CN103986771A - High-availability cluster management method independent of shared storage - Google Patents
High-availability cluster management method independent of shared storage Download PDFInfo
- Publication number
- CN103986771A CN103986771A CN201410218141.0A CN201410218141A CN103986771A CN 103986771 A CN103986771 A CN 103986771A CN 201410218141 A CN201410218141 A CN 201410218141A CN 103986771 A CN103986771 A CN 103986771A
- Authority
- CN
- China
- Prior art keywords
- data
- node
- write
- host node
- magnetic disk
- 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.)
- Pending
Links
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a high-availability cluster management method independent of shared storage. According to the method, no additional storage device needs to be added, a cluster is configured according to the requirements, disk data are synchronously or asynchronously transmitted by means of block devices of cluster nodes and network bandwidth limitation, and disk data transmission and share between nodes are ensured. According to the high-availability cluster management method, a disk data transmission device independent of shared storage is provided, the disk data transmission device comprises a synchronous tracking module, a block data synchronization module, a data change mapping module and a synchronization link module. According to the high-availability cluster management method, disk data transmission and share between the nodes are achieved through the block devices of the cluster nodes, cost caused by storage devices can be greatly reduced, and configuration complexity and cluster management risks are reduced.
Description
Technical field
The present invention relates to the shared management method of data in magnetic disk in highly available cluster system, specifically a kind of high availability cluster management method of sharing storage that do not rely on.
Background technology
Highly available cluster system refers to and can ensure that business do not interrupt for 24 hours, and in the time of system generation problem or application software generation problem, quick-recovery business soon, is controlled at downtime the system of minute rank as far as possible.The service resources of managing in highly available cluster system is of a great variety, comprise process resource, Floating IP address resource, Service Source, script resource, disk resource, database monitoring resource and middleware monitoring resource etc., various resources and its supervision resource can customize configuration and management according to application scenarios.Wherein, disk resource is applied required total data for stored key type.
In dual computer group management system, be to dispose in the following way and configure disk resource at present: add an extra memory device, and connect with memory device on two-shipper, ensure can access this storage on each node.In the time that cluster arranges, while adding disk resource, fill in the drive that this memory device is corresponding, after starting, cluster can ensure the exclusive reference of two-shipper to this memory device by disk filtration drive, thus the consistency of protected data and integrality.But this by adding the method for storage device management data in magnetic disk, need additionally to drop into higher cost and remove to buy memory device, and require in cluster each node normal with being connected of memory device, require the many factors of extraneous guarantee.Therefore, need a kind of data in magnetic disk management method of sharing storage that do not rely on, carry out data transfer access by the disk of node itself, thereby ensure consistency and the integrality of data.And the each node of cluster has the block device of oneself, can utilize the block device of self realize the storage of total data and enjoy completely.Therefore invent a kind of high availability cluster management method that does not rely on shared storage and just seem very necessary.
Summary of the invention
The weak point existing for prior art, the present invention proposes a kind of high availability cluster management method of sharing storage that do not rely on.
A kind of high availability cluster management method of sharing storage that do not rely on of the present invention, the technical scheme that solves the problems of the technologies described above employing is as follows: described high availability cluster management method does not need to add extra memory device, allocation cluster according to demand, by the block device of clustered node self, and network bandwidth limitations, synchronous or asynchronous transmission data in magnetic disk, ensures between each node data in magnetic disk transmission and shares; This high availability cluster management method has proposed a kind of data in magnetic disk transmitting device of sharing storage that do not rely on, and this data in magnetic disk transmitting device comprises synchronous tracking module, blocks of data synchronization module, data change mapping block and synchronization link module; Wherein,
Described synchronous tracking module is a cyclic buffer, is conventionally realized by round-robin queue, for recording the io that will write designated disk receiving from application program, to maintain write sequence fidelity;
Described blocks of data synchronization module is to carry out the transmission of two-shipper node data in magnetic disk, and realizes the module of data dispatch function, belongs to the nucleus module of whole data in magnetic disk management;
Described data change mapping block, for follow the tracks of write operation in the time that synchronous tracking module overflows, can be avoided the data on supplementary set group node to re-start completely synchronously like this;
Described synchronization link module is the module that connects host node and auxiliary node, for set up link between host node and auxiliary node, and provides the medium of transfer of data between two nodes.
