CN106569732A - Data migration method and device - Google Patents
Data migration method and device Download PDFInfo
- Publication number
- CN106569732A CN106569732A CN201510655794.XA CN201510655794A CN106569732A CN 106569732 A CN106569732 A CN 106569732A CN 201510655794 A CN201510655794 A CN 201510655794A CN 106569732 A CN106569732 A CN 106569732A
- Authority
- CN
- China
- Prior art keywords
- data
- hot spot
- cache
- address space
- migrated
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Memory System Of A Hierarchy Structure (AREA)
Abstract
The invention provides a data migration method and device, so that a problem that the internal performance of the system is reduced because of large hard disc reading frequency during a hierarchical stored data migration process can be solved. The data migration method comprises: a high-speed cache memory cache module obtains address space of to-be-migrated hotspot data from a hierarchical storage module; the cache module writes in the to-be-migrated hotspot data based on the address space; and the cache module informs the hierarchical storage module of the address space of the to-be-migrated hotspot data in the cache module. Therefore, the performance of the storage system can be improved.
Description
Technical field
The present invention relates to communication field, more particularly to data migration method and device.
Background technology
The storage media types and quantity supported with the progress of seagate, storage system are dramatically increased.
The storage medium that storage system is supported includes:SSD (Solid StateDrive, solid state hard disc), SAS (Serial
Attached SCSI, Serial Attached SCSI (SAS)) hard disk and NL (Near Line) SAS hard disks.Storage resource
(SSD, SAS, NL_SAS) allotment is the costly and time consuming longer task of repeatability, by the rule of early stage
Draw make storage system reach between performance requirement and capacity requirement balance be a relatively difficult thing, i.e.,
The demand that the initial stage reached balance, performance and capacity is made also to change, the resource allocation of storage system is needed
Constantly to adjust, therefore, industry storage system is typically all supported to be classified store function, the purpose of the function
Exactly by the deposit position of the management data for enlivening rank intelligence of statistical data, storage system entirety is improved
Performance simultaneously reduces user cost, wherein, user cost generally includes equipment purchase, space, energy consumption and management
Cost.Automatically Resource allocation and smoothing is carried out by being classified storage, can substantially reduce the difficulty of preconsolidation stress,
The maintenance and dilatation in later stage also becomes easier to.
At present, realization of the industry to classification storage is substantially entered by statistics, analysis, migration three phases
Capable, namely counted by the visiting frequency to certain area data, whether periodically analyze is focus
Data, are then realized from a data spatial migration to another data space by the means of Data Migration.
And the mode for migrating is usually that, on the premise of main frame IO is not affected, storage system is carried out by background program
Data are moved, and such as the data for needing migration are read out from hard disk, and being then then written to needs the purpose of migration
The mode of hard disk address is realized.For specific to storage system, as classification storage is under Cache,
Traditional flow process is as shown in figure 1, the focus migration of classification storage is to be classified memory module to hard disk hot spot region
Read-write operation.Classification memory module counts the hot spot region of hard disk, migrate process cycle to focus
Region carries out data resettlement, and generally moving process is from the disk for needing resettlement to read a hop count first
According to, read data and deposited by interim memory source, then this block number evidence is write needs by classification memory module
On the purpose hard disk address space of migration, interim memory source is discharged after writing, is so once normally removed
Move action to finish, now moving process can be related to a disk-read, a memory source application discharges and once writes
Disk.Simultaneously, it is contemplated that in storage system, the granularity in focus statistics region and the granularity once moved are mismatched,
Such as Fig. 2, typically (typically upper layer logic rolls up address to hard disk address of cache grain to hot spot region statistics granularity
Degree) more than once moving granularity, so, only after the completion of all data resettlements of hot spot region, ability
Modification logical volume address is to hard disk address of cache.So before hot spot region data have not been moved, if
Now just have when writing I O access to the partial data content moved, this time stage storage is processed should
Writing IO needs while write two number evidences, and portion writes original old hard disk areas, it is a need to write it is new
Hard disk areas, the purpose of do so is that, in order to ensure data consistency, this process is as classified memory module
Double to write flow process, summary is exactly the data migration process of classification storage traditionally.It can be seen that, traditional classification is deposited
It is more to the reading times of hard disk in storage data migration process, reduce storage system internal performance.
