CN102024044B - Distributed file system - Google Patents
Distributed file system Download PDFInfo
- Publication number
- CN102024044B CN102024044B CN2010105872357A CN201010587235A CN102024044B CN 102024044 B CN102024044 B CN 102024044B CN 2010105872357 A CN2010105872357 A CN 2010105872357A CN 201010587235 A CN201010587235 A CN 201010587235A CN 102024044 B CN102024044 B CN 102024044B
- Authority
- CN
- China
- Prior art keywords
- metadata
- data
- storage system
- module
- backup
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/10—File systems; File servers
- G06F16/18—File system types
- G06F16/182—Distributed file systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
本发明实施例提供一种分布式文件系统。本包括分布式文件系统接入模块、元数据管理单元，主用存储系统、备用存储系统和外部存储系统；其中，所述接入模块、所述元数据管理单元、所述主用存储系统和所述备用存储系统之间通过系统总线连接；所述外部存储系统通过网络与所述主用存储系统和所述备用存储系统连接。本发明实施例提供的分布式文件系统中，通过将数据写入到主用存储系统中，采用异步备份机制备份到备用存储系统，不影响高速设备的读写性能；而且可以实现在线恢复数据，数据对外服务过程中自动实现数据恢复，加快恢复过程。
An embodiment of the present invention provides a distributed file system. This includes a distributed file system access module, a metadata management unit, a main storage system, a backup storage system and an external storage system; wherein, the access module, the metadata management unit, the main storage system and The backup storage systems are connected through a system bus; the external storage system is connected with the main storage system and the backup storage system through a network. In the distributed file system provided by the embodiment of the present invention, by writing data into the main storage system, the asynchronous backup mechanism is used to back up the backup storage system, which does not affect the read and write performance of high-speed devices; and it can realize online recovery of data, Data recovery is automatically realized during the process of data external service, which speeds up the recovery process.
Description
Technical field
The embodiment of the invention relates to the data backup technology, relates in particular to a kind of distributed file system.
Background technology
Along with the internet develops to wider direction, all trades and professions applied more and more, particularly Streaming Media are used, content delivery network (Content Delivery Network; Hereinafter to be referred as: Content supply server performance and reliability CDN) are more and more important, and the core document system of carrying these application datas requires also increasingly high to Performance And Reliability.
For reliability, prior art mainly adopts raid-array (Redundant Arrayof Independent Disk; Hereinafter to be referred as: RAID) technology uses the redundancy backup characteristic to guarantee.Briefly, RAID be a kind of polylith independently hard disk (physical hard disk) combine by different modes and form a hard disk groups (logic hard disk), thereby provide than the higher memory property of single hard disk with the data backup technology is provided.And for performance, prior art adopts the storage striping of distributed file system to manage the polymerization hard disk bandwidth that superposes usually.Described band turns to a kind of management function, and its effect is that data are dispersed on a plurality of memory devices by a fixed step size, and making walks abreast when reading obtains from a plurality of physical storage devices, realizes the stack of many physical storage devices performance.The data redundancy and the data striping characteristic of RAID technology have guaranteed the high reliability and the high-performance of distributed file system.
In general, distributed file system is appreciated that and becomes to be based upon on the network store system.Under constantly increase form of high performance demands, hard disk is gradually by the substituted trend development of solid state hard disc, because solid-state disk is very expensive, adopts this moment the RAID technology of solid state hard disc backup solid state hard disc to cause server cost expense to increase severely.Normal RAID1 of employing of RAID technology and RAID5 realize.Wherein, RAID1 is the 1st a grade of RAID technology, adopts a kind of complete mirror back-up, needs the storage system of two homogeneities to carry out read-write operation synchronously, mirror image each other, even a disk failures is arranged, system still can operate as normal.RAID5 is the storage solution that a kind of memory property, data security and carrying cost are taken into account; The data to storage do not back up; But store data and corresponding parity information on each disk of forming RAID5 into, and parity information is stored in respectively on the different disks with corresponding data.After the data in magnetic disk of RAID5 is damaged, utilize remaining data and corresponding parity information to go to recover impaired data.
In realizing process of the present invention; The inventor finds to have following problem in the prior art at least: in the CDN network; Use the RAID technology to have following shortcoming: equal capacity, forming RAID needs more disk, particularly RAID1 to need double disk; And solid-state disk is very expensive, thereby carrying cost causes system cost too high; The modification of any data of RAID5 need rewrite verification, and it is slow slightly to cause writing data, and data are of a specified duration release time, possibly influence business; And, the be damaged restriction of quantity of disk of the recovery of data backup.
Summary of the invention
The embodiment of the invention provides a kind of distributed file system; Comprise: AM access module, the metadata management unit that is connected with said AM access module; And be connected with said metadata management unit respectively main with storage system and accumulator backup system, said AM access module also is connected with said accumulator backup system with storage system with said leading respectively; Said AM access module, said metadata management unit, said main with connecting through system bus between storage system and the said accumulator backup system; Also comprise external storage system, said external storage system is connected with said accumulator backup system with storage system with said master through network; Wherein:
Said AM access module; Be used to receive the read/write data request; Send the data metadata corresponding that metadata request is asked to obtain to said metadata management unit, and use said metadata and read and write data with storage system or said accumulator backup system to said master;
Said metadata management unit is used for when the said metadata of said AM access module request, search the data of being asked said master with the position on storage system or the said accumulator backup system, and construct said metadata, return to said AM access module; Said master uses storage system, and being used for when said distributed file system is in normal condition is that said AM access module provides the data of being asked; Said accumulator backup system is used for when said distributed file system is in ERST or returns to form, for said master provides data backup with storage system; Said external storage system is used to said master and with storage system data backup is provided.
