solid-state hard disk SSD is selected, and set the selected SSD to the main force and SSD is written；When selected main force write-in SSD scraps, removable disk operation is executed；After current main force write-in SSD scraps, reselects a SSD and SSD is written as the main force.Scheme through the invention so that the case where bad SSD disks controllably, in an orderly manner, compartment of terrain occur, to which for smoothly removable disk, smoothly migrating data creates the time, and influence business as small as possible, while ensureing the storage service q＆r of system.
Description
Technical field
The present invention relates to computer field of storage, espespecially a kind of storage system balanced management method and apparatus.
Background technology
Solid state disk (SSD) is gradually deposited as computer in recent years because it is in the advantage of performance, volume, energy consumption etc.
The Primary memory part of storage system substitutes disk and has become trend of the times.SSD is by a flash controller and a large amount of flash cells
This two parts forms.Wherein, flash cell has the erasable number of limited number of time, and there are the upper limits in service life, and storage density is bigger
Its service life of flash cell it is shorter.Just because of this, the method that flash controller uses abrasion equilibrium so that each sudden strain of a muscle inside SSD
The erasable number of memory cell is close as possible, avoid wherein a small number of flash cells concentrated it is erasable, reach the erasable limit, appearance first
Exception reports an error and scraps totally.Obviously, the abrasion equilibrium of each flash cell in the inside SSD, is conducive to the maximization in single-deck service life, has
Conducive to the enhancing of single-deck reliability.
It is and previously described all using SSD as in the storage system (hereinafter referred to as full flash memory storage system) of memory device
Single solid-state disk there is a situation where similar but not exactly the same.It (includes but not limited to distributed file system, distribution to store software
Block storage system, distributed objects storage system, distributed data base, flash memory RAID system etc.) also indiscriminately imitate the side of abrasion equilibrium
Method so that erasable number between each SSD is close as possible, the degree of wear is close, avoids wherein minority SSD from being concentrated erasable, first
First reach and scraps totally.Obviously, the abrasion equilibrium between SSD, the whole reliability services time for being conducive to system maximize, and will be
The appearance of low-quality disk is delayed as possible for the first time in system, and the original intention of this kind of method is to be worth affirmative.
But the macroscopic view between polydisc indiscriminately imitates the abrasion equilibrium method in the case of microcosmic in single-deck, there is weight
Big defect：Can occur the case where low-quality disk simultaneously when bulk life time reaches aging, have little time the flash memory device more renewed, cause
Storage service is promptly unavailable, unreliable.
For example, as shown in Figure 1, being each SSD service life versus time curve in conventional wear equalization methods.From Fig. 1
It can be seen that using the conventional method of abrasion equilibrium, the service life decline curve of each SSD is sufficiently close in system, used in cluster
When by 20th month, there is the case where concentrating low-quality disk, have little time the replacement for replacing new and old disk.
For example at most tolerate the damage of 1 piece of SSD disk in the configuration of typical two copy again, the 2nd piece of SSD adjustment debit occurs
Before bad, it is necessary to complete to replace SSD disks, that is, include：(1) new building changes the manual operations of low-quality disk；(2) Data Migration is soft to new building
Part operates.Replacing the principal risk that SSD dribblings come includes：(1) improperly manual operations easily causes new low-quality disk；(2) data are moved
It moves past journey and generates a large amount of I/O operations and easily cause new low-quality disk；(3) Data Migration speed is unsuitable too fast, because will be provided with front-end business
Source competes, and can not ensure storage service quality.Once occurring the 2nd piece of SSD damage before replacing SSD disks and finishing, system will be complete
Full collapse, goes out the serious error of active data.User has to make between service quality and reliability to be selected without mercy
Select --- rapidly removable disk will cause service quality low, and the risk that slowly removable disk will cause reliability to lack.It is regarded from this
From the point of view of angle, the generation of low-quality disk must be there are certain time interval, to provide the sufficient time to replace SSD disks.Macroscopic view
Although abrasion equilibrium delays the appearance of low-quality disk for the first time as possible, final low-quality disk, which will be concentrated, to be occurred.Have little time more to renew
In the case of flash memory device, storage service can be caused promptly unavailable, unreliable.
Invention content
To solve the above-mentioned problems, the present invention proposes a kind of storage system balanced management method and apparatus, can improve
The storage service q＆r of system.