The content of the invention
The present invention provides a kind of data migration method and device, to solve traditional classification data storage migration
During it is more to hard disk reading times, reduce the problem of internal system performance.
According to the first aspect of the invention, there is provided a kind of data migration method, including:Speed buffering is deposited
Reservoir Cache modules know the address space of hot spot data to be migrated from classification memory module;Cache moulds
Tuber writes hot spot data to be migrated according to address space;Cache modules are by hot spot data to be migrated in Cache
Address space in module notifies to give classification memory module.
Wherein, Cache modules write hot spot data to be migrated according to address space, including:When not in Cache
In the case of the address space of all hot spot datas to be migrated is found in module, Cache modules are deposited to classification
Storage module issues pre- read request;Cache modules receive the hot spot data to be migrated that classification memory module sends,
And the hot spot data for receiving is write into Cache modules.
According to the second aspect of the invention, there is provided a kind of data migration method includes:Classification memory module
The address space of hot spot data to be migrated is informed into cache memory Cache modules;Classification memory module
Knowing that hot spot data, after the address space in Cache modules, write new medium by data to be migrated
Address space;Classification memory module changes logical volume with medium mapping table according to the address space of new medium
Corresponding relation.
Wherein, be classified memory module and data to be migrated known after the address space in Cache modules, incite somebody to action
Hot spot data writes the address space of new medium, including:When reception is from the focus of Cache module forwards
During data access request, judge to ask to need whether the data for reading are according to the migration progress of hot spot data
Migrating data, obtains judged result;Data are read according to judged result, after digital independent is finished, is notified
Cache module datas read and finish.
Further, said method also includes:Cache is stated the address space of hot spot data to be migrated is informed
Before module, the hard disk address space of the level that memory module is located to logical volume LUN is classified according to classification
Storage statistics granularity carries out hot spot data statistics;Classification memory module goes out to be migrated according to the data analysiss of statistics
The space address of hot spot data.
According to the third aspect of the present invention, there is provided a kind of data migration device, including:Know unit,
For hot spot data to be migrated being known by cache memory Cache modules from classification memory module
Address space;First writing unit, for writing to be migrated focus according to address space by Cache modules
Data;Notification unit, for by ground of the Cache modules by hot spot data to be migrated in Cache modules
Location space notifies to give classification memory module.
Wherein, above-mentioned first writing unit, including:Subelement is pre-read, for ought not in Cache modules
In the case of finding the address space of all hot spot datas to be migrated, Cache modules are to classification memory module
Issue pre- read request;Write subelement, for receiving treating for classification memory module transmission by Cache modules
Migration hot spot data, and the hot spot data for receiving is write into Cache modules.
According to the fourth aspect of the present invention, there is provided a kind of data migration device, including:Informing unit,
For the address space of hot spot data to be migrated is informed cache memory by being classified memory module
Cache modules;Second writing unit, for knowing data to be migrated in Cache by being classified memory module
After address space in module, hot spot data is write into the address space of new medium;Modification unit, is used for
Classification memory module changes the corresponding relation of logical volume and medium mapping table according to the address space of new medium.
Wherein, above-mentioned second writing unit, including:Judgment sub-unit, for when reception is from Cache moulds
During the hot spot data access request of block forwarding, judge that request needs what is read according to the migration progress of hot spot data
Whether data are data of having moved, and obtain judged result;Subelement is notified, for reading according to judged result
Data, after digital independent is finished, notify that Cache module datas read and finish.
Further, said apparatus also include:Statistic unit, for by the address of hot spot data to be migrated
Before Cache modules are informed in space, the hard disk ground of the level that memory module is located to logical volume LUN is classified
Location space carries out hot spot data statistics according to classification storage statistics granularity;Analytic unit, for being classified storage mould
Tuber data analysiss according to statistics go out the space address of hot spot data to be migrated.
The present invention has the beneficial effect that:
Scheme provided in an embodiment of the present invention, can reduce by a disk-read operation in data migration process is carried out
With an EMS memory occupation, a writing disk manipulation is pertained only in brush transition process, is improve inside storage system
Performance.