In the distributed file system that the embodiment of the invention provides, main through writing data into in the storage system, adopt async backup mechanism to backup to accumulator backup system, do not influence the readwrite performance of high-speed equipment; And can be implemented in the line restore data, and realize the data recovery in the external service process of data automatically, accelerate rejuvenation.
Description of drawings
In order to be illustrated more clearly in the embodiment of the invention or technical scheme of the prior art; To do one to the accompanying drawing of required use in embodiment or the description of the Prior Art below introduces simply; Obviously, the accompanying drawing in describing below is some embodiments of the present invention, for those of ordinary skills; Under the prerequisite of not paying creative work, can also obtain other accompanying drawing according to these accompanying drawings.
Fig. 1 forms synoptic diagram for one embodiment of the invention distributed file system;
Fig. 2 forms synoptic diagram for another embodiment of the present invention distributed file system;
Fig. 3 is embodiment of the invention distributed file system treatment scheme synoptic diagram under normal condition;
Fig. 4 is that embodiment of the invention distributed file system is under ERST and be the treatment scheme synoptic diagram under the read data request situation;
Fig. 5 is that embodiment of the invention distributed file system is under ERST and be the treatment scheme synoptic diagram under the write data requests situation;
Fig. 6 is returning to form response readings according to the processing of request schematic flow sheet for embodiment of the invention distributed file system;
Fig. 7 embodiment of the invention distributed file system is at the treatment scheme synoptic diagram of response write data requests down that returns to form;
Fig. 8 is the treatment scheme synoptic diagram of embodiment of the invention distributed file system data recovery procedure under the system recovery state.
Embodiment
For the purpose, technical scheme and the advantage that make the embodiment of the invention clearer; To combine the accompanying drawing in the embodiment of the invention below; Technical scheme in the embodiment of the invention is carried out clear, intactly description; Obviously, described embodiment is the present invention's part embodiment, rather than whole embodiment.Based on the embodiment among the present invention, those of ordinary skills are not making the every other embodiment that is obtained under the creative work prerequisite, all belong to the scope of the present invention's protection.
Fig. 1 forms synoptic diagram for one embodiment of the invention distributed file system; As shown in Figure 1; This distributed file system 1 comprises AM access module 11, the metadata management unit 12 that is connected with AM access module 11; And be connected with metadata management unit 12 respectively main with storage system 13 and accumulator backup system 14, AM access module 11 also is connected with accumulator backup system 14 with storage system 13 with leading respectively; Wherein, AM access module 11, metadata management unit 12, master all are arranged in internal network with storage system 13 and accumulator backup system 14, connect through system bus between each functional module; This distributed file system 1 also comprises the external storage system 15 that is arranged in external network, external storage system 15 through main with storage system 13 and accumulator backup system 14 networks be connected.
Metadata management unit 12 is used for when the said metadata of AM access module 11 requests, searching the data of being asked and leading with the position on storage system 13 or the accumulator backup system 14, and construct said metadata, returns to AM access module 11; Also be used for the system state with equipment state incident conversion this distributed file system 1 of storage system 13 according to the master.Distributed file system 1 has three kinds of states: normal condition, ERST and return to form: wherein:
Normal condition refers to that the memory device of leading with storage system 13 does not break down, and data storage can backup on the accumulator backup system 14 on leading with storage system whenever necessary.
ERST refers to that the master has memory device to break down with storage system 13, and this moment need be main with storage system 13, accumulator backup system 14 co-operation, coordinates to accomplish data storage, and preservation coordination result, coordinates the result and is used for data recovery procedure.
Return to form, refer to main recovered failed storage equipment with storage system 13 after, triggering system is carried out the system data recovery, mainly uses the Backup Data on the accumulator backup system 14 to return to the data that are stored in originally on the failed storage equipment.
Further, metadata management unit 12 is responsible for when AM access module 11 request metadata, search data on storage system the position and construct metadata, return to AM access module 11.Metadata management unit 12 also is in charge of system state, and it receives main with storage system 13 equipment state incidents, according to event information, and converting system state, selection storage system, determination data memory device (position); Also be responsible for the reliability of system, it is according to data access information, automatically backup data.When equipment failure, according to main DATA DISTRIBUTION situation, form available metadata automatically with, subsequent use, external storage system, guarantee the availability of system; And responsible system data automatic on-line recovers, and guarantees the availability and the data consistency of data.Data Management Unit 12 can be deployed in separately on the server.
Main with storage system 13, be used under described normal condition, the data of being asked being provided for AM access module 11.Particularly, main is main storages of distributed file system 1 with storage system 13, is used for all data of saved system; With high readwrite performance is purpose; Use under system's normal condition mainly to read and write data and total data all is kept at and mainly is made up of high speed storing equipment with on the storage system 13 with storage system 13, striping technical support memory bandwidth polymerization raising readwrite performance is stored in use; The block access mode is provided, and AM access module 11 can be with the data of block mode direct read storage on it.
Above-mentioned internal network is to connect AM access module 11, metadata management unit 12, main with the direct-connected network of storage system 13 and accumulator backup system 14, can be that Ethernet also can be internal bus (like the PCIe bus).Metadata mainly is the information of data of description attribute; Being used for supporting as indication memory location, historical information, resource information and file logging etc., for example can be the memory address (can be to comprise memory device number, storage block number, data-bias) of the data (length of request and skew) of the request of preserving.Metadata also can be the inode number of data.The index node of data is kept in the memory device, and the memory address of identification data.According to the inode number of data, can calculate the memory address of index node.