The beneficial effect that the high availability cluster management method that does not rely on shared storage of the present invention has:
This high availability cluster management method can solve because adding the high and connection constraints problem of cost that extra memory device brings; On the one hand, the block device storage data that carry by clustered node itself, have reduced extra memory device expense; On the one hand, reduce clustered node and be connected complexity with equipment room, reduced the risk of high availability cluster management.
Brief description of the drawings
Accompanying drawing 1 is not for relying on the high availability cluster management method module composition diagram of sharing storage;
Accompanying drawing 2 is copy data flow process figure under synchronous mode;
Accompanying drawing 3 is copy data flow process figure under asynchronous mode.
Embodiment
With reference to Figure of description, a kind of high availability cluster management method that does not rely on shared storage of the present invention is described in detail below.
High availability cluster management method of the present invention does not need to add extra memory device, allocation cluster according to demand, by the block device of clustered node self, and network bandwidth limitations, synchronous or asynchronous transmission data in magnetic disk, ensures between each node data in magnetic disk transmission and shares; This high availability cluster management method has proposed a kind of data in magnetic disk transmitting device of sharing storage that do not rely on, and this data in magnetic disk transmitting device comprises synchronous tracking module, blocks of data synchronization module, data change mapping block and synchronization link module.
Respectively above-mentioned synchronous tracking module, blocks of data synchronization module, data change mapping block and synchronization link module are elaborated below:
Described synchronous tracking module is a cyclic buffer, is conventionally realized by round-robin queue, for recording the io that will write designated disk receiving from application program, to maintain write sequence fidelity;
Described blocks of data synchronization module is the module of carrying out the transmission of two-shipper node data in magnetic disk and realizing data dispatch function, belongs to the nucleus module of whole data in magnetic disk management;
Described data change mapping block, for follow the tracks of write operation in the time that synchronous tracking module overflows, can be avoided the data on supplementary set group node to re-start completely synchronously like this;
Described synchronization link module is the module that connects host node and auxiliary node, for set up link between host node and auxiliary node, and provides the medium of transfer of data between two nodes.
Described in this high availability cluster management method, in blocks of data synchronization module, provide the data in magnetic disk method of synchronization between two kinds of nodes, synchronous mode and asynchronous mode;
Under synchronous mode, carry out data when synchronous, the data on auxiliary node are up-to-date completely, if host node generation disaster can be recovered data from the auxiliary node of the normal operation of any continuation by milli without loss; If auxiliary node must reflect be successfully completed on host node all and write, choice for use synchronous mode is proper; The advantage of synchronous mode is the major-minor synchronisation of nodes of data, keeps up-to-date; Shortcoming is to affect application program capacity in the environment of high latency or Bandwidth-Constrained, and when auxiliary node is more, the slowest node of speed becomes the bottleneck that application program postpones;
Under asynchronous mode, carry out data when synchronous, when the renewal of host node blocks of data is recorded in the synchronous tracking module of host node, complete renewal, follow-uply by synchronization link module, application data is synchronized to auxiliary node; Asynchronous mode does not ensure that data are up-to-date all the time, but the impact of its application programs performance is less, and can use and have more cost-benefit communication; This pattern can ensure on auxiliary node disk, to carry out all renewals that host node disk completed, and has some delays; Therefore, the advantage of asynchronous mode is that application program capacity impact is little, has write and has returned soon; Shortcoming is that auxiliary node data are not up-to-date in real time, and fashionable when there being application data to write, auxiliary node Data Update has delay.