The present invention proposes a kind of storage system balanced management method, the method includes：
When system is in ageing state, a solid-state hard disk SSD is selected, and set the selected SSD to the main force
SSD is written；
When selected main force write-in SSD scraps, removable disk operation is executed；
After current main force write-in SSD scraps, reselects a SSD and SSD is written as the main force.
The invention also provides a kind of storage system balance management device, including memory, processor and it is stored in described
On memory and the computer program that can run on the processor, which is characterized in that the processor executes the calculating
The processing of either method of the present invention is realized when machine program.
Compared with prior art, technical solution provided by the invention includes：When system is in ageing state, one is selected
SSD is written as the main force in SSD；When the selected write-in load main force scraps, removable disk operation is executed；It is write in the current main force
Enter after SSD scraps, reselects a SSD and SSD is written as the main force.Scheme through the invention so that bad SSD disks
The case where controllably, in an orderly manner, compartment of terrain occur, to which for smoothly removable disk, smoothly migrating data creates the time, and as possible
Business is influenced smallly, while ensureing the storage service q＆r of system.
Description of the drawings
The attached drawing in the embodiment of the present invention is illustrated below, the attached drawing in embodiment be for the present invention into one
Step understands, for explaining the present invention together with specification, does not constitute limiting the scope of the invention.
Fig. 1 is each SSD service life versus time curve in abrasion equilibrium method in the related technology；
Fig. 2 is each SSD service life versus time curve in equalization methods provided in an embodiment of the present invention；
Fig. 3 is each SSD service life versus time curve provided by one embodiment of the present invention；
Fig. 4 is the structural schematic diagram that another embodiment of the invention provides；
Fig. 5 is system mode transition figure provided in an embodiment of the present invention；
Fig. 6 is flow chart provided in an embodiment of the present invention；
Fig. 7 is module relation diagram provided in an embodiment of the present invention.
Specific implementation mode
For the ease of the understanding of those skilled in the art, the invention will be further described below in conjunction with the accompanying drawings, not
It can be used for limiting the scope of the invention.It should be noted that in the absence of conflict, the embodiment in the application and reality
The various modes applied in example can be combined with each other.
The present invention proposes a kind of storage system balanced management method, the method includes：
Step 10 determines whether system is in ageing state；
Step 20, when system is in non-ageing state, each SSD balancedly undertakes writing in setting system；
Step 30, when system is in ageing state, a SSD is selected to load the main force as write-in, and it is negative that write-in is arranged
It carries main force SSD and preferentially undertakes writing.
In the embodiment of the present invention, when system is in ageing state and non-ageing state, different abrasion equilibriums can be taken
Strategy；
It is using traditional abrasion equilibrium method, i.e., each in system when system is in younger state, i.e. non-ageing state
SSD balancedly undertakes writing.When system is in ageing state, a SSD is selected as the write-in load main force.In this way
Control method, disperse the time that each SSD scraps, improve multiple SSD failure times and get too close to, or even occur one simultaneously
The case where SSD more than block scraps improves the performance of storage system.
When younger state includes but not limited to system initial on-line running, entirety SSD is when being changed to new building etc..At this time
The time that SSD distances are scrapped is longer, according to traditional abrasion equilibrium method so that system performance is in excellent condition.
The storage system balanced management method provided through the embodiment of the present invention proposes reverse in a kind of full flash memory storage system
The method of abrasion equilibrium so that full flash memory storage system accomplishes abrasion equilibrium macroscopically as possible at bulk life time initial stage, and whole
The case where body life time middle and later periods does the best reverse abrasion equilibrium, bad SSD disks controllably, in an orderly manner, compartment of terrain occurs.To be smooth
Ground removable disk, smoothly migrating data create the time, and influence business as small as possible, at the same ensure system storage service quality and
Reliability.
In the embodiment of the present invention, step 30 specifically includes：
Step 31, when system is in ageing state, select a SSD as the main force be written SSD；
Wherein, SSD is written as the main force in one SSD of at least one selection as follows：
It randomly selects, in order of numbers are chosen, backward is chosen, selects the maximum SSD of degree of aging by number.
Step 32, setting write-in load main force SSD preferentially undertake writing；Wherein
SSD is written by the main force and undertakes whole writing；Or
The writing that SSD undertakes is written by the main force and is more than the writing that other non-main forces' write-in SSD undertake；Or
SSD is written by the main force and undertakes most of writing, is equably undertaken by other non-main forces' write-in SSD surplus
Remaining writing.