Fig. 2 forms synoptic diagram for another embodiment of the present invention distributed file system; Based on the foregoing description; As shown in Figure 2, this distributed file system 1 comprises AM access module 11, metadata management unit 12, main with storage system 13 and accumulator backup system 14 and external storage system 15.Further, metadata management unit 12 comprises metadata operation module, backup module and abnormal restoring module; Wherein:
Said metadata operation module, respectively with AM access module 11, main be connected with storage system 13, be used to receive said metadata request, return said metadata to leading with storage system 13 or abnormal restoring module request metadata and to AM access module 11; Also be used for upgrading said system state according to the main equipment state incident that reports with storage system 13 that receives; Also be responsible for the condition managing of system, write down the data control information in the said read/write data request.Particularly; Said metadata operation module receives all metadata request of AM access module 11; At first retrieve its metadata cache, if do not find metadata, then need be to main with storage system 13 or accumulator backup system 14 request metadata; Buffer memory obtains metadata, returns metadata at last and gives AM access module 11.
Under normal condition, the retrieval metadata cache if do not have, is then noted the data control information in the metadata request that receives to main with storage system 13 request metadata simultaneously, returns metadata then and gives AM access module 11.Data control information can record internal memory, also can be persisted to database.In the present embodiment, metadata management unit 12 can comprise a data control information logging modle, is connected between metadata operation module and the backup module, is used for the described data control information of permanent storage.
Under ERST, the metadata request that the metadata operation module is sent receiver module 11 is transmitted to the abnormal restoring module, is responsible for obtaining metadata information to main with storage system 13, accumulator backup system 14 by the abnormal restoring module.
Returning to form down, the write data process is with the same under the normal condition, the metadata operation module to main with storage system 13 request metadata; Reading data course, request abnormal restoring module is responsible for processing.
The metadata operation module is responsible for condition managing, main is responsible for the reporting equipment state events with storage system 13, when receiving the equipment failure incident, and the record trouble device id, system state is by normally transferring to unusually; When the equipment that receives recovered incident, system state was by transferring recovery unusually to; When recovery operation was accomplished, after abnormal restoring module circular metadata operation module, system state was transferred to normal by recovery.
Said backup module; Be connected with accumulator backup system 13 with said metadata operation module; Be used to read the said data control information that said metadata operation module records is got off; Generate the data backup control information operation and send to accumulator backup system 14, backup in the accumulator backup system 14 will lead with the data on the storage system 13.Particularly, said backup module moves with background thread under normal condition; Read the data control information that metadata operation module records is for the previous period got off; Analyze the data behaviour in service and generate data backup requests according to backup policy, on accumulator backup system 14, send backup operation control information (action type, file destination path then; The source file path), requirement will be led with the assigned address to the slack storage server of the file backup on the storage system.Said backup module combines backup policy, can realize various data backup scenario flexibly, comprises the full backup scheme, only analyzes data writing operation when analyzing data, in case be to have write data just to need to generate backup request; Comprise the focus backup scenario, an assay readings is according to operating position (read data request number of times, read data frequency), according to the data focus condition in the strategy (can access times, read data frequency) when analyzing data; Comprise backup specific data scheme, can analyze the specific data condition code and generate backup request when analyzing data by tactful specific data condition code; Comprise the backup request information in the data that analyzes, generate backup request according to backup request information.
Said abnormal restoring module; Be connected with accumulator backup system 14 with storage system 13 with said metadata operation module, said backup module, master respectively, be used under said ERST and said returning to form, obtaining said metadata and return to said metadata operation module.Validity unusual, that return to form down and obtain metadata information and safeguard metadata cache that said abnormal restoring module is responsible for, and control is main with storage system 13 data restore operation is to support the master with storage system 13 system availability that breaks down down.
Said abnormal restoring module can comprise following submodule:
First processing sub; Be used under said ERST; For read data request; To lacking the piece detection,, then send the address that solicited messages comprise disappearance block message and external storage system 14 to accumulator backup system 14 if detecting said metadata has the disappearance piece from buffer memory or from the main metadata that obtains with storage system 13; After accumulator backup system 14 returns said metadata, carry out the heavy bindings of block address, the available metadata of recombinating is also sent.Particularly, under ERST, for read data request; Said abnormal restoring module is at first retrieved its metadata cache through first processing sub, if do not find metadata, to leading with storage system 13 request metadata; Then metadata is lacked piece (data storage detects its memory device ID on failed storage equipment) and detect, if find to have the disappearance piece; Again to accumulator backup system 14 request metadata, solicited message comprises disappearance block message (piece number, data ID; Skew), after accumulator backup system 14 returns metadata, carries out the heavy bindings of block address with external storage system 15 addresses.The heavy bindings of block address refers to replace the disappearance piece metadata corresponding information on the accumulator backup system 14 main with the disappearance piece metadata corresponding information on the storage system 13, reassembles into this metadata of available metadata and buffer memory and obtains to accelerate metadata.
Second processing sub is used under said ERST, for write data requests, only to accumulator backup system 14 request metadata and return to said metadata operation module; Also be used to write down data writing operation, in data recovery procedure, to operate data sync to leading with in the storage system 13 according to the write data of record.Particularly; Under ERST, for write data requests, said abnormal restoring module selects accumulator backup system 14 to be the storage target through second processing sub automatically; Only to accumulator backup system 14 request metadata and return request results to the metadata operation module; Like this, AM access module 11 can all be write data on the accumulator backup system 14 according to metadata information, and the abnormal restoring module also can be noted data writing operation; In the data recovery procedure, can arrive data sync main according to these records with storage system 13.