Description of the drawings
Fig. 1 is the flow chart of the hot spot data migration that storage is classified in prior art;
Fig. 2 is that be classified in prior art in the hot spot data transition process of storage double write flow chart;
Fig. 3 is a kind of flow chart of data migration method that the embodiment of the present invention 1 is provided;
Fig. 4 is the ultimate principle figure of Cache in the embodiment of the present invention 1;
Fig. 5 is the flow chart of the data migration method that the embodiment of the present invention 2 is provided;
Fig. 6 is to hit the Data Migration in the case of all migrating datas in the embodiment of the present invention 3 in Cache
Flow chart;
Fig. 7 is the Data Migration in the embodiment of the present invention 3 in the case of the miss all migrating datas of Cache
Flow chart;
Fig. 8 be I O access migration region is write in the embodiment of the present invention 3 in data migration process in the case of
Data Migration flow chart;
Fig. 9 be have reading I O access migration region in the embodiment of the present invention 3 in data migration process in the case of
Data Migration flow chart;
Figure 10 is the structured flowchart of the data migration device that the embodiment of the present invention 4 is provided;
Figure 11 is the structured flowchart of the data migration device that the embodiment of the present invention 5 is provided.
Specific embodiment
It is more to hard disk reading times in traditional classification data storage transition process in order to solve, reduce and be
The problem of system internal performance, the invention provides a kind of data migration method and device, below in conjunction with accompanying drawing with
And embodiment, the present invention will be described in further detail.It should be appreciated that described herein be embodied as
Example limits the present invention only to explain the present invention, not.
Embodiment 1
Present embodiments provide a kind of data migration method, the method is applied to Cache modules, i.e., by
Realizing, Fig. 3 is a kind of flow chart of data migration method that the embodiment of the present invention 1 is provided to Cahche modules,
As shown in figure 3, the method comprises the steps:
Step 301:Cache modules know the address space of hot spot data to be migrated from classification memory module;
Step 302:Cache modules write hot spot data to be migrated according to address space;
Step 303:Address space of the Cache modules by hot spot data to be migrated in Cache modules is notified
Give classification memory module.
Wherein, the step of Cache modules write hot spot data to be migrated according to address space can specifically include:
In the case of the address space of all hot spot datas to be migrated is not found in Cache modules,
Cache modules issue pre- read request to classification memory module;Cache modules receive classification memory module and send
Hot spot data to be migrated, and by the hot spot data for receiving write Cache modules.
As shown in figure 4, the data space LUN that main frame is accessed first is expressed as in storage system Cache
One tree, and the function of Cache is exactly to carry out certain management to this tree, the access of main frame IO is in fact
It is exactly the access of this tree, such as this tree is hit, is pre-read.Simultaneously Cache also to this set into
Superseded and brush dirty data etc. of going is operated.So as to realize feedbacks of the Cache to the access of main frame IO.Furthermore it is main
Machine IO can be resident a period of time, so as to reach in Cache once reaching the storage system Cache stage
Effect is hit to Cache, this is also the meaning that Cache is present.
In the present embodiment, the migration action of former classification memory module is given cache modules to implement, is decomposed
To cache modules, it is that a brush is operated, the brush operation triggering classification storage transferring module of cache modules
Singly write, reach the effect of Data Migration.Based on this, cache modules can utilize what is itself existed
Cache resources, if hit, it is possible to reduce a disk-read operation and once interim EMS memory occupation, directly by Cache
Classification memory module is brushed under data in module, destination address is write data into by being classified memory module,
Now simply destination address is write data into without writing former media address.Therefore, reduce and once write behaviour
Make.If be not hit by, Cache modules are needed to classification memory module pre-reads data, then by brush
Mode write data into destination address, be now, with traditional flow process as, to need disk-read to operate,
But the use to interim internal memory still has advantage, meanwhile, the page read in advance increased the life of Cache
Middle rate.In the present embodiment, the cooperation being classified between memory module and cache modules completes Data Migration
Whole flow process, substitutes the migration operation of classification storage using the brush operation of cache, therefrom reduces migrating
Disk-read of journey and EMS memory occupation.
Embodiment 2
Present embodiments provide a kind of data migration method, the method is applied to be classified memory module side, by point
Realizing, Fig. 5 is the flow chart of the data migration method that the embodiment of the present invention 2 is provided to level memory module, such as
Shown in Fig. 5, the method comprises the steps:
Step 501:The address space of hot spot data to be migrated is informed Cache modules by classification memory module;
Step 502:Classification memory module knowing data to be migrated after the address space in Cache modules,
Hot spot data is write into the address space of new medium;
Step 503:Classification memory module is changed logical volume according to the address space of new medium and is mapped with medium
The corresponding relation of table.