Accompanying drawing 2 is copy data flow process figure under synchronous mode, and as shown in Figure 2, under described synchronous mode, copy data flow process is:
Step 1: data in magnetic disk transmitting device receives writing of application program on host node;
Step 2: will write and write the synchronous tracking module of host node;
Step 3: will write and write on host node disk, and simultaneously by synchronization link module, send to auxiliary node and wait for the network validation from auxiliary node writing;
Step 4: on auxiliary node, data in magnetic disk transmitting device receives and process writing information, then network validation is sent to host node;
Step 5: will write in the disk sending on auxiliary node;
Step 6: after data are write to the disk being written on auxiliary node, the data in magnetic disk transmitting device on auxiliary node can send a data validation to host node;
Step 7: after host node is received the data validation from all auxiliary node main frames, data in magnetic disk transmitting device can will write to be labeled as and complete in synchronous tracking module;
Under synchronous mode, in the flow process of copy data, as long as receive writing information in data in magnetic disk transmitting device kernel internal memory, auxiliary node will send network validation, will from application program postpones, eliminate the required time of auxiliary node disk that writes like this.On host node, data in magnetic disk transmitting device can not waited for and write data into auxiliary node disk, can improve like this application program capacity.But data in magnetic disk transmitting device can be followed the tracks of all this type of and not yet write the confirmation writing information of disk.If auxiliary node collapse before the disk writing on auxiliary node, or host node collapsed before receiving data validation, the data in magnetic disk transmitting device writing information of these tracking of can resetting.
Accompanying drawing 3 is copy data flow process figure under asynchronous mode, and as shown in Figure 2, under described asynchronous mode, copy data flow process is:
Step 1: data in magnetic disk transmitting device receives writing of application program on host node;
Step 2: will write and write the synchronous tracking module of host node;
Step 3: confirm to have write to application program;
Step 4: will write and send to asynchronous auxiliary node main frame according to receive the order writing at host node, will write on host node and write local disk simultaneously;
Step 5: auxiliary node receives that writing of host node is fashionable, sends network validation to host node; In the time that host node is received network validation, it is known in auxiliary node data in magnetic disk transmitting device inter-sync tracking module and has received that this writes;
Step 6: data in magnetic disk transmitting device will write the disk sending on auxiliary node;
Step 7: after to be written completing, auxiliary node sends data validation to host node;
Step 8: in the time that host node is received data validation, data in magnetic disk transmitting device can will write to be labeled as and complete in synchronous tracking module.
Above-mentioned embodiment is only concrete case of the present invention; scope of patent protection of the present invention includes but not limited to above-mentioned embodiment; suitable variation or replacement that person of an ordinary skill in the technical field any claims according to the invention and any does it, all should fall into scope of patent protection of the present invention.
Claims (6)
1. one kind does not rely on the high availability cluster management method of sharing storage, it is characterized in that, described high availability cluster management method allocation cluster according to demand, by the block device of clustered node self, and network bandwidth limitations, synchronous or asynchronous transmission data in magnetic disk, ensures between each node data in magnetic disk transmission and shares; This high availability cluster management method has proposed a kind of data in magnetic disk transmitting device of sharing storage that do not rely on, and this data in magnetic disk transmitting device comprises synchronous tracking module, blocks of data synchronization module, data change mapping block and synchronization link module; Wherein,
Described synchronous tracking module is a cyclic buffer, is conventionally realized by round-robin queue, for recording the io that will write designated disk receiving from application program, to maintain write sequence fidelity;
Described blocks of data synchronization module is to carry out the transmission of two-shipper node data in magnetic disk, and realizes the module of data dispatch function, belongs to the nucleus module of whole data in magnetic disk management;
Described data change mapping block for following the tracks of write operation in the time that synchronous tracking module overflows;
Described synchronization link module is the module that connects host node and auxiliary node, for set up link between host node and auxiliary node, and provides the medium of transfer of data between two nodes.
2. the high availability cluster management method of sharing storage that do not rely on according to claim 1, is characterized in that, in described blocks of data synchronization module, provides the data in magnetic disk method of synchronization between two kinds of nodes: synchronous mode and asynchronous mode.
3. the high availability cluster management method of sharing storage that do not rely on according to claim 2, it is characterized in that, under synchronous mode, carry out data when synchronous, data on auxiliary node are up-to-date completely, if host node generation disaster, can recover data from the auxiliary node of the normal operation of any continuation by milli without loss; By synchronous mode, auxiliary node can reflect that be successfully completed on host node all write.
