CN101873237B - Method and system for dynamically maintaining member servers in cluster - Google Patents
Method and system for dynamically maintaining member servers in cluster Download PDFInfo
- Publication number
- CN101873237B CN101873237B CN2010102079927A CN201010207992A CN101873237B CN 101873237 B CN101873237 B CN 101873237B CN 2010102079927 A CN2010102079927 A CN 2010102079927A CN 201010207992 A CN201010207992 A CN 201010207992A CN 101873237 B CN101873237 B CN 101873237B
- Authority
- CN
- China
- Prior art keywords
- cluster
- member server
- controller
- server
- topology information
- 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
Abstract
The invention provides a method for dynamically maintaining member servers in a cluster. A cluster controller monitors states of all the member servers in the cluster on public address and port; the cluster controller revises cluster topological structure information when discovering that the state of any one member server in the cluster changes; and before establishing communication connection with other member servers, a member server in the cluster obtains addresses and ports of other member servers by inquiring the cluster topological structure information from the cluster controller so as to establish communication channels with other member servers. Based on the method, the cluster members are managed and maintained by introducing the cluster controller and through the traditional mechanism of register-discovery-binding, and a system has the advantages of simple architecture, convenient implementation and maintenance as well as good monitor and control on the stability. The invention also provides a system for dynamically maintaining the member servers in the cluster.
Description
Technical field
The present invention relates to the maintenance of cluster, relate more specifically to the method and system of the member server in a kind of Dynamic Maintenance cluster.
Background technology
Cloud computing and intelligent cluster have become one of current most important computing architecture.Be that cloud or the cluster in the intelligent cluster in the cloud computing all is to be made up of some members with operational capability; These members' computing capability possibly provided by independent physical equipment, also possibly be based on virtual machine, technical operation some logical nodes on physical equipment such as virtual.The characteristics of cloud computing and intelligent cluster are the dynamics of height, and one of them most important parts is that cluster member increases dynamically, deletes.Traditional mode, cluster all are the configurations through static state, carry out member's distribution, and the formation of cluster is static, increase if desired, removing members need be through reconfiguring cluster, the change underaction of cluster member.Present more popular dynamic cluster maintenance scheme based on multicast and P-2-P technology, the communication technology very complicated that relates to, the complexity of maintenance and configuration is higher, is difficult to assurance for large-scale cluster topology stability.The static configuration that cluster constitutes can not satisfy the requirement of highly dynamic computing environment, and too complicated through multicast and the point-to-point dynamic cluster framework of technical maintenance that waits, and therefore needs more simply dynamic cluster maintenance mode of a kind of framework.
Summary of the invention
In view of the above problems; The present invention proposes the method for the member server in a kind of Dynamic Maintenance cluster; This method is carried out administering and maintaining of cluster member through introducing cluster controller through the mechanism of traditional registration → discovery → binding, and its framework is simple; Implement easy to maintenancely, stability is hard to bear to good monitoring and control.The invention allows for the system of the member server in a kind of Dynamic Maintenance cluster.
According to an aspect of the present invention, the method for the member server in a kind of Dynamic Maintenance cluster is provided, has comprised: cluster controller is monitored the state of all member servers in the cluster on a disclosed address and port; When the state of the arbitrary member server of cluster controller in finding cluster changes, revise the cluster topology information; And before the member server in the cluster will establish a communications link with other member servers, through obtain the address and the port of other member servers to cluster controller inquiry cluster topology information, to set up communication port with other member servers.
Preferably, cluster controller after revising the cluster topology information, the event notice that all the member server signalling of bouquet members in cluster change.
Preferably, the member server in the cluster when starting or stoping to cluster controller signalling of bouquet incident; And cluster controller revises the cluster topology information according to the cluster incident, and the event notice that changes of all the member server signalling of bouquet members in cluster.
Preferably, the member server in the cluster upgrades its local cluster topology information after receiving event notice.
Preferably, the member server in the cluster selects communication port to communicate according to the cluster topology information of this locality when the cluster controller fault.
Preferably, member server in the cluster is regularly got in touch with the cluster controller of fault, in case cluster controller during from fault recovery, just the information with member server is registered to cluster controller.
Preferably, cluster controller initiatively connects each member server in the cluster according to the cluster topology information of its record from fault recovery the time, and after successful connection, the information of each member server is registered to cluster controller.
Preferably, the cluster controller state variation of keeping watch on member server through the cluster management address and the port of member server.