Step 33, when the selected write-in load main force scraps, execute removable disk and operate；
Step 34, after current main force's write-in SSD scraps, reselect a SSD as main force's write-in
SSD, that is, go to step 31.
Wherein, step 10 determination system whether in ageing state include：
The ageing state of each SSD in step 11, acquisition system；
When step 12, the quantity for the SSD for being in ageing state in systems are more than predetermined threshold, determine that system is in aging
State.
Wherein, the ageing state of each SSD includes in acquisition system in step 11：
Obtain the writing of SSD；
The remaining life percentage of the SSD is obtained according to the writing of the SSD；
When the remaining life percentage of the SSD is less than or equal to scheduled single-deck aging threshold β；Determine the SSD
In ageing state.
Wherein it is possible to one initial value is set to single-deck aging threshold β, it can also be during equilibrium to single-deck aging
Threshold value beta executes dynamic and adjusts；The initial value of single-deck aging threshold could be provided as 40%~70%.For example, by single-deck aging threshold
The initial value of value is set as 50%, or the initial value of single-deck aging threshold is set as 60%.The β settings of single-deck aging threshold
Size determines that the line of demarcation of two different abrasion equilibrium strategy executions, single-deck aging threshold β are arranged smaller, executes each
The time of the abrasion equilibrium strategy for undertaking read-write amount of SSD equalitys is longer, too small, two pieces of the appearance of single-deck aging threshold β settings
The risk that SSD is in state of scrapping is also bigger, can be set to single-deck aging threshold β according to the requirement of system performance and stability
Set suitable initial value.
In one embodiment, the writing according to the SSD obtains the remaining life percentage packet of the SSD
It includes：
Wherein, left_life is remaining life percentage；KB_wrtn is the amount of having been written into；Max_wrt is theoretical writing；
Wherein, the calculation formula of theoretical writing Max_wrt is：
Wherein, dynamic is executed to the single-deck aging threshold β to adjust；
The dynamic adjusts：
Detect the time interval y of adjacent 2 low-quality disks；
Wherein, the value range of y is：y∈(γ1,γ2), then adjustment need not be executed to single-deck aging threshold β, if y
Value in section (γ1,γ2) except, then it needs to execute adjustment to single-deck aging threshold β.
Wherein, γ1For low-quality disk interval first threshold, γ2For low-quality disk interval second threshold；The γ1And γ2And removable disk
Operation duration is proportionate；
Wherein, first threshold γ1It is the lower limit of y value zone of reasonableness, second threshold γ2It is the upper of y value zone of reasonableness
Limit, first threshold γ1With second threshold γ2It is pre-set, the first threshold γ of size progress that duration is operated according to removable disk1
With second threshold γ2It is disposed as being more than removable disk operation duration.And leave risk-aversion surplus, it is preferable that first threshold γ1If
2~3 times of removable disk operation duration are set to, second threshold γ2It is set as 3~5 times of removable disk operation duration.
In y≤γ1In the case of, increase single-deck aging threshold β；
In y >=γ2In the case of, reduce single-deck aging threshold β.
If y values are too small, have little time to complete the risk increase that removable disk operation occurs as soon as next piece of low-quality disk, therefore, it is necessary to
Increase single-deck aging threshold β, to extend the time interval y of low-quality disk twice., whereas if y values are excessive, then illustrate the single-deck main force
The time of read-write is longer, without largest optimization system read-write performance, therefore, it is necessary to reduce single-deck aging threshold β, to reduce two
The time interval y of secondary low-quality disk, while the strategy execution time of each SSD equilibriums read-write in system is also increased, it is utmostly excellent
Change system read-write performance.
Single-deck aging threshold β can be adjusted according to the single adjusting step of setting.Single adjusting step could be provided as
4%~10%, for example, single adjusting step is set as 8%, in this way, increasing or reducing single-deck aging threshold β each
When, the amplitude increased or reduced is 8%.
It is illustrated with reference to specific implement scene.
It changes with time song as shown in Fig. 2, being exemplary each SSD service life in the reverse abrasion equilibrium method of the present invention
Line.From fig. 2 it can be seen that the ageing step is stepped into system bulk life time, using the method for reverse abrasion equilibrium so that twice
The interval time of low-quality disk is longer, to there is the sufficient time to complete the replacement of new and old disk.