Wherein, be classified memory module and data to be migrated known after the address space in Cache modules, incite somebody to action
Hot spot data writes the address space of new medium, can specifically include:
When receiving from the hot spot data access request of Cache module forwards, according to the migration of hot spot data
Progress judges to ask whether the data for needing request to access are data of having moved, and obtains judged result;According to sentencing
Disconnected result reads data, and pending data is read after finishing, and notifies that Cache module datas read and finishes.
Further, said method can also include:
Before the address space of hot spot data to be migrated is informed Cache modules, classification memory module is to patrolling
Volume volume LUN be located level hard disk address space according to classification storage statistics granularity carry out hot spot data system
Meter;Classification memory module goes out the space address of hot spot data to be migrated according to the data analysiss of statistics.
In the present embodiment, in classification storage, the statistics of focus is the access frequency to certain sector address space on disk,
And carry out mainly Cache brushes dirty data for disk and Cache tracings from the point of view of Cache within the storage system
For the access frequency of disk.Now, when such as on fruit tray, certain sector address space is considered as hot spot data by statistics,
Be then brush is carried out to this sector address space due to Cache caused by, or by the main frame frequent tracing region
Caused by address space.And now be also possible to reside these hot spot datas in Cache.Once these focuses
Data are not eliminated in Cache, then just can be by being classified memory function module by this part focus
Data write the hard disk areas of new capability.But if be now considered hot spot data do not had in Cache,
The data that Cache not hits this part are then needed to be read out by being classified memory module in advance, then by dividing
Level memory function module writes new performance hard disk region, meanwhile, if now storage system is in tracing
Focus caused by state, then the data that this pre-reads up just meet Cache and hit so as to improve system
Performance.
Classification memory module then receives the hot spot data brushed under Cache, needs basis to migrate into scale and shows certainly
It is fixed that Cache data are write into new medium address area.Because focus statistics is needed commonly greater than migration granularity,
Therefore, before hot spot region data are not migrated and completed, classification storage is can not to change logical volume address to disk
The mapping table of location, only just changes the address mapping table that logic is rolled onto disk after the completion of hot spot data is all migrated.
When so next I O access is to the logical volume region, access be then new medium region.For this heat
Point data is not migrated before completing, if read request has access to hot spot region, if the then read request data
Just in Cache, then directly access Cache and return, if the request data is no in Cache,
Or part is in Cache, then Cache is needed to issue pre- read request, classification storage mould to classification memory module
Block judges it is to read old media data or reading new medium data by migrating message bit pattern, so as to reach number
According to the purpose for accessing.
Embodiment 3
The present embodiment respectively by migrating data Cache hit when migration, migrating data Cache is miss
When migration, I O access migration region and transition process are write in transition process in have reading I O access migration
These four situations of region to get off enter the flow process of cooperative achievement Data Migration between Cache and classification memory module
Row description:
During migrating data, the Data Migration flow process in the case of all migrating datas is hit in Cache:
As shown in fig. 6, specifically including following process:
The level hard disk address space that classification memory function module is located to LUN is according to classification storage statistics grain
Degree carries out hot spot data statistics;
Which address area is classification memory function module go out for hot spot region according to grain size statistics data analysiss;
Classification memory function module notifies which address of Cache modules needs to carry out focus migration;
Cache functional modules carry out data writing disk manipulation according to the address space that classification storage is provided, and now write
Just hit in Cache in disk address region;
The address space of brush classification memory module request under Cache;
Classification storage writes the address space that disk request writes data into new medium according to what Cache was issued;
Classification memory module modification logic is rolled onto medium mapping table for new medium address space.