According to a further aspect in the invention; The system of the member server in a kind of Dynamic Maintenance cluster is provided; Comprise: cluster controller; On a disclosed address and port, monitor the state of all member servers in the said cluster, when the state of the arbitrary member server in finding said cluster changes, revise the cluster topology information; And at least one member server; Be arranged in said cluster; Before arbitrary member server will establish a communications link with other member servers; Through inquire about address and the port that said cluster topology information obtains said other member servers to said cluster controller, to set up communication port with said other member servers.
Preferably, cluster controller also is used for after revising said cluster topology information, the event notice that all the member server signalling of bouquet members in said cluster change.
Preferably, the arbitrary member server in the cluster when starting or stoping to cluster controller signalling of bouquet incident; And cluster controller revises the cluster topology information according to said cluster incident, and the event notice that changes of all the member server signalling of bouquet members in cluster.
Preferably, the member server in the cluster upgrades its local cluster topology information after receiving event notice.
Preferably, the member server in the cluster selects communication port to communicate according to the cluster topology information of this locality when the cluster controller fault.
Preferably, member server in the cluster is regularly got in touch with the cluster controller of fault, in case cluster controller during from fault recovery, just the information with member server is registered to cluster controller.
Preferably, cluster controller initiatively connects each member server in the cluster according to the cluster topology information of its record from fault recovery the time, and after successful connection, the information of each member server is registered to cluster controller.
Preferably, the cluster controller state variation of keeping watch on member server through the cluster management address and the port of member server.
Through adopting the method and system of the member server in the Dynamic Maintenance cluster proposed by the invention, can carry out Dynamic Maintenance to cluster, simultaneously because this dynamic member's maintenance is transparent to using, thereby the availability that has guaranteed system is unaffected.This technology makes aspects such as the autgmentability, reliability, maintainability, stability of system obtain comprehensive raising, has solved too complicated problems of available technology adopting static configuration or Dynamic Maintenance.
Description of drawings
Fig. 1 shows the flow chart according to the method for the member server in the Dynamic Maintenance cluster of the present invention;
Fig. 2 shows the sketch map of the system of the member server in the Dynamic Maintenance cluster according to an embodiment of the invention;
Fig. 3 shows the sketch map of the communication between the member server in the cluster according to an embodiment of the invention;
Fig. 4 shows the sketch map of the communication between the member server under the cluster controller failure condition according to an embodiment of the invention.
Embodiment
Below will be described in detail with reference to the attached drawings specific embodiment of the present invention.
Fig. 1 shows the flow chart according to the method for the member server in the Dynamic Maintenance cluster of the present invention, and is as shown in Figure 1, and this method may further comprise the steps:
Step S102, cluster controller monitor the state of all member servers in the cluster on a disclosed address and port;
Step S104 when the state of the arbitrary member server of cluster controller in finding cluster changes, revises the cluster topology information; And
Step S106; Before member server in the cluster will establish a communications link with other member servers; Through obtain the address and the port of other member servers to cluster controller inquiry cluster topology information, to set up communication port with other member servers.
Preferably, cluster controller after revising the cluster topology information, the event notice that all the member server signalling of bouquet members in cluster change.
Preferably, the member server in the cluster when starting or stoping to cluster controller signalling of bouquet incident; And cluster controller revises the cluster topology information according to the cluster incident, and the event notice that changes of all the member server signalling of bouquet members in cluster.
Preferably, the member server in the cluster upgrades its local cluster topology information after receiving event notice.
Preferably, the member server in the cluster selects communication port to communicate according to the cluster topology information of this locality when the cluster controller fault.
Preferably, member server in the cluster is regularly got in touch with the cluster controller of fault, in case cluster controller during from fault recovery, just the information with member server is registered to cluster controller.
Preferably, cluster controller initiatively connects each member server in the cluster according to the cluster topology information of its record from fault recovery the time, and after successful connection, the information of each member server is registered to cluster controller.
Preferably, the cluster controller state variation of keeping watch on member server through the cluster management address and the port of member server.
This method is carried out administering and maintaining of cluster member through introducing the cluster registrar through the mechanism of traditional registration → discoverys → binding, and its framework is simple, implements easy to maintenancely, and stability is hard to bear to arrive good monitoring and control.