Return to form down, the abnormal restoring module is responsible for online data and is recovered, and can carry out data, services and data simultaneously and recover.The abnormal restoring module representes to lack the piece recovery situation through failed storage equipment bitmap bitmap, and in the rejuvenation, the abnormal restoring module is responsible for safeguarding the renewal of bitmap.When detecting the disappearance piece, earlier according to memory device ID, whether the pairing bitmap of correlation data piece recovers again.
The abnormal restoring module can also comprise the 3rd processing sub, is used for returning to form down said, for read data request; To lacking the piece detection from buffer memory or from the main metadata that obtains with storage system 13; If detecting said metadata has the disappearance piece, then send solicited message and comprise the disappearance block message, after accumulator backup system 14 returns metadata to accumulator backup system 14; Structural deficiency piece recovery operation information is carried out data for main storage system 13 and is recovered; And the failed storage equipment bitmap of renewal main storage system 13, return said metadata.
Particularly, returning to form down, for read data request; The abnormal restoring module is at first retrieved its metadata cache through the 3rd processing sub, if do not find metadata, then to leading with storage system 13 request metadata; Then to lacking the piece detection, if find that the disappearance piece is arranged, again to accumulator backup system 14 request metadata from buffer memory or the main metadata that obtains with storage system 13; Transmit disappearance block message (piece number; Data ID, skew), after accumulator backup system 14 returns metadata; Structure is main with storage system 13 disappearance piece recovery operation information, main with storage system 13 according to this information perhaps to accumulator backup system 14 perhaps to external storage system 15 restore data.After the recovery, upgrade failed storage equipment bitmap bitmap, return metadata.
The abnormal restoring module can also comprise that manages submodule everywhere; Be used for returning to form, for write data requests, to leading with storage system 13 request metadata said; After obtaining said metadata; If detecting said metadata has the disappearance piece, then upgrade the failed storage equipment bitmap of main storage system 13, return said metadata.Particularly; Returning to form down, for write data requests, the abnormal restoring module through the manage everywhere submodule to main with storage system 13 request metadata; After obtaining metadata; Detect the disappearance piece with the failed storage device id,, return request results then to the metadata operation module if the disappearance piece is arranged then directly upgrade failed storage equipment bitmap bitmap.
Returning to form down, the abnormal restoring module starts a backstage and recovers thread.Recover the type of thread, select rejuvenation according to restorer.If memory device is not is not read and write fault (possibly be to turn back to again after the memory device hot plug), only need be saved in the master to the data of between anomalistic period, writing accumulator backup system and use storage system, and delete the metadata cache of heavily binding.
If memory device read-write fault; The main data of traversal with storage system, its rejuvenation is following: the abnormal restoring module is at first retrieved its metadata cache, if find metadata cache; Whether will check has data at accumulator backup system; If data are arranged accumulator backup system (metadata cache of heavily binding is just arranged), then structure is main with storage disappearance piece recovery operation information, main storage system according to this information to the backup storage system restore data.After the recovery, delete the metadata cache of heavily binding and upgrade failed storage equipment bitmap bitmap.
If do not find metadata,, then detect lacking piece from the main metadata that obtains with storage system then to the main storage system request metadata of using at metadata cache; If finding has the disappearance piece, again to the accumulator backup system request metadata, solicited message comprises disappearance block message (piece number; Data ID; Skew), but do not comprise the exterior storage address, after accumulator backup system returns metadata; Structure is main with storage system disappearance piece recovery operation information, and main storage system is according to this information or to backup storage system or to the external storage system restore data.After the recovery, upgrade failed storage equipment bitmap bitmap.
Thread is recovered after the main data with storage system of traversal in the backstage; Also travel through data writing operation under abnormal restoring module records between anomalistic period; The data that responsible handle writes on accumulator backup system are saved in mainly uses storage system, upgrades failed storage equipment bitmap bitmap then and deletes metadata cache.
As shown in Figure 2; Can also comprise the main data management module of using in this distributed file system; Be connected the master with between storage system 13 and the metadata management unit 12, be used to manage main data of being stored, response element request of data and data operation request with storage system 13.Particularly, main be responsible for the response element request of data with data management module, also be responsible for receiving recovery operation information and realize that also data recover, recovery operation information has two types, a kind of have an accumulator backup system address information, and a kind of have an external storage system address information.Lead with data management module according to recovery operation information, perhaps to accumulator backup system or external storage system recover data blocks.
Can also comprise the preliminary data administration module in this distributed file system, be connected between 12 yuan of accumulator backup system 14 and the metadata management lists, be used to manage the data that accumulator backup system 14 is stored, response element request of data and data operation request.Particularly, the preliminary data administration module is responsible for the response element request of data, also is responsible for receiving backup request and realizing data backup.Metadata request has two types; The first kind has the exterior storage address information; Second type does not have the external storage system address; The preliminary data administration module to the difference of these two types of metadata request is: if in accumulator backup system 14 retrieval less than metadata, can and store to external storage system 15 all data of request when then handling the first kind, return the metadata after the storage.When accumulator backup system 14 is handled backup request,, directly also preserve with storage system 13 request msgs to main according to backup request.
In the above-mentioned distributed file system, the master comprises several high speed storing equipment with storage system, and described high speed storing equipment includes but not limited to SCSI hard disk, SATA hard disk, the SSD of high data transmission rate.Main with the also responsible monitoring self of storage system storage device status, and reporting equipment incident (like equipment failure, equipment recovers).The said equipment fault comprises memory device read-write fault, hot plug storage equipment.Accumulator backup system comprises several high speed storing equipment and/or low speed storage device, and wherein, high speed storing equipment includes but not limited to SCSI hard disk, SATA hard disk, the SSD of high data transmission rate; Low speed storage device includes but not limited to the memory device of low data transmission rate.