The application scenarios of embodiment 1 are full sudden strain of a muscle distributed file system；2 application scenarios of embodiment are that the full distributed block that dodges is deposited
Storage system；3 application scenarios of embodiment are full sudden strain of a muscle RAID5 systems；Embodiment 4 is full sudden strain of a muscle distributed objects storage system.
Embodiment 1
In the present embodiment, the hardware module part of storage system includes：
It is made of 5 pieces of SSD on the distributed file system hardware, is new building.It can with the guarantee of two copy version of file-level
By property, metadata and data are distributed in different SSD.Service life aging threshold is 63%.Document size 512KB-16MB.
The processing step of flow elements is as follows：
(1) when the system initialization, balancedly by write IO is assigned to each SSD disks in a manner of adding every time.Due to file
Size is different, write-in size is different every time, therefore it is very balanced that writing, which can not be completely secured,.
(2) service life monitoring modular counts the remaining life situation of each SSD as unit of the moon.
Table 1 monthly counts the remaining life situation of each SSD
Testing result shows, when August, system steps into aging.
(3) it at this point, choosing No. 1 SSD as the write-in load main force according to number order by inverse balance module, undertakes as far as possible
No. 1 SSD is written in more writing, all primary copies, and secondary copy is written in other SSD and as uniform as possible.
(4) during the 9-12 months, the service life of No. 1 SSD is drastically reduced, and is had occurred completely in the first some day in December
It scraps, at the same time, removable disk flow is started by low-quality disk processing module, and Data Migration is completed in some day by the end of December, newly
No. 1 SSD remaining life be 99%.At the same time, No. 2 SSD are chosen as write-in according to number order by inverse balance module
Load the main force.During the 13-16 months, the service life of No. 2 SSD is drastically reduced, and is had occurred completely in some day in the 16th middle of the month
Scrap, at the same time, removable disk flow is started by low-quality disk processing module, and completes Data Migration some day at 16 months bottoms,
The remaining life of No. 2 new SSD is 99%.Subsequently repeat no more.
Table 2 monthly counts the remaining life situation of each SSD
(5) by the 25th month when, No. 5 SSD complete new and old replacement, and so far, which is restored to younger state.
Table 3 monthly counts the remaining life situation of each SSD
(6) subsequently, the newly-increased write-in IO of each SSD are balancedly distributed as possible.Since initiation life gradient, system are follow-up
No longer there is system entirety ageing state, removable disk is executed when SSD occur and scrapping.
Embodiment 2
In the present embodiment, hardware module part includes：
It is made of 8 pieces of SSD on the distributed block storage system hardware, wherein 4 pieces are old disk, 4 pieces are new building.With three copies
Form ensures reliability.Service life aging threshold is 42%.Data slicer size is 1MB.Low-quality disk interval first threshold is 20 days.
The processing step of flow elements is as follows：
(1) when the system initialization, writing is not distributed for old disk, writing is balancedly distributed for new building.
New and old SSD disks have reached abrasion equilibrium at 7th month.
Table 4 monthly counts the remaining life situation of each SSD
(2) over time, 10th month when, system reaches aging.At this point, by inverse balance module according to " minimum
Service life principle " chooses No. 4 SSD as the write-in load main force, undertakes writing as much as possible, all primary copies are written 4
Number SSD, two copies are written in other SSD and as far as possible uniformly.During the 10-14 months, the service life of No. 4 SSD drastically subtracts
It is few, it had occurred in the 14th month low some day and completely scraps, at the same time, removable disk flow is started by low-quality disk processing module, and
Data Migration is completed in some day at the beginning of 15 months, the remaining life of No. 4 new SSD is 99%.At the same time, by inverse equilibrium model
Block chooses No. 1 SSD as the write-in load main force according to " least life principle ".Subsequently repeat no more.
Table 5 monthly counts the remaining life situation of each SSD
(3) with the further passage of time, continuously there is low-quality disk at the 21st month and the 22nd month, and the time is only spaced 17
It, is less than low-quality disk interval first threshold, therefore, at the 22nd month, improves service life aging threshold to 50%, and suspend to number and be
Data are written in 2 and 8 aging single-deck.At the 23rd month, No. 5 SSD completed removable disk, chose No. 8 SSD as the write-in load main force.
At the 26th month, No. 2 SSD completed removable disk, and so far, which is restored to younger state.