Fig. 2 shows the sketch map of the system of the member server in the Dynamic Maintenance cluster according to an embodiment of the invention.As shown in Figure 2, in this system, introduced cluster controller 202, it is in charge of, controls and safeguard whole cluster, each server that adds this cluster all is called as member server 204.In the example shown in Fig. 2, for succinctly, only show a member server, it should be appreciated by those skilled in the art can have a plurality of member servers in the whole cluster at this.Cluster controller 202 is only safeguarded the structural information that cluster is relevant and the service relevant with these information is provided as a main control server.The structure of cluster is dynamic, and the member server 204 that constitutes cluster can add and delete dynamically, and when interpolation and removing members server, its dynamic course of work is following:
1. cluster controller 202 is monitored on a disclosed address and port;
2. member server 204, starting/add/leave the cluster incident to cluster controller 202 transmissions when stopping;
3. cluster controller 202, according to this incident, revise the cluster topology information in the cluster controller;
4. cluster controller 202 is to the event notice of cluster member server 204 signalling of bouquet members variation;
5. each member server 204 is updated to its local up-to-date cluster topology information to its cluster of seeing through this incident.
In general, member server comprises to the log-on message that cluster controller sends:
The id of member server;
The address of service of member server and port;
The address of carrying out cluster management and the port of member server;
The information that other possibly used in trunking communication is like member's grouping, load capacity etc.
After cluster topology was set up, cluster controller 202 was kept watch on the variation of cluster member through the cluster management address and the port of member server 204, and the update notifications cluster information.
Communicate by letter if desired between each member server in the cluster, its course of work is generally described like Fig. 3.Two member servers 302 and 304 have been shown among Fig. 3, but have it should be appreciated by those skilled in the art, this only is exemplary, can have a plurality of member servers.
1. member server 302 hopes to initiate an inter-cluster communication, this communication possibly be point-to-point also possibly be one-to-many;
2. member server 302 is to the address of service and the port of cluster controller 306 query aim servers, and the hypothetical target server is a member server 304;
3. member server 302 communicates through setting up communication port with the address of service of member server 304 and port.
Above situation all is to consider under the situation of cluster controller and member server operate as normal, will describe the situation that cluster controller fault, member server are unusual and cluster controller recovers below respectively.
Cluster controller lost efficacy
If cluster controller fault or maintenance; Then the cluster topological structure can not obtain upgrading, and new member server can not add, and member server leaves and can not obtain upgrading; But the method for deduction according to the present invention, it can not have influence on the communication between the member server that has constituted cluster.Fig. 4 shows the sketch map of the communication between the member server under the cluster controller failure condition according to an embodiment of the invention, specifically describes as follows:
1. each member server 402 and 404 has all been safeguarded the cluster topology information an of this locality, and in cluster controller 406 faults or when safeguarding, the understanding of member server 402 and 404 pairs of cluster topological structures is from its local cluster topology information;
2. member server 402 and 404 is periodically checked the activity of cluster controller 406, if can not be communicated with, it can constantly attempt communicating by letter with cluster controller 406, with after cluster controller 406 activities, recovers cluster topology;
3. if cluster controller 406 activities, and after the cluster topology recovery, it can notify member server 402 and 404 to upgrade its local cluster topology informations;
4. when cluster controller 406 inertias, communicate by letter between the cluster if desired, member server 402 can be selected suitable mode and member server 404 mutual communication according to its local cluster topology information.
Member server is unusual
Consider unusual the leaving of cluster member server possibility, therefore, in order to safeguard cluster topology, carry out following operation between cluster controller and the member server:
1. cluster controller is checked the life state of each member server termly;
2. when cluster controller finds that certain member server is no longer movable, the member server of notifying other also to survive through cluster member server altering event;
3. member server according to this event notice, upgrades its local cluster topology information.
Cluster controller recovers
Cluster controller was from losing efficacy to recovering the time, and the cluster topological structure can be restored automatically, and its course of work is described as follows:
1. cluster controller can initiatively connect each member server and check recovery cluster topological structure according to the cluster topological structure of historical record, if on connecting, registers its information and gets into cluster topology;
2. member server is constantly attempted getting in touch with the cluster controller that lost efficacy, in case get in touch with, is registered to cluster controller to its information.
To sum up, through technical scheme provided by the invention, more stable through the cluster topology of introducing cluster controller foundation, available as long as cluster controller keeps, the communication between the cluster member, the topology information of cluster can both obtain good management.The communications framework of the cluster on the basis of such cluster topology management has been realized communication point-to-point between the cluster, one-to-many, can not occur the communication mode that can not expect owing to the difference that each member is familiar with the cluster topological structure.Not only overcome the shortcoming of underaction in the traditional static configuration mode, also solved technological very complicated in the conventional dynamic mode, the problem that stability is difficult to guarantee.