Recovery operation information described in the foregoing description has two types, a kind of have an accumulator backup system address information, and a kind of have an external storage system address information.Lead with data management module according to recovery operation information, perhaps to accumulator backup system or external storage system recover data blocks.That is to say that the recovery operation information of accumulator backup system address information is arranged, the main recovery operation information that the external storage system address information is arranged from the accumulator backup system recover data blocks with data management module; Main with data management module from the external storage system recover data blocks returning to form down, during read data, will construct top recovery operation information; Like this; The data that the read data service not only were provided but also recovered to be visited at once, according to the principle of locality of data, the data paid close attention to of the just most of users of the data of visit recently; These data of priority restores help improving performance.
About the exterior storage address: under ERST; During read data; Need be to the accumulator backup system request metadata, this comprises the exterior storage address, when this hopes metadata that accumulator backup system itself do not have to be asked; Accumulator backup system returns metadata more earlier from external storage system request msg (being equivalent to backup to accumulator backup system to all data from external storage system).Like this, not only guarantee the read data service but also backup to data the accumulator backup system from external storage system, and can be from accumulator backup system when recovering restore data, thereby expedited data recovers.Returning to form down; During read data or back-end data when recovering thread, may be to the accumulator backup system request metadata, this is not comprise the exterior storage address; Like this; If accumulator backup system itself does not have the metadata of being asked, accumulator backup system can be to outside storage memory system request msg, and accumulator backup system returns sky.Then main with storage system use exterior storage address, to external storage system request msg (the only request disappearance data that piece comprised only need be recovered to lack the data that piece comprises when data are recovered).
The distributed file system that the embodiment of the invention provides writes data into the master with in the storage system, adopts async backup mechanism to backup to accumulator backup system, does not influence the readwrite performance of high-speed equipment; Can online restore data, realize automatically in the external service process of data that data recover, accelerate rejuvenation; And do not need during restore data to calculate, even not backup of data still can recover through exterior storage; Under the CDN environment, usage policy Backup Data mechanism is obtained with the available external storage does not have backed up data, but implementation part data backup and do not influence availability; In addition, can use cheap memory device to form accumulator backup system, reduce cost of products.
Fig. 3 is embodiment of the invention distributed file system treatment scheme synoptic diagram under normal condition, and as shown in Figure 3, this flow process comprises:
Step 4, master receive backup request with data management module, to leading with the memory device reading of data in the storage system;
Step 6, preliminary data administration module are write data on the memory device in the accumulator backup system;
Fig. 4 be embodiment of the invention distributed file system under ERST and be the treatment scheme synoptic diagram under the read data request situation, as shown in Figure 4, this flow process comprises:
Step 4, abnormal restoring module turned to for the 6th step earlier in metadata information cache lookup metadata if find;
Whether step 6, abnormal restoring module check have the disappearance piece to exist from metadata cache or the main metadata information that returns with storage system; Have then and initiate metadata request and be with a last external backup control information (to comprise the external storage system position to accumulator backup system; Data Position information); Then do not return metadata information to the metadata operation module, and turned to for the 9th step;
If these data of step 7 have backed up at accumulator backup system; Then the preliminary data administration module returns metadata; Otherwise the preliminary data administration module to the external storage system request msg, at accumulator backup system, and returns metadata information to the gained data storage according to the external backup control information;
Step 12, AM access module response data return to application program.
Fig. 5 be embodiment of the invention distributed file system under ERST and be the treatment scheme synoptic diagram under the write data requests situation, as shown in Figure 5, this flow process comprises:
Step 4, abnormal restoring module are directly initiated metadata request to accumulator backup system;
Step 6, abnormal restoring module are returned metadata information and are given the metadata operation module;
Response readings is according to the processing of request schematic flow sheet under returning to form for embodiment of the invention distributed file system for Fig. 6, and as shown in Figure 6, this flow process comprises:
Step 4, abnormal restoring module retrieval metadata cache if do not find, are then used the storage system request metadata to main, jump to for the 6th step if find;
Whether step 6, abnormal restoring module check lack piece, if having then to the accumulator backup system request metadata, if do not have, then jump to for the 11st step;
The metadata that step 7, abnormal restoring module are returned according to accumulator backup system, structure are recovered control information and are issued the main storage system of using;
Step 12, metadata operation module are returned metadata and are given AM access module;
Fig. 7 embodiment of the invention distributed file system is at the treatment scheme synoptic diagram of response write data requests down that returns to form, and as shown in Figure 7, this flow process comprises:
Step 4, abnormal restoring module are to the main storage system request metadata of using;
Whether step 6, abnormal restoring module check lack piece, if having, then upgrade the disappearance piece bitmap bitmap that has recovered, and the pairing bitmap of disappearance piece is set to 1;
Fig. 8 is the treatment scheme synoptic diagram of embodiment of the invention distributed file system data recovery procedure under the system recovery state, and is as shown in Figure 8, and this flow process comprises:
If step 2 finds metadata cache, whether back-end data is recovered the thread inspection has data on accumulator backup system, if do not have data on accumulator backup system, jumps to for the 11st step;
If step 3 has data on accumulator backup system, back-end data is recovered the thread constructing master and is lacked piece recovery operation information and issue the main storage system of using with storage;
Step 4, main receiving with storage system lack piece recovery operation information and execution; Accumulator backup system data
to the disappearance piece, are returned disappearance piece recovery operation result;
If step 6 does not find metadata at metadata cache, then back-end data is recovered thread to the main storage system request metadata of using;
If step 8 has the disappearance piece, back-end data is recovered thread request to the accumulator backup system restore data;
If the 8th step of step 9 backs up successfully, then jumped to for the 11st step;
If the 8th step of step 10 backup failure, then back-end data is recovered thread request to the external storage system restore data;
One of ordinary skill in the art will appreciate that: all or part of step that realizes said method embodiment can be accomplished through the relevant hardware of programmed instruction; Aforesaid program can be stored in the computer read/write memory medium; This program the step that comprises said method embodiment when carrying out; And aforesaid storage medium comprises: various media that can be program code stored such as ROM, RAM, magnetic disc or CD.