4. the high availability cluster management method of sharing storage that do not rely on according to claim 2, it is characterized in that, under asynchronous mode, carry out data when synchronous, the renewal of host node blocks of data is recorded in the synchronous tracking module of host node and completes renewal, follow-uply by synchronization link module, application data is synchronized to auxiliary node.
5. the high availability cluster management method of sharing storage that do not rely on according to claim 3, is characterized in that, under described synchronous mode, copy data flow process is:
Step 1: data in magnetic disk transmitting device receives writing of application program on host node;
Step 2: will write and write the synchronous tracking module of host node;
Step 3: will write and write on host node disk, and simultaneously by synchronization link module, send to auxiliary node and wait for the network validation from auxiliary node writing;
Step 4: on auxiliary node, data in magnetic disk transmitting device receives and process writing information, then network validation is sent to host node;
Step 5: will write in the disk sending on auxiliary node;
Step 6: after data are write to the disk being written on auxiliary node, the data in magnetic disk transmitting device on auxiliary node can send a data validation to host node;
Step 7: after host node is received the data validation from all auxiliary node main frames, data in magnetic disk transmitting device can will write to be labeled as and complete in synchronous tracking module.
6. the high availability cluster management method of sharing storage that do not rely on according to claim 4, is characterized in that, under described asynchronous mode, copy data flow process is:
Step 1: data in magnetic disk transmitting device receives writing of application program on host node;
Step 2: will write and write the synchronous tracking module of host node;
Step 3: confirm to have write to application program;
Step 4: send to asynchronous auxiliary node main frame according to receiving the order writing at host node, writing, will write on host node and write local disk simultaneously;
Step 5: auxiliary node receives that writing of host node is fashionable, sends network validation to host node; In the time that host node is received network validation, it is known in auxiliary node data in magnetic disk transmitting device inter-sync tracking module and has received that this writes;
Step 6: data in magnetic disk transmitting device will write the disk sending on auxiliary node;
Step 7: after to be written completing, auxiliary node sends data validation to host node;
Step 8: in the time that host node is received data validation, data in magnetic disk transmitting device can will write to be labeled as and complete in synchronous tracking module.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410218141.0A CN103986771A (en) | 2014-05-22 | 2014-05-22 | High-availability cluster management method independent of shared storage |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410218141.0A CN103986771A (en) | 2014-05-22 | 2014-05-22 | High-availability cluster management method independent of shared storage |
Publications (1)
Publication Number | Publication Date |
---|---|
CN103986771A true CN103986771A (en) | 2014-08-13 |
Family
ID=51278594
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410218141.0A Pending CN103986771A (en) | 2014-05-22 | 2014-05-22 | High-availability cluster management method independent of shared storage |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103986771A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105487929A (en) * | 2015-11-19 | 2016-04-13 | 山东大学 | Method for managing shared data of lens in cluster rendering process |
CN106101208A (en) * | 2016-06-10 | 2016-11-09 | 北京银信长远科技股份有限公司 | The method building cross-platform high-availability system based on Ethernet |
CN106412011A (en) * | 2016-08-30 | 2017-02-15 | 广州鼎甲计算机科技有限公司 | High-availability cluster system without shared storage among multiple nodes, and implementation |
CN111372100A (en) * | 2020-04-21 | 2020-07-03 | 白杨 | End-to-end content distribution network system and distribution method based on distributed election |
CN113504933A (en) * | 2021-05-28 | 2021-10-15 | 曲阜师范大学 | Workshop machine updating method and system |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1512376A (en) * | 2002-12-31 | 2004-07-14 | 联想(北京)有限公司 | Central control method for large machine group system |