Data Migration flow process during migrating data, in the case of the miss all migrating datas of Cache:
As shown in fig. 7, specifically including following process:
The level hard disk address space that classification memory function module is located to LUN is according to classification storage statistics grain
Degree carries out hot spot data statistics;
Which address area is classification memory function module go out for hot spot region according to grain size statistics data analysiss;
Classification memory function module notifies which address of Cache needs to carry out focus migration;
Cache functional modules carry out data writing disk manipulation according to the address space that classification storage is provided, and now write
Disk address region is not in Cache or part is in Cache;
Cache issues pre- read request to classification memory module, and classification memory module has pre-read data and returned Cache;
Cache continues the address space for issuing classification memory requirement to classification memory module;
Classification storage writes the address space that disk request writes data into new medium according to what Cache was issued;
Classification memory module modification logic is rolled onto medium mapping table for new medium address space.
During migrating data, there are the data write in the case of I O access migrates region in data migration process
Migration flow process:
As shown in figure 8, specifically including following process:
The level hard disk address space that classification memory function module is located to LUN is according to classification storage statistics grain
Degree carries out hot spot data statistics;
Which address area is classification memory function module go out for hot spot region according to grain size statistics data analysiss;
Classification memory function module notifies which address of Cache needs to carry out focus migration;
Cache functional modules carry out data writing disk manipulation according to the address space that classification storage is provided, and now write
Disk address area is just in Cache;
Cache continues the address space for issuing classification memory requirement to classification memory module;
Classification storage writes the address space that disk request writes data into new medium according to what Cache was issued;
Classification memory module modification logic is rolled onto medium mapping table for new medium address space;
If now just there is write request to access hot spot region data, now request accesses the Cache pages, just
In good lot, then main frame is directly returned.
During migrating data, there are the data read in the case of I O access migrates region in data migration process
Migration flow process:
As shown in figure 9, specifically including following process:
The level hard disk address space that classification memory function module is located to LUN is according to classification storage statistics grain
Degree carries out hot spot data statistics;
Which address area is classification memory function module go out for hot spot region according to grain size statistics data analysiss;
Classification memory function module notifies which address of Cache needs to carry out focus migration;
Cache functional modules carry out data writing disk manipulation according to the address space that classification storage is provided, and now write
Disk address area is just in Cache;
Cache continues the address space for issuing classification memory requirement to classification memory module;
Classification storage writes the address space that disk request writes data into new medium according to what Cache was issued.But
Partial data has been write only, whole hot spot region data do not write;
Now just there is read request access hot spot data, and be the area data for having migrated;
If cache hits, directly Cache is returned, and if being not hit by, Cache issues request to classification
Memory module;
Classification memory module is to have migrated according to migrate into that scale designation judges that this read request needs reading
New data or legacy data, if migrating data has been write be new data, be otherwise legacy data;
Classification storage reads data to Cache, and Cache returns data to host read request;
Classification memory module modification logic is rolled onto medium mapping table for new medium address space.
Embodiment 4
A kind of data migration device is present embodiments provided, the device is arranged at Cache sides, Tu10Shi
The structured flowchart of the data migration device that the embodiment of the present invention 4 is provided, as shown in Figure 10, the device 100
Including following ingredient:
Know unit 101, for knowing focus number to be migrated by Cache modules from classification memory module
According to address space;
First writing unit 102, for writing to be migrated focus number according to address space by Cache modules
According to;
Notification unit 103, for by Cache modules by hot spot data to be migrated in Cache modules
Address space notifies to give classification memory module.
Wherein, above-mentioned first writing unit 102 can specifically include:Subelement is pre-read, for not existing
In the case of the address space of all hot spot datas to be migrated is found in Cache modules, Cache modules to
Classification memory module issues pre- read request;Write subelement, for receiving classification storage by Cache modules
The hot spot data to be migrated that module sends, and the hot spot data for receiving is write into Cache modules.
Embodiment 5
A kind of data migration device is present embodiments provided, the device is arranged at classification memory module side, figure
11 is the structured flowchart of the data migration device that the embodiment of the present invention 5 is provided, as shown in figure 11, the device
110 specifically include following ingredient:
Informing unit 111, for being informed the address space of hot spot data to be migrated by being classified memory module
Cache modules;
Second writing unit 112, for knowing data to be migrated in Cache moulds by being classified memory module
After address space in block, hot spot data is write into the address space of new medium;
Modification unit 113, for be classified memory module according to the address space of new medium change logical volume with
The corresponding relation of medium mapping table.