What should explain at last is: above embodiment is only in order to explaining technical scheme of the present invention, but not to its restriction; Although with reference to previous embodiment the present invention has been carried out detailed explanation, those of ordinary skill in the art is to be understood that: it still can be made amendment to the technical scheme that aforementioned each embodiment put down in writing, and perhaps part technical characterictic wherein is equal to replacement; And these are revised or replacement, do not make the spirit and the scope of the essence disengaging various embodiments of the present invention technical scheme of relevant art scheme.
Claims (6)
1. distributed file system; It is characterized in that; Comprise: AM access module, the metadata management unit that is connected with said AM access module; And be connected with said metadata management unit respectively main with storage system and accumulator backup system, said AM access module also is connected with said accumulator backup system with storage system with said leading respectively; Said AM access module, said metadata management unit, said main with connecting through system bus between storage system and the said accumulator backup system; Also comprise external storage system, said external storage system is connected with said accumulator backup system with storage system with said master through network; Wherein:
Said AM access module; Be used to receive the read/write data request; Send the data metadata corresponding that metadata request is asked to obtain to said metadata management unit, and use said metadata and read and write data with storage system or said accumulator backup system to said master;
Said metadata management unit is used for when the said metadata of said AM access module request, search the data of being asked said master with the position on storage system or the said accumulator backup system, and construct said metadata, return to said AM access module; Said master uses storage system, and being used for when said distributed file system is in normal condition is that said AM access module provides the data of being asked; Said accumulator backup system is used for when said distributed file system is in ERST or returns to form, for said master provides data backup with storage system; Said external storage system is used to said master and with storage system data backup is provided;
Wherein, said metadata management unit comprises: metadata operation module, backup module and abnormal restoring module; Wherein:
Said metadata operation module; Be connected with storage system with said AM access module, said master respectively; Be used to receive said metadata request, return said metadata with storage system or said abnormal restoring module request metadata and to said AM access module to said master; Also be used for the equipment state incident that reports with storage system according to the said master who receives, upgrade the system state of distributed file system; Also be used for writing down the data control information of said read/write data request;
Said backup module; Be connected with said accumulator backup system with said metadata operation module; Be used to read the said data control information that said metadata operation module records is got off; Generate the data backup control information operation and send to said accumulator backup system, so that said master is backuped in the said accumulator backup system with the data on the storage system;
Said abnormal restoring module; Be connected with said accumulator backup system with storage system with said metadata operation module, said backup module, said master respectively, be used under said ERST and said returning to form, obtaining said metadata and return to said metadata operation module.
2. distributed file system according to claim 1 is characterized in that, said metadata management unit also comprises:
The data control information logging modle is connected between said metadata operation module and the said backup module, is used to store said data control information.
3. distributed file system according to claim 1 is characterized in that, said abnormal restoring module comprises:
First processing sub; Be used under said ERST; For read data request; To lacking the piece detection,, then send the address that solicited message comprises disappearance block message and said external storage system to said accumulator backup system if detecting said metadata has the disappearance piece from buffer memory or from the metadata that said master obtains with storage system; After said accumulator backup system returns said metadata, carry out the heavy bindings of block address, the available metadata of recombinating is also sent;
Second processing sub is used under said ERST, for write data requests, only to said accumulator backup system request metadata and return to said metadata operation module; Also be used to write down data writing operation, in data recovery procedure, to operate data sync to said main with in the storage system according to the write data of record;
The 3rd processing sub is used for returning to form down said, for read data request; To lacking the piece detection from buffer memory or from the metadata that said master obtains with storage system; If detecting said metadata has the disappearance piece, then send solicited message and comprise the disappearance block message, after said accumulator backup system returns metadata to said accumulator backup system; Structural deficiency piece recovery operation information is carried out data for said master with storage system and is recovered; And upgrade said main failed storage equipment bitmap with storage system, return said metadata;
Manages submodule everywhere; Be used for returning to form down, for write data requests, to the said main storage system request metadata of using said; After obtaining said metadata; If detecting said metadata has the disappearance piece, then upgrade said main failed storage equipment bitmap with storage system, return said metadata.
4. according to claim 1 or 2 or 3 described distributed file systems, it is characterized in that, also comprise:
The master uses data management module, is connected said master with between storage system and the said metadata management unit, is used to manage said main data of storing with storage system, response element request of data and data operation request; And
The preliminary data administration module is connected between said accumulator backup system and the said metadata management unit, is used to manage the data that said accumulator backup system is stored, response element request of data and data operation request.
5. distributed file system according to claim 4 is characterized in that, said master comprises several high speed storing equipment with storage system, and said high speed storing equipment includes but not limited to SCSI hard disk, SATA hard disk, the SSD of high data transmission rate;
Said accumulator backup system comprises several high speed storing equipment and/or low speed storage device, and wherein, said high speed storing equipment comprises SCSI hard disk, SATA hard disk, SSD.