CN103118121A (en) * | 2013-02-19 | 2013-05-22 | 浪潮电子信息产业股份有限公司 | Application method of high availability cluster in virtualization technology |
CN103152416A (en) * | 2013-03-04 | 2013-06-12 | 浪潮电子信息产业股份有限公司 | Dynamic management method for improving high availability of online clustering group |
CN103312541A (en) * | 2013-05-28 | 2013-09-18 | 浪潮电子信息产业股份有限公司 | Management method of high-availability mutual backup cluster |
CN103631903A (en) * | 2013-11-22 | 2014-03-12 | 曙光信息产业股份有限公司 | System for synchronizing data of database |
-
2014
- 2014-05-22 CN CN201410218141.0A patent/CN103986771A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1512376A (en) * | 2002-12-31 | 2004-07-14 | 联想(北京)有限公司 | Central control method for large machine group system |
CN103118121A (en) * | 2013-02-19 | 2013-05-22 | 浪潮电子信息产业股份有限公司 | Application method of high availability cluster in virtualization technology |
CN103152416A (en) * | 2013-03-04 | 2013-06-12 | 浪潮电子信息产业股份有限公司 | Dynamic management method for improving high availability of online clustering group |
CN103312541A (en) * | 2013-05-28 | 2013-09-18 | 浪潮电子信息产业股份有限公司 | Management method of high-availability mutual backup cluster |
CN103631903A (en) * | 2013-11-22 | 2014-03-12 | 曙光信息产业股份有限公司 | System for synchronizing data of database |
Non-Patent Citations (1)
Title |
---|
姚杰: ""基于共享存储的高可用性系统设计及管理模块实现"", 《中国优秀博硕士学位论文全文数据库 信息科技辑》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105487929A (en) * | 2015-11-19 | 2016-04-13 | 山东大学 | Method for managing shared data of lens in cluster rendering process |
CN106101208A (en) * | 2016-06-10 | 2016-11-09 | 北京银信长远科技股份有限公司 | The method building cross-platform high-availability system based on Ethernet |
CN106412011A (en) * | 2016-08-30 | 2017-02-15 | 广州鼎甲计算机科技有限公司 | High-availability cluster system without shared storage among multiple nodes, and implementation |
CN111372100A (en) * | 2020-04-21 | 2020-07-03 | 白杨 | End-to-end content distribution network system and distribution method based on distributed election |
CN111372100B (en) * | 2020-04-21 | 2023-07-14 | 白杨 | Distributed election-based end-to-end content distribution network system and distribution method |
CN113504933A (en) * | 2021-05-28 | 2021-10-15 | 曲阜师范大学 | Workshop machine updating method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108228082B (en) | Storage system and method for storage control | |
US8718065B2 (en) | Transmission using multiple physical interface | |
CN103986771A (en) | High-availability cluster management method independent of shared storage | |
CN106648896B (en) | Method for dual-core sharing of output peripheral by Zynq chip under heterogeneous-name multiprocessing mode | |
CN101808137B (en) | Data transmission method, device and system | |
CN109376197B (en) | Data synchronization method, server and computer storage medium | |
CN103793291B (en) | Distributed data copying method and device | |
CN103051687B (en) | A kind of system and method that applied business is deployed to cloud storage virtual machine | |
CN108733506A (en) | Method, equipment and the computer-readable medium synchronized for data | |
CN106446159B (en) | A kind of method of storage file, the first virtual machine and name node | |
CN102129434B (en) | Method and system for reading and writing separation database | |
CN102521038A (en) | Virtual machine migration method and device based on distributed file system | |
CN103780638A (en) | Data synchronization method and system | |
CN101299757A (en) | Data sharing method and communication system as well as correlation equipment | |
CN103891223A (en) | Methods and apparatus for transporting data through network tunnels | |
CN114546913B (en) | Method and device for high-speed data interaction between multiple hosts based on PCIE interface | |
RU2653254C1 (en) | Method, node and system for managing data for database cluster | |
CN106095622A (en) | Data back up method and device | |
CN104750428A (en) | Block storage access and gateway module, storage system and method, and content delivery apparatus | |
CN105740248B (en) | A kind of method of data synchronization, apparatus and system | |
US20200012438A1 (en) | Method and apparatus for caching data | |
CN106919473A (en) | A kind of data disaster recovery and backup systems and method for processing business | |
CN102790801B (en) | Asynchronous remote copy system and method for maintaining data consistency thereof | |
CN101790230A (en) | Precision time protocol node, time stamp operation method and time synchronization system | |
US8595452B1 (en) | System and method for streaming data conversion and replication |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20140813 |