Table 6 monthly counts the remaining life situation of each SSD
(4) at the 27th and the 28th month, writing is distributed by the way of abrasion equilibrium, system reached equal at the 29th month
The service life of weighing apparatus.Due to the effect similar with step (1) formation, subsequently repeat no more.
As shown in figure 3, being each SSD service life versus time curve in the present embodiment.
Table 7 monthly counts the remaining life situation of each SSD
Embodiment 3
In the present embodiment, hardware module part includes：
It is made of 13 pieces of SSD on flash memory RAID5 system hardwares, is new building.The concrete configuration of RAID5 is 3+1 forms,
That is 3 parts of data blocks and 1 part of check block.As shown in figure 4, for the structural schematic diagram of the present embodiment, service life aging threshold is set as
50%.Reverse abrasion equilibrium controller controls the data block write-in of its band for managing 13 pieces of SSD.
The processing step of flow elements is as follows：
(1) when the system initialization, under the scene of full band write-in, the theoretical writing of each SSD is uniform, is accounted for
8 each item of table takes data block write-in situation wherein, and the correspondence that data block is numbered in table 8 is：
| ① | Data block 1 |
| ② | Data block 2 |
| ③ | Data block 3 |
| ⊙ | Check block |
The correspondence of 9 data block of table number
(2) over time, 14th month when, system reaches aging.At this point, suitable according to numbering by inverse balance module
Sequence chooses No. 1 SSD as the write-in load main force, undertakes writing as much as possible.
Under the scene of full band write-in, theoretically loads the main force and undertakeWriting, other SSD respectively undertakeWrite
Enter amount.
10 each item of table takes data block write-in situation
(3) with the further passage of time, when by the 17th month, No. 1 SSD scraps, and carries out removable disk operation at this time, by
Other 12 pieces of disks carry out data heavy duty by calculating.At the same time, No. 2 SSD are chosen as the load main force, theoretically undertake 1/4
Writing, other SSD respectively undertake 1/16 writing.
(4) and so on, it repeats no more, until system entirety completes removable disk, returns to younger state.
As shown in figure 4, being the system structure of the present embodiment 3.
Implement use-case 4
In the present embodiment, hardware module part includes：
It is made of 10 pieces of SSD on the full sudden strain of a muscle distributed objects storage system hardware, is new building.Service life aging threshold is
60%.The storage system storage data object in, by access times preceding 10% data object tag be dsc data, will visit
Ask that data object tag of the number in preceding 10%-30% is warm data, other data object tags are cold data.Due to the storage
System allows to lose as content distributing network (CDN) node, data, therefore using single copy.
The processing step of flow elements is as follows：
(1) when the system initialization, log-on data hot statistics execute primary per hour.Dsc data is balancedly divided
It is fitted in each SSD, warm data balancing is assigned in each SSD.If data hot statistics module finds certain SSD dsc data collection
In, then Data Migration is executed, the part dsc data in the SSD is moved in the less SSD of dsc data.
(2) over time, system stepped into aging at 9th month.At this point, in the remaining life such as table of each SSD
It is shown.Service life minimum No. 3 SSD are chosen as the load main force, undertake whole dsc datas.Service life secondary 5, No. 8 low SSD are chosen to make
To load time main force, all warm data are undertaken, and be uniformly distributed in the two SSD.Other SSD undertake cold data, and uniformly
Distribution.
Table 11 monthly counts the remaining life situation of each SSD
(3) with the further passage of time, at the 11st middle of the month, No. 3 SSD scrap, and start removable disk.Reelect
Fixed No. 5 SSD select 8, No. 9 SSD as the load time main force as the load main force.
(4) it so repeats no more, until completing within the 22nd month all removable disks, system is restored to younger state.
As shown in figure 5, being the system mode transition figure of the present invention.Wherein, rectangle indicates that state, circle indicate operation.
When initial, system is in younger state, at this time by the way of abrasion equilibrium.Over time, data content is continuous
Write-in, system steps into ageing state, at this time by the way of reverse abrasion equilibrium.With the further passage of time, between each SSD
Reach disenabling stage every ground, carries out removable disk one by one.As all old disk replacements finish, system is detached from ageing state, returns again
Return younger state.
As shown in fig. 6, being a kind of method flow diagram provided in an embodiment of the present invention, as shown in fig. 6, this method includes：
1st step：When system is in younger state, using conventional wear equalization methods.Younger state includes but not limited to be
Unite initial on-line running when, entirety SSD is when being changed to new building etc..Service life monitoring modular periodically obtains the writing of each SSD, and counts
Calculate remaining life.