Wherein, above-mentioned second writing unit 112 can specifically include:Judgment sub-unit, for next when receiving
From the hot spot data access request of Cache module forwards when, according to the migration progress of hot spot data judge ask
Need whether the data for reading are data of having moved, obtain judged result;Subelement is notified, is sentenced for basis
Disconnected result reads data, continues after fetching data and finishing, notifies that Cache module datas read and finish.
Further, said apparatus 110 can also include:Statistic unit, for by focus number to be migrated
According to address space inform Cache modules before, be classified the level that is located to logical volume LUN of memory module
Hard disk address space according to classification storage statistics granularity carry out hot spot data statistics;Analytic unit, for dividing
Level memory module goes out the space address of hot spot data to be migrated according to the data analysiss of statistics.
Scheme provided in an embodiment of the present invention, reduces a disk-read operation and one in data migration process is carried out
Secondary EMS memory occupation, pertains only to a writing disk manipulation, and legacy migration method is in brush in brush transition process
Waiting needs carries out double writing so as to ensure data consistency.Meanwhile, relative to legacy migration method, the present invention is real
The scheme of example is applied in Cache under the scene of miss migrating data, the data read in advance can improve system
Hit rate, so as to a certain degree lift system performance.
Although being example purpose, the preferred embodiments of the present invention are had been disclosed for, those skilled in the art will
Recognize various improvement, increase and replacement be also it is possible, therefore, the scope of the present invention should be not limited to
State embodiment.
Claims (10)
1. a kind of data migration method, it is characterised in that include:
Cache memory Cache modules know the address of hot spot data to be migrated from classification memory module
Space;
The Cache modules write the hot spot data to be migrated according to the address space;
Address space of the Cache modules by hot spot data to be migrated in the Cache modules notify to
The classification memory module.
2. method according to claim 1, it is characterised in that the Cache modules are according to described
Address space writes the hot spot data to be migrated, including:
In the case of the address space of all hot spot datas to be migrated is not found in the Cache modules,
The Cache modules issue pre- read request to the classification memory module;
The Cache modules receive the hot spot data to be migrated that the classification memory module sends, and will receive
The hot spot data for arriving writes the Cache modules.
3. a kind of data migration method, it is characterised in that include:
The address space of hot spot data to be migrated is informed cache memory Cache moulds by classification memory module
Block;
The classification memory module knowing data to be migrated after the address space in the Cache modules,
The hot spot data is write into the address space of new medium;
The classification memory module is changed logical volume according to the address space of the new medium and is mapped with medium
The corresponding relation of table.
4. method according to claim 3, it is characterised in that the classification memory module exists
Know that data to be migrated, after the address space in the Cache modules, the hot spot data are write new
The address space of medium, including:
When receiving from the hot spot data access request of the Cache module forwards, according to hot spot data
Migration progress judges that the request needs whether the data for reading are data of having moved, and obtains judged result;
Data are read according to judged result, after digital independent is finished, notifies that the Cache module datas are read
Take complete.
5. the method according to claim 3 or 4, it is characterised in that methods described also includes:
The address space of hot spot data to be migrated is being informed before stating Cache modules, the classification storage mould
The hard disk address space of the level that block is located to logical volume LUN carries out focus according to classification storage statistics granularity
Data statisticss;
The classification memory module goes out the space address of hot spot data to be migrated according to the data analysiss of statistics.
6. a kind of data migration device, it is characterised in that include:
Know unit, treat for being known from classification memory module by cache memory Cache modules
The address space of migration hot spot data;
First writing unit, for waiting to move according to address space write by the Cache modules
Move hot spot data;
Notification unit, for by the Cache modules by hot spot data to be migrated in the Cache modules
Interior address space is notified to the classification memory module.
7. device according to claim 6, it is characterised in that first writing unit, including:
Subelement is pre-read, for all hot spot datas to be migrated ought not be found in the Cache modules
In the case of address space, the Cache modules issue pre- read request to the classification memory module;
Write subelement, for by the Cache modules receive it is described classification memory module transmission wait move
Hot spot data is moved, and the hot spot data for receiving is write into the Cache modules.
8. a kind of data migration device, it is characterised in that include:
Informing unit, for being informed the address space of hot spot data to be migrated at a high speed by being classified memory module
Buffer storage Cache module;
Second writing unit, for knowing data to be migrated described by the classification memory module
After address space in Cache modules, the hot spot data is write into the address space of new medium;
Modification unit, changes logic according to the address space of the new medium for the classification memory module
The corresponding relation of volume and medium mapping table.