The above is merely the preferred embodiments of the present invention, is not limited to the present invention, and for a person skilled in the art, the present invention can have various changes and variation.All within spirit of the present invention and principle, any modification of being done, be equal to replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (12)
1. the method for the member server in the Dynamic Maintenance cluster is characterized in that, comprising:
Cluster controller is monitored the state of all member servers in the said cluster on a disclosed address and port;
When the state of the arbitrary member server of said cluster controller in finding said cluster changes, revise the cluster topology information; And
Before member server in the said cluster will establish a communications link with other member servers; Through inquire about address and the port that said cluster topology information obtains said other member servers to said cluster controller, to set up communication port with said other member servers.
2. method according to claim 1 is characterized in that, further comprises:
Said cluster controller after revising said cluster topology information, the event notice that all the member server signalling of bouquet members in said cluster change.
3. method according to claim 1 and 2 is characterized in that, further comprises:
Member server in the said cluster when starting or stoping to said cluster controller signalling of bouquet incident; And
Said cluster controller is revised the cluster topology information according to said cluster incident, and all member servers in said cluster send the event notice that said cluster member changes.
4. method according to claim 3 is characterized in that, further comprises:
Member server in the said cluster upgrades its local cluster topology information after receiving said event notice.
5. method according to claim 4 is characterized in that, the member server in the said cluster selects communication port to communicate according to the cluster topology information of said this locality when said cluster controller fault; And
Member server in the said cluster is regularly got in touch with the said cluster controller of fault, in case said cluster controller during from fault recovery, just the information with said member server is registered to said cluster controller.
6. method according to claim 5; It is characterized in that; Said cluster controller is from fault recovery the time; Said cluster topology information according to its record initiatively connects each member server in the said cluster, and after successful connection, the information of each member server is registered to said cluster controller.
7. method according to claim 1 is characterized in that, the state variation that said cluster controller is kept watch on said member server through the cluster management address and the port of said member server.
8. the system of the member server in the Dynamic Maintenance cluster is characterized in that, comprising:
Cluster controller, the state of all member servers on a disclosed address and port in the said cluster of monitoring when the state of the arbitrary member server in finding said cluster changes, is revised the cluster topology information; And
At least one member server; Be arranged in said cluster; Before arbitrary member server will establish a communications link with other member servers; Through inquire about address and the port that said cluster topology information obtains said other member servers to said cluster controller, to set up communication port with said other member servers.
9. system according to claim 8 is characterized in that, said cluster controller also is used for after revising said cluster topology information, the event notice that all the member server signalling of bouquet members in said cluster change.
10. according to Claim 8 or 9 described systems, it is characterized in that,
Arbitrary member server in the said cluster when starting or stoping to said cluster controller signalling of bouquet incident;
Said cluster controller is revised the cluster topology information according to said cluster incident, and all member servers in said cluster send the event notice that said cluster member changes; And
Member server in the said cluster upgrades its local cluster topology information after receiving said event notice.
11. according to Claim 8 or 9 described systems, it is characterized in that the member server in the said cluster selects communication port to communicate according to the cluster topology information of this locality when said cluster controller fault;
Member server in the said cluster is regularly got in touch with the said cluster controller of fault, in case said cluster controller during from fault recovery, just the information with said member server is registered to said cluster controller; And
Said cluster controller initiatively connects each member server in the said cluster according to the said cluster topology information of its record from fault recovery the time, and after successful connection, the information of each member server is registered to said cluster controller.