2nd step：If the system of detecting steps into the ageing step, single SSD is selected by inverse balance module and is loaded as write-in
The main force.
3rd step：When this in step 2, which is selected SSD, to be scrapped, operated by low-quality disk processing module triggering removable disk, and select
Fixed another SSD is as the write-in load main force.
4th step：Low-quality disk processing module detects the time interval of nearest 2 low-quality disks, as needed to service life aging threshold β into
Row adjustment.
5th step：And so on, until all single-decks have been changed to new building, system is detached from aging, locates again step 2-4
In younger state.
6th step：The 1st step is come back to, until system closing terminates.
Based on same as the previously described embodiments or similar design, the embodiment of the present invention also provides a kind of storage system equalizer line
Device is managed, referring to Fig. 7, a kind of storage system balance management device proposed by the present invention includes：
Inverse balance module, for when system is in ageing state, selecting a solid-state hard disk SSD, and the institute that will be selected
It states SSD and is set as main force write-in SSD；
Low-quality disk processing module, for when selected main force write-in SSD scraps, executing removable disk operation；
The inverse balance module is additionally operable to after current main force write-in SSD scraps, and is reselected a SSD and is done
SSD is written for the main force.
In the embodiment of the present invention, described device further includes：Service life monitoring modular, in one solid state disk of the selection
Before SSD, determine whether system is in ageing state；
The service life monitoring modular determine system whether in ageing state include：
The ageing state of each SSD in acquisition system；
When being in the quantity of the SSD of ageing state in systems more than predetermined threshold, determine that system is in ageing state.
In the embodiment of the present invention, the ageing state of each SSD includes in the service life monitoring modular acquisition system：
Obtain the writing of SSD；
The remaining life percentage of the SSD is obtained according to the writing of the SSD；
When the remaining life percentage of the SSD is less than or equal to scheduled single-deck aging threshold β；Determine the SSD
In ageing state.
In the embodiment of the present invention, the inverse balance module is additionally operable to：
One or more SSD is selected, and sets the selected SSD to time main force and SSD is written；
Wherein, the writing that secondary main force's write-in SSD undertakes is set smaller than main force write-in SSD；
For not being the SSD of the write-in SSD or secondary main forces of main force write-in SSD, it is set as not undertaking writing.
In the embodiment of the present invention, the service life monitoring modular obtains the remaining longevity of the SSD according to the writing of the SSD
Ordering percentage includes：
Left_life=100%- (kB_wrtn)/(Max_wrt)；
Wherein, left_life is remaining life percentage；KB_wrtn is the amount of having been written into；Max_wrt is theoretical writing；
Wherein, the calculation formula of theoretical writing Max_wrt is：
Max_wrt=(the erasable number upper limit of single-deck total capacity * flash cells)/flash cell size.
In the embodiment of the present invention, the service life monitoring modular is additionally operable to execute dynamic adjustment to the single-deck aging threshold β；
The dynamic adjusts：
The dynamic adjusts：
Detect the time interval y of adjacent low-quality disk twice；
Confirm the value of y whether in section (γ1,γ2) except；Wherein, γ1For low-quality disk interval first threshold, γ2It is bad
Disk interval second threshold；The γ1And γ2It is proportionate with removable disk operation duration；
In y≤γ1In the case of, increase single-deck aging threshold β；
In y >=γ2In the case of, reduce single-deck aging threshold β.
In the embodiment of the present invention, the service life monitoring modular each SSD in acquisition system at predetermined intervals
Ageing state.
In the embodiment of the present invention, described device further includes metadata management module, for one SSD of the selection as
The main force is written after SSD,
SSD is written by the main force and undertakes whole writing；Or
The writing that SSD undertakes is written by the main force and is more than the writing that other non-main forces' write-in SSD undertake.
In the embodiment of the present invention, one SSD of at least one selection of the inverse balance module as follows is as the main force
SSD is written：
It randomly selects, in order of numbers are chosen, backward is chosen, selects the maximum SSD of degree of aging by number.
Based on same as the previously described embodiments or similar design, the embodiment of the present invention also provides a kind of storage system equalizer line
Device is managed, including memory, processor and is stored in the computer journey that can be run on the memory and on the processor
Sequence, the processor realize any storage system balanced management side provided in an embodiment of the present invention when executing the computer program
The processing of method.