9. device according to claim 8, it is characterised in that second writing unit, including:
Judgment sub-unit, for when receive from the hot spot data access request of the Cache module forwards when,
Judge that the request needs whether the data for reading are data of having moved according to the migration progress of hot spot data, obtain
To judged result;
Subelement is notified, for data being read according to judged result, pending data is read after finishing, and notifies described
Cache module datas read and finish.
10. device according to claim 8 or claim 9, it is characterised in that described device also includes:
Statistic unit, for before the address space of hot spot data to be migrated to be informed the Cache modules,
The hard disk address space of the level that the classification memory module is located to logical volume LUN is according to classification storage system
Meter granularity carries out hot spot data statistics;
Analytic unit, goes out to be migrated hot spot data according to the data analysiss of statistics for the classification memory module
Space address.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510655794.XA CN106569732B (en) | 2015-10-12 | 2015-10-12 | Data migration method and device |
PCT/CN2016/099934 WO2017063495A1 (en) | 2015-10-12 | 2016-09-23 | Data migration method and apparatus |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510655794.XA CN106569732B (en) | 2015-10-12 | 2015-10-12 | Data migration method and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106569732A true CN106569732A (en) | 2017-04-19 |
CN106569732B CN106569732B (en) | 2021-04-20 |
Family
ID=58506322
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510655794.XA Active CN106569732B (en) | 2015-10-12 | 2015-10-12 | Data migration method and device |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN106569732B (en) |
WO (1) | WO2017063495A1 (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108519860A (en) * | 2018-03-31 | 2018-09-11 | 北京联想核芯科技有限公司 | A kind of SSD reads the treating method and apparatus of hit |
CN109376094A (en) * | 2018-09-27 | 2019-02-22 | 郑州云海信息技术有限公司 | A kind of method of the cold and hot Data Migration of garbage reclamation in storage system |
CN110162376A (en) * | 2018-02-12 | 2019-08-23 | 杭州宏杉科技股份有限公司 | A kind of data read-write method and device |
CN111273861A (en) * | 2020-01-15 | 2020-06-12 | 苏州浪潮智能科技有限公司 | Data volume Bitmap destaging method and system |
CN112214169A (en) * | 2020-09-28 | 2021-01-12 | 深圳大学 | Data storage device and stored data migration method |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111124307B (en) * | 2019-12-20 | 2022-06-07 | 北京浪潮数据技术有限公司 | Data downloading and brushing method, device, equipment and readable storage medium |
CN111722797B (en) * | 2020-05-18 | 2021-06-29 | 西安交通大学 | SSD and HA-SMR hybrid storage system oriented data management method, storage medium and device |
CN112667525A (en) * | 2020-12-23 | 2021-04-16 | 北京浪潮数据技术有限公司 | Used space measuring method and component of persistent memory |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102799390A (en) * | 2012-06-14 | 2012-11-28 | 国家超级计算深圳中心(深圳云计算中心) | Cooperative adaptive pre-fetching method for sensing energy consumption and middleware |
CN102831087A (en) * | 2012-07-27 | 2012-12-19 | 国家超级计算深圳中心(深圳云计算中心) | Data reading-writing processing method and device based on mixing memory |
CN103186350A (en) * | 2011-12-31 | 2013-07-03 | 北京快网科技有限公司 | Hybrid storage system and hot spot data block migration method |
US8688904B1 (en) * | 2005-05-23 | 2014-04-01 | Hewlett-Packard Development Company, L.P. | Method for managing data storage |
CN104321754A (en) * | 2012-12-21 | 2015-01-28 | 华为技术有限公司 | Method and device for setting working mode of Cache |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7756943B1 (en) * | 2006-01-26 | 2010-07-13 | Symantec Operating Corporation | Efficient data transfer between computers in a virtual NUMA system using RDMA |
CN102215268A (en) * | 2011-07-14 | 2011-10-12 | 北京飞杰信息技术有限公司 | Method and device for transferring file data |
CN102968498B (en) * | 2012-12-05 | 2016-08-10 | 华为技术有限公司 | Data processing method and device |