12. according to Claim 8 or 9 described systems, it is characterized in that the state variation that said cluster controller is kept watch on said member server through the cluster management address and the port of said member server.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010102079927A CN101873237B (en) | 2010-06-13 | 2010-06-13 | Method and system for dynamically maintaining member servers in cluster |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010102079927A CN101873237B (en) | 2010-06-13 | 2010-06-13 | Method and system for dynamically maintaining member servers in cluster |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101873237A CN101873237A (en) | 2010-10-27 |
| CN101873237B true CN101873237B (en) | 2012-02-22 |
Family
ID=42997913
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2010102079927A Active CN101873237B (en) | 2010-06-13 | 2010-06-13 | Method and system for dynamically maintaining member servers in cluster |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101873237B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107786373A (en) * | 2017-10-13 | 2018-03-09 | 广州供电局有限公司 | Generate method, apparatus, storage medium and the computer equipment of topology server relation |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102724104B (en) * | 2011-03-30 | 2015-04-29 | 北京东方通科技股份有限公司 | Apparatus and method for automatically configuring Java EE application cluster |
| CN102571960A (en) * | 2012-01-12 | 2012-07-11 | 浪潮(北京)电子信息产业有限公司 | Method and device for monitoring high-availability cluster state |
| CN102752146B (en) * | 2012-07-13 | 2015-07-15 | 中国科学院微电子研究所 | Cluster topological graph generation method and server |
| CN102769675B (en) * | 2012-08-13 | 2015-04-22 | 广州杰赛科技股份有限公司 | Method used for keeping host resource stable and based on cloud computing platform |
| US10454997B2 (en) | 2012-09-07 | 2019-10-22 | Avigilon Corporation | Distributed physical security system |
| CN104301158B (en) * | 2014-11-06 | 2018-01-16 | 中国移动通信集团广东有限公司 | Group management and system based on GMS frameworks |
| CN105743671B (en) * | 2014-12-10 | 2020-07-14 | 华为技术有限公司 | Capacity expansion method, controller and system |
| CN105204910B (en) * | 2015-10-15 | 2018-11-06 | 北京乐动卓越科技有限公司 | A kind of hot update method of script and system |
| CN110278133B (en) * | 2019-07-31 | 2021-08-13 | 中国工商银行股份有限公司 | Checking method, device, computing equipment and medium executed by server |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7636917B2 (en) * | 2003-06-30 | 2009-12-22 | Microsoft Corporation | Network load balancing with host status information |
| US8291108B2 (en) * | 2007-03-12 | 2012-10-16 | Citrix Systems, Inc. | Systems and methods for load balancing based on user selected metrics |
| CN101415023A (en) * | 2008-08-08 | 2009-04-22 | 西安瑞友信息技术资讯有限公司 | Intelligent dynamic load equilibrium technology for server |
-
2010
- 2010-06-13 CN CN2010102079927A patent/CN101873237B/en active Active
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107786373A (en) * | 2017-10-13 | 2018-03-09 | 广州供电局有限公司 | Generate method, apparatus, storage medium and the computer equipment of topology server relation |
| CN107786373B (en) * | 2017-10-13 | 2021-08-31 | 广东电网有限责任公司广州供电局 | Method and device for generating server topological relation, storage medium and computer equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101873237A (en) | 2010-10-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101873237B (en) | Method and system for dynamically maintaining member servers in cluster | |
| US9806983B2 (en) | System and method for control flow management in software defined networks | |
| US7904052B2 (en) | Sensor net management method | |
| CN103036736B (en) | A kind of configuration equipment monitoring system based on data source and method | |
| US20140018002A1 (en) | Network, master, hub and method for providing a bluetooth infrastructure | |
| CN103609167A (en) | System and method for session restoration at Geo-redundant gateways | |
| CN105227385B (en) | A kind of method and system of troubleshooting | |
| CN103888277A (en) | Gateway disaster recovery backup method, apparatus and system | |
| CN103475722A (en) | Implement system for business collaboration platform | |
| CN108282540A (en) | A kind of subway monitoring system and its monitoring method | |
| CN106027270B (en) | On-demand power management in a networked computing environment | |
| CN107925619A (en) | System and method for disposing the loss of link in network | |
| CN102215123A (en) | Multi-ring-network-topology-structure-based large-scale trunking system | |
| CN103036719A (en) | Cross-regional service disaster method and device based on main cluster servers | |
| CN112003716A (en) | Data center dual-activity implementation method | |
| CN104468633A (en) | SDN southing security proxy product | |
| CN106357723A (en) | Synchronous system and method for multi-cluster information caching based on cloud host | |
| US8560874B2 (en) | Method and system for power saving management | |
| CN102316018B (en) | Method for updating routing table in network node equipment and network node equipment | |
| CN104125079A (en) | Method and device for determining double-device hot-backup configuration information | |
| CN106411574B (en) | Management control method and device | |
| CN113824801B (en) | Intelligent integration terminal unified access management component system | |
| CN102104490B (en) | Comprehensive network management system for railway communication network | |
| CN107222520B (en) | Distributed system based on directional diffusion algorithm | |
| CN102185720A (en) | North notification management interface device and management method thereof |
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 | ||
| C56 | Change in the name or address of the patentee |
Owner name: YONYOU NETWORK TECHNOLOGY CO., LTD. Free format text: FORMER NAME: UFIDA SOFTWARE CO., LTD. |
|
| CP03 | Change of name, title or address |
Address after: 100094 Haidian District North Road, Beijing, No. 68 Patentee after: Yonyou Network Technology Co., Ltd. Address before: 100094 Beijing city Haidian District North Road No. 68, UFIDA Software Park Patentee before: UFIDA Software Co., Ltd. |