6. distributed file system according to claim 1 is characterized in that,
Said normal condition refers to main not break down with storage system;
Said ERST refers to main break down with storage system, and is main with storage system and accumulator backup system co-operation, coordinates to accomplish data storage, preservation coordination result, and wherein, said coordination result is used for data recovery procedure;
Said returning to form is meant through after the ERST, uses the data on the accumulator backup system to recover main with the data in the storage system.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010105872357A CN102024044B (en) | 2010-12-08 | 2010-12-08 | Distributed file system |
| PCT/CN2011/079685 WO2012075845A1 (en) | 2010-12-08 | 2011-09-15 | Distributed file system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010105872357A CN102024044B (en) | 2010-12-08 | 2010-12-08 | Distributed file system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102024044A CN102024044A (en) | 2011-04-20 |
| CN102024044B true CN102024044B (en) | 2012-11-21 |
Family
ID=43865341
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2010105872357A Active CN102024044B (en) | 2010-12-08 | 2010-12-08 | Distributed file system |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN102024044B (en) |
| WO (1) | WO2012075845A1 (en) |
Families Citing this family (26)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102024044B (en) * | 2010-12-08 | 2012-11-21 | 华为技术有限公司 | Distributed file system |
| CN103095767B (en) * | 2011-11-03 | 2019-04-23 | 中兴通讯股份有限公司 | Distributed cache system and data reconstruction method based on distributed cache system |
| CN103220162B (en) * | 2012-01-19 | 2016-08-31 | 百度在线网络技术（北京）有限公司 | The fault-tolerant optimization method and device of SCSI based on HDFS |
| GB2503016B (en) * | 2012-06-14 | 2017-10-04 | Draeger Safety Uk Ltd | A telemetry monitoring system and a data recovery method for a telemetry monitoring system |
| CN103516736A (en) * | 2012-06-20 | 2014-01-15 | 中兴通讯股份有限公司 | Data recovery method of distributed cache system and a data recovery device of distributed cache system |
| CN102867035B (en) * | 2012-08-28 | 2015-09-23 | 浪潮(北京)电子信息产业有限公司 | A kind of distributed file system cluster high availability method and device |
| CN102890716B (en) * | 2012-09-29 | 2017-08-08 | 南京中兴新软件有限责任公司 | The data back up method of distributed file system and distributed file system |
| CN103049390B (en) * | 2012-12-14 | 2016-03-09 | 华为技术有限公司 | The data processing method of apply metadata and storage system |
| CN103902349B (en) * | 2012-12-27 | 2017-05-31 | 中国移动通信集团江西有限公司 | A kind of virtual platform storage managing server and its management method |
| CN103167026B (en) * | 2013-02-06 | 2016-05-18 | 数码辰星科技发展（北京）有限公司 | A kind of cloud store environmental data processing method, system and equipment |
| CN103207894A (en) * | 2013-03-14 | 2013-07-17 | 深圳市知正科技有限公司 | Multipath real-time video data storage system and cache control method thereof |
| CN103347086B (en) * | 2013-07-11 | 2016-06-29 | 南京大学 | Collaborative kernel construction method based on Distributed Coordination algorithm |
| WO2015015502A1 (en) * | 2013-07-29 | 2015-02-05 | Hewlett-Packard Development Company, L.P. | Writing to files and file meta-data |
| CN103473184B (en) * | 2013-08-01 | 2016-08-10 | 记忆科技(深圳)有限公司 | The caching method of file system and system |
| CN104657392B (en) * | 2013-11-25 | 2020-02-11 | 腾讯科技（深圳）有限公司 | Method and device for realizing retrieval abnormity restoration |
| CN104135539B (en) | 2014-08-15 | 2018-03-16 | 华为技术有限公司 | Date storage method, SDN controllers and distributed network storage system |
| CN104202387B (en) * | 2014-08-27 | 2017-11-24 | 华为技术有限公司 | A kind of metadata restoration methods and relevant apparatus |
| CN105915600A (en) * | 2016-04-13 | 2016-08-31 | 乐视控股（北京）有限公司 | Data writing-in method based CDN network system and CDN network system thereof |
| CN108108422A (en) * | 2017-12-15 | 2018-06-01 | 郑州云海信息技术有限公司 | A kind of metadata acquisition methods, device and the medium of Ceph file system |
| CN110096220B (en) * | 2018-01-31 | 2020-06-26 | 华为技术有限公司 | Distributed storage system, data processing method and storage node |
| CN108388604B (en) * | 2018-02-06 | 2022-06-10 | 平安科技（深圳）有限公司 | User authority data management apparatus, method and computer readable storage medium |
| CN109327539A (en) * | 2018-11-15 | 2019-02-12 | 上海天玑数据技术有限公司 | A kind of distributed block storage system and its data routing method |
| CN110659157A (en) * | 2019-08-30 | 2020-01-07 | 安徽芃睿科技有限公司 | Distributed multi-language retrieval platform and method for lossless recovery |
| CN111026432A (en) * | 2019-12-06 | 2020-04-17 | 中国建设银行股份有限公司 | Big data processing platform, platform construction method and storage medium |
| CN112532525B (en) * | 2020-11-25 | 2022-11-25 | 北京金山云网络技术有限公司 | Processing method, device and system for equipment recovery service |
| JP7232281B2 (en) * | 2021-03-29 | 2023-03-02 | 株式会社日立製作所 | Storage system and storage system control method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1955939A (en) * | 2006-10-13 | 2007-05-02 | 清华大学 | Backup and Recovery Method Based on Virtual Memory Disk |
| CN101394424A (en) * | 2008-10-27 | 2009-03-25 | 中国科学院计算技术研究所 | System and method for disk-level network data backup |