It should be noted that embodiment described above be merely for convenience of it will be understood by those skilled in the art that, and
It is not used in and limits the scope of the invention, under the premise of not departing from the inventive concept of the present invention, those skilled in the art couple
Any obvious replacement and improvement that the present invention is made etc. are within protection scope of the present invention.
Claims (10)
1. a kind of storage system balanced management method, which is characterized in that the method includes：
When system is in ageing state, a solid-state hard disk SSD is selected, and set the selected SSD to main force's write-in
SSD；
When selected main force write-in SSD scraps, removable disk operation is executed；
After current main force write-in SSD scraps, reselects a SSD and SSD is written as the main force.
2. storage system balanced management method according to claim 1, which is characterized in that hard in one solid-state of the selection
Before disk SSD, the method further includes：
Determine whether system is in ageing state；
The determining system whether in ageing state include：
The ageing state of each SSD in acquisition system；
When being in the quantity of the SSD of ageing state in systems more than predetermined threshold, determine that system is in ageing state.
3. storage system balanced management method according to claim 2, which is characterized in that each in the acquisition system
The ageing state of SSD includes：
Obtain the writing of SSD；
The remaining life percentage of the SSD is obtained according to the writing of the SSD；
When the remaining life percentage of the SSD is less than or equal to scheduled single-deck aging threshold β；Determine that the SSD is in
Ageing state.
4. storage system balanced management method according to claim 1, which is characterized in that the method further includes：
One or more SSD is selected, and sets the selected SSD to time main force and SSD is written；
Wherein, the writing that secondary main force's write-in SSD undertakes is set smaller than main force write-in SSD；
For not being the SSD of the write-in SSD or secondary main forces of main force write-in SSD, it is set as not undertaking writing.
5. storage system balanced management method according to claim 1, which is characterized in that the writing according to the SSD
Enter amount and obtain the remaining life percentage of the SSD and includes：
Left_life=100%- (kB_wrtn)/(Max_wrt)；
Wherein, left_life is remaining life percentage；KB_wrtn is the amount of having been written into；Max_wrt is theoretical writing；
Wherein, the calculation formula of theoretical writing Max_wrt is：
Max_wrt=(the erasable number upper limit of single-deck total capacity * flash cells)/flash cell size.
6. storage system balanced management method according to claim 3, which is characterized in that the single-deck aging threshold β
Dynamic is executed to adjust；
The dynamic adjusts：
Detect the time interval y of adjacent low-quality disk twice；
Confirm the value of y whether except section (γ 1, γ 2)；Wherein, γ1For low-quality disk interval first threshold, γ2Between low-quality disk
Every second threshold；The γ1And γ2It is proportionate with removable disk operation duration；
In y≤γ1In the case of, increase single-deck aging threshold β；
In y >=γ2In the case of, reduce single-deck aging threshold β.
7. storage system balanced management method according to claim 3, which is characterized in that obtain at predetermined intervals
Take the ageing state of each SSD in system.
8. storage system balanced management method according to claim 1, which is characterized in that done in one SSD of the selection
After SSD is written for the main force, further include：
SSD is written by the main force and undertakes whole writing；Or
The writing that SSD undertakes is written by the main force and is more than the writing that other non-main forces' write-in SSD undertake.
9. storage system balanced management method according to claim 1, which is characterized in that at least one as follows
Select a SSD that SSD is written as the main force：
It randomly selects, in order of numbers are chosen, backward is chosen, selects the maximum SSD of degree of aging by number.