CN103106045A (en) * | 2012-12-20 | 2013-05-15 | 华为技术有限公司 | Data migration method, system and device at host machine end |
CN104917784B (en) * | 2014-03-10 | 2018-06-05 | 华为技术有限公司 | A kind of data migration method, device and computer system |
-
2015
- 2015-10-12 CN CN201510655794.XA patent/CN106569732B/en active Active
-
2016
- 2016-09-23 WO PCT/CN2016/099934 patent/WO2017063495A1/en active Application Filing
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8688904B1 (en) * | 2005-05-23 | 2014-04-01 | Hewlett-Packard Development Company, L.P. | Method for managing data storage |
CN103186350A (en) * | 2011-12-31 | 2013-07-03 | 北京快网科技有限公司 | Hybrid storage system and hot spot data block migration method |
CN102799390A (en) * | 2012-06-14 | 2012-11-28 | 国家超级计算深圳中心(深圳云计算中心) | Cooperative adaptive pre-fetching method for sensing energy consumption and middleware |
CN102831087A (en) * | 2012-07-27 | 2012-12-19 | 国家超级计算深圳中心(深圳云计算中心) | Data reading-writing processing method and device based on mixing memory |
CN104321754A (en) * | 2012-12-21 | 2015-01-28 | 华为技术有限公司 | Method and device for setting working mode of Cache |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110162376A (en) * | 2018-02-12 | 2019-08-23 | 杭州宏杉科技股份有限公司 | A kind of data read-write method and device |
CN108519860A (en) * | 2018-03-31 | 2018-09-11 | 北京联想核芯科技有限公司 | A kind of SSD reads the treating method and apparatus of hit |
CN108519860B (en) * | 2018-03-31 | 2021-01-19 | 深圳忆联信息***有限公司 | SSD read hit processing method and device |
CN109376094A (en) * | 2018-09-27 | 2019-02-22 | 郑州云海信息技术有限公司 | A kind of method of the cold and hot Data Migration of garbage reclamation in storage system |
CN111273861A (en) * | 2020-01-15 | 2020-06-12 | 苏州浪潮智能科技有限公司 | Data volume Bitmap destaging method and system |
CN112214169A (en) * | 2020-09-28 | 2021-01-12 | 深圳大学 | Data storage device and stored data migration method |
Also Published As
Publication number | Publication date |
---|---|
WO2017063495A1 (en) | 2017-04-20 |
CN106569732B (en) | 2021-04-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106569732A (en) | Data migration method and device | |
CN105930282B (en) | A kind of data cache method for NAND FLASH | |
CN110515859B (en) | Parallel processing method for read-write requests of solid state disk | |
CN103136121B (en) | Cache management method for solid-state disc | |
US10049040B2 (en) | Just in time garbage collection | |
CN104834607B (en) | A kind of hit rate for improving distributed caching and the method for reducing solid state hard disc abrasion | |
CN105159622B (en) | A kind of method and system reducing SSD read-write IO time delay | |
CN103777905B (en) | Software-defined fusion storage method for solid-state disc | |
TWI446345B (en) | Method for performing block management, and associated memory device and controller thereof | |
CN104794064B (en) | A kind of buffer memory management method based on region temperature | |
CN107885456A (en) | Reduce the conflict that I/O command accesses NVM | |
CN109710184B (en) | Hierarchical hybrid storage method and system for tile record disk perception | |
CN106294197A (en) | A kind of page frame replacement method towards nand flash memory | |
CN109240939B (en) | Method for rapidly processing solid state disk TRIM | |
CN103902475B (en) | Solid state disk concurrent access method and device based on queue management mechanism | |
WO2017132056A1 (en) | Hot-spot adaptive garbage collection | |
CN110515552A (en) | A kind of method and system of storage device data no write de-lay | |
CN107797772A (en) | A kind of garbage retrieving system and method based on flash media | |
CN103677670A (en) | Method and device for reading data | |
CN109164976A (en) | Optimize storage device performance using write buffer | |
CN109144885A (en) | The rubbish recovering method and solid storage device of solid storage device | |
CN106527987A (en) | Non-DRAM SSD master control reliability improving system and method | |
CN107145301B (en) | Data access method and device based on hybrid disk | |
CN109558334A (en) | Junk data recovery method and solid storage device | |
CN107391035B (en) | A method of being perceived by misprogrammed reduces solid-state mill damage |
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 |