| CN101408855A (en) * | 2008-11-07 | 2009-04-15 | 北京威视数据系统有限公司 | Method for protecting remote backup equipment of temporary abnormal by continuous data protective system |
| CN101436151A (en) * | 2008-12-01 | 2009-05-20 | 成都索贝数码科技股份有限公司 | Data real time backup method and system based on file system |
| CN101436149A (en) * | 2008-12-19 | 2009-05-20 | 华中科技大学 | Method for rebuilding data of magnetic disk array |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4982651B2 (en) * | 2000-02-04 | 2012-07-25 | リアルネットワークス・インコーポレイテッド | System including distributed media network and metadata server |
| US6728849B2 (en) * | 2001-12-14 | 2004-04-27 | Hitachi, Ltd. | Remote storage system and method |
| DE10393771T5 (en) * | 2002-11-20 | 2006-03-30 | Filesx Ltd. | Fast backup storage and fast data recovery (FBSRD) |
| CN101539873B (en) * | 2009-04-15 | 2011-02-09 | 成都市华为赛门铁克科技有限公司 | Data recovery method, data node and distributed file system |
| CN102024044B (en) * | 2010-12-08 | 2012-11-21 | 华为技术有限公司 | Distributed file system |
-
2010
- 2010-12-08 CN CN2010105872357A patent/CN102024044B/en active Active
-
2011
- 2011-09-15 WO PCT/CN2011/079685 patent/WO2012075845A1/en active Application Filing
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1955939A (en) * | 2006-10-13 | 2007-05-02 | 清华大学 | Backup and Recovery Method Based on Virtual Memory Disk |
| CN101394424A (en) * | 2008-10-27 | 2009-03-25 | 中国科学院计算技术研究所 | System and method for disk-level network data backup |
| CN101408855A (en) * | 2008-11-07 | 2009-04-15 | 北京威视数据系统有限公司 | Method for protecting remote backup equipment of temporary abnormal by continuous data protective system |
| CN101436151A (en) * | 2008-12-01 | 2009-05-20 | 成都索贝数码科技股份有限公司 | Data real time backup method and system based on file system |
| CN101436149A (en) * | 2008-12-19 | 2009-05-20 | 华中科技大学 | Method for rebuilding data of magnetic disk array |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2012075845A1 (en) | 2012-06-14 |
| CN102024044A (en) | 2011-04-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102024044B (en) | Distributed file system | |
| US9146684B2 (en) | Storage architecture for server flash and storage array operation | |
| US6691209B1 (en) | Topological data categorization and formatting for a mass storage system | |
| US8689047B2 (en) | Virtual disk replication using log files | |
| US8438332B2 (en) | Apparatus and method to maintain write operation atomicity where a data transfer operation crosses a data storage medium track boundary | |
| US9946655B2 (en) | Storage system and storage control method | |
| US7975168B2 (en) | Storage system executing parallel correction write | |
| CN102521072B (en) | Virtual tape library equipment and data recovery method | |
| EP2879040B1 (en) | Data storage method, data storage apparatus, and storage device | |
| CN101755257B (en) | Managing the copying of writes from primary storages to secondary storages across different networks | |
| CN105183581A (en) | Database disaster tolerance system | |
| CN101667181A (en) | Method, device and system for data disaster tolerance | |
| CN103037004A (en) | Implement method and device of cloud storage system operation | |
| CN103942112A (en) | Magnetic disk fault-tolerance method, device and system | |
| US20090198942A1 (en) | Storage system provided with a plurality of controller modules | |
| US11868625B2 (en) | Alert tracking in storage | |
| CN102177496A (en) | System and method for transferring data between different RAID data storage types for current data and replay data | |
| CN104424052A (en) | Automatic redundant distributed storage system and method | |
| CN100524235C (en) | Recovery operations in storage networks | |
| CN111309245A (en) | Layered storage writing method and device, reading method and device and system | |
| CN102135862B (en) | Disk storage system and data access method thereof | |
| US10642508B2 (en) | Method to limit impact of partial media failure of disk drive and detect/report the loss of data for objects due to partial failure of media | |
| US20060190682A1 (en) | Storage system, method for processing, and program | |
| US10915405B2 (en) | Methods for handling storage element failures to reduce storage device failure rates and devices thereof | |
| CN101390059A (en) | Apparatus for concurrent raid array relocation |
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 | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20210423 Address after: Unit 3401, unit a, building 6, Shenye Zhongcheng, No. 8089, Hongli West Road, Donghai community, Xiangmihu street, Futian District, Shenzhen, Guangdong 518040 Patentee after: Honor Device Co.,Ltd. Address before: 518129 Bantian HUAWEI headquarters office building, Longgang District, Guangdong, Shenzhen Patentee before: HUAWEI TECHNOLOGIES Co.,Ltd. |
|
| TR01 | Transfer of patent right | ||
| CP03 | Change of name, title or address |
Address after: Unit 3401, unit a, building 6, Shenye Zhongcheng, No. 8089, Hongli West Road, Donghai community, Xiangmihu street, Futian District, Shenzhen, Guangdong 518040 Patentee after: Honor Terminal Co.,Ltd. Country or region after: China Address before: 3401, unit a, building 6, Shenye Zhongcheng, No. 8089, Hongli West Road, Donghai community, Xiangmihu street, Futian District, Shenzhen, Guangdong Patentee before: Honor Device Co.,Ltd. Country or region before: China |
|
| CP03 | Change of name, title or address |