10. a kind of storage system balance management device, including memory, processor and it is stored on the memory and can be in institute
State the computer program run on processor, which is characterized in that the processor is realized when executing the computer program as weighed
Profit requires the processing of the method described in any claim in 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711350651.3A CN108334278B (en) | 2017-12-15 | 2017-12-15 | Storage system balance management method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711350651.3A CN108334278B (en) | 2017-12-15 | 2017-12-15 | Storage system balance management method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108334278A true CN108334278A (en) | 2018-07-27 |
| CN108334278B CN108334278B (en) | 2021-05-07 |
Family
ID=62922486
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711350651.3A Active CN108334278B (en) | 2017-12-15 | 2017-12-15 | Storage system balance management method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108334278B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113535082A (en) * | 2021-06-09 | 2021-10-22 | 杭州电子科技大学 | Method for realizing wear inverse equilibrium |
| CN115934007A (en) * | 2023-03-15 | 2023-04-07 | 浪潮电子信息产业股份有限公司 | Data storage method, system, equipment and storage medium of distributed storage system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101814010A (en) * | 2010-01-11 | 2010-08-25 | 浪潮(北京)电子信息产业有限公司 | Disk array implementation method and device |
| CN103309816A (en) * | 2013-05-31 | 2013-09-18 | 苏州亮智科技有限公司 | Solid hard disk, electronic device with solid hard disk and dynamic capacity control method of solid hard disk |
| CN206178765U (en) * | 2016-11-07 | 2017-05-17 | 北京得瑞领新科技有限公司 | SSD unit and SSD equipment |
| CN107193486A (en) * | 2016-03-14 | 2017-09-22 | 三星电子株式会社 | Data storage device and the data handling system including the data storage device |
-
2017
- 2017-12-15 CN CN201711350651.3A patent/CN108334278B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101814010A (en) * | 2010-01-11 | 2010-08-25 | 浪潮(北京)电子信息产业有限公司 | Disk array implementation method and device |
| CN103309816A (en) * | 2013-05-31 | 2013-09-18 | 苏州亮智科技有限公司 | Solid hard disk, electronic device with solid hard disk and dynamic capacity control method of solid hard disk |
| CN107193486A (en) * | 2016-03-14 | 2017-09-22 | 三星电子株式会社 | Data storage device and the data handling system including the data storage device |
| CN206178765U (en) * | 2016-11-07 | 2017-05-17 | 北京得瑞领新科技有限公司 | SSD unit and SSD equipment |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113535082A (en) * | 2021-06-09 | 2021-10-22 | 杭州电子科技大学 | Method for realizing wear inverse equilibrium |
| CN115934007A (en) * | 2023-03-15 | 2023-04-07 | 浪潮电子信息产业股份有限公司 | Data storage method, system, equipment and storage medium of distributed storage system |
| CN115934007B (en) * | 2023-03-15 | 2023-05-23 | 浪潮电子信息产业股份有限公司 | Data storage method, system, equipment and storage medium of distributed storage system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108334278B (en) | 2021-05-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102789423B (en) | Four-pool flash wear leveling method | |
| CN101777026B (en) | Memory management method, hard disk and memory system | |
| US7499960B2 (en) | Adaptive memory allocation | |
| CN100399301C (en) | Mass storage device and method for dynamically managing a mass storage device | |
| CN103399713B (en) | The data buffering method of balance dynamic data attemper performance and solid-state disk service life | |
| CN102640120B (en) | Management system for calculating storage capacity to be increased/decreased | |
| CN104778126B (en) | Transaction Information storage optimization method and system in non-volatile main | |
| CN101645309A (en) | Non-volatile memory device and control method thereof | |
| CN108334278A (en) | A kind of storage system balanced management method and apparatus | |
| CN108255419A (en) | A kind of abrasion equilibrium method and SSD for TLC types SSD | |
| CN105551522A (en) | Management method and management apparatus of flash memory storage device | |
| CN101923448A (en) | Method for reading and writing conversion layer of NAND flash memory | |
| CN101419573A (en) | Storage management method, system and storage apparatus | |
| CN102541470B (en) | Data object processing method and device | |
| CN103095805A (en) | Cloud storage system of data intelligent and decentralized management | |
| US20110035605A1 (en) | Method for optimizing performance and power usage in an archival storage system by utilizing massive array of independent disks (MAID) techniques and controlled replication under scalable hashing (CRUSH) | |
| CN106371916B (en) | A kind of thread optimized method and device thereof of storage system IO | |
| CN106502587A (en) | Data in magnetic disk management method and magnetic disk control unit | |
| CN109960471A (en) | Date storage method, device, equipment and storage medium | |
| CN110096350A (en) | Cold and hot region division energy saving store method based on the prediction of clustered node load condition | |
| CN108563397A (en) | Storage device and data save method | |
| CN103150123A (en) | Volume copy management method on thin provisioning pool of storage subsystem | |
| CN108089994A (en) | Storage device and data save method | |
| CN105117175B (en) | A kind of variable resistance type memory abrasion equilibrium method based on software translating layer | |
| CN107797936A (en) | A kind of method for realizing solid state hard disc abrasion equilibrium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |