CN109995842A

CN109995842A - A kind of group technology and device for distributed server cluster

Info

Publication number: CN109995842A
Application number: CN201810002956.3A
Authority: CN
Inventors: 陶涛
Original assignee: Beijing Qihoo Technology Co Ltd
Current assignee: Beijing Qihoo Technology Co Ltd
Priority date: 2018-01-02
Filing date: 2018-01-02
Publication date: 2019-07-09
Anticipated expiration: 2038-01-02
Also published as: CN109995842B

Abstract

The present invention provides a kind of group technology and device for distributed server cluster, the distributed server cluster includes multiple server farms, method includes: the grouping strategy for obtaining distributed server cluster, and grouping strategy is pre-stored under the corresponding node of distributed server cluster；Obtain the distributed lock of the distributed server cluster；Distributed lock is such as got, according to the server farm that grouping strategy determination should be added, itself will be registered in corresponding server farm；When cluster needs to be grouped, server to be registered obtains grouping strategy to node, when itself gets distributed lock, the server farm that can should be added according to the grouping strategy determination, it itself will be registered in corresponding server farm, distributed server cluster can be thus divided into multiple server farms, each server farm provides different business services, improves the efficiency of distributed server cluster.

Description

A kind of group technology and device for distributed server cluster

Technical field

The present invention relates to network technique field more particularly to a kind of group technologies and dress for distributed server cluster It sets.

Background technique

With the development of internet, the scale of website application constantly expands, and conventional vertical applications framework can not cope with, Distributed Services framework is imperative.

When establishing Distributed Services, realized generally by a ZooKeeper cluster is established.In the prior art, Due to that cannot be grouped to distributed server cluster, a distributed server cluster can only provide a kind of service, Cause distributed server cluster that expected efficiency is not achieved.

Summary of the invention

In view of the above problems, propose the present invention in order to provide a kind of group technology for distributed server cluster and Device must not be grouped distributed server cluster with solution, cause a distributed server cluster that can only provide one The technical issues of planting service function, and then causing distributed server cluster that expected efficiency is not achieved.

The first aspect of the present invention provides a kind of group technology for distributed server cluster, the distributed clothes Business device cluster includes multiple server farms, which comprises

The grouping strategy of the distributed server cluster is obtained, the grouping strategy is pre-stored in the distributed clothes It is engaged under the corresponding node of device cluster；

The distributed lock of distributed server cluster described in request；

The distributed lock is such as got, then the server farm that should be added according to the grouping strategy determination, by itself It is registered in corresponding server farm.

Optionally, the grouping strategy includes: that average packet rule and the distributed server cluster are divided into M clothes The information of Wu Qi group；The server farm that should be added according to the grouping strategy determination, itself will be registered to accordingly In server farm, comprising:

Being determined according to the average packet rule itself should be registered to the current registered least service of server total quantity In device group；

Know the registered server total quantity that each server farm includes in presently described M server farm, leads to It crosses to compare and determines the current registered least server farm of server total quantity；

It itself will be registered in the current registered least server farm of server total quantity of the determination.

Optionally, the grouping strategy includes: that weight rule of classification, the distributed server cluster are divided into M service The weighted value of the information of device group and each server farm；The server that should be added according to the grouping strategy determination Group itself will be registered in corresponding server farm, comprising:

Being determined according to the weight rule of classification, which itself should be registered to current registered server total quantity, is not up to the group In the server farm of preset quantity, the preset quantity is according to the weighted value and the distributed server collection of server farm The destination server total quantity of group determines；

Know the registered server total quantity that each server farm includes in presently described M server farm, leads to It crosses and is compared with corresponding preset quantity, determine that current registered server total quantity is also not up to this group of preset quantity Server farm；

The current registered server total quantity for itself being registered to the determination is also not up to the clothes of this group of preset quantity In Wu Qi group.

Optionally, the grouping strategy includes: that load rule of classification and the distributed server cluster are divided into M clothes The information of Wu Qi group；The server farm that should be added according to the grouping strategy determination, itself will be registered to accordingly In server farm, comprising:

Being determined according to the load rule of classification itself should be registered in the highest server farm of current operating load；

The operating load for knowing each server farm in presently described M server farm, determines to work as by comparing The preceding highest server farm of operating load；

It itself will be registered in the highest server farm of current operating load of the determination.

Optionally, the grouping strategy includes: that designated packet rule, the distributed server cluster are divided into M service The information of device group and in each server farm group identification；The clothes that should be added according to the grouping strategy determination Wu Qi group itself will be registered in corresponding server farm, comprising:

Being determined according to the designated packet rule, which itself should be registered to designated group, identifies in corresponding server farm；

The group identification for knowing each server farm in designated group mark and presently described M server farm, leads to It crosses comparison and determines that the designated group identifies corresponding server farm；

The designated group for itself being registered to the determination is identified in corresponding server farm.

Optionally, each server farm in the distributed server cluster is for providing different business services.

The second aspect of the present invention provides a kind of apparatus for grouping for distributed server cluster, the distributed clothes Business device cluster includes multiple server farms, and described device includes:

Tactful acquiring unit, for obtaining the grouping strategy of the distributed server cluster, the grouping strategy is preparatory It is stored under the corresponding node of the distributed server cluster；

Distributed lock acquiring unit, the distributed lock for distributed server cluster described in request；

Registering unit, for when determining that the distributed lock acquiring unit gets the distributed lock, according to described The server farm that grouping strategy determination should be added itself will be registered in corresponding server farm.

Optionally, the grouping strategy includes: that average packet rule and the distributed server cluster are divided into M clothes The information of Wu Qi group；The registering unit includes:

First determines subelement, for determining that itself should be registered to current registered service according to the average packet rule In the least server farm of device total quantity；

Second determines subelement, and for knowing, each server farm includes in presently described M server farm Registrar total quantity determines the current registered least server farm of server total quantity by comparing；

Registration subunit, for the current registered least service of server total quantity of the determination itself will to be registered to In device group.

Optionally, the grouping strategy includes: that weight rule of classification, the distributed server cluster are divided into M service The weighted value of the information of device group and each server farm；The registering unit includes:

First determines subelement, for determining that itself should be registered to current registered service according to the weight rule of classification Device total quantity is not up in the server farm of this group of preset quantity, the preset quantity according to the weighted value of server farm and The destination server total quantity of the distributed server cluster determines；

Second determines subelement, and for knowing, each server farm includes in presently described M server farm Registrar total quantity determines current registered server total quantity also by being compared with corresponding preset quantity The server farm of not up to this group preset quantity；

Registration subunit, for the current registered server total quantity for itself being registered to the determination also not up to should In the server farm of group preset quantity.

Optionally, the grouping strategy includes: that load rule of classification and the distributed server cluster are divided into M clothes The information of Wu Qi group；The registering unit includes:

First determines subelement, for determining that itself should be registered to current operating load most according to the load rule of classification In high server farm；

Second determines subelement, and the operation for knowing each server farm in presently described M server farm is negative Lotus determines the highest server farm of current operating load by comparing；

Registration subunit, for itself will be registered in the highest server farm of current operating load of the determination.

Optionally, the grouping strategy includes: that designated packet rule, the distributed server cluster are divided into M service The information of device group and in each server farm group identification；The registering unit includes:

First determines subelement, itself should be registered to designated group mark pair for determining according to the designated packet rule In the server farm answered；

Second determines subelement, for knowing each service in designated group mark and presently described M server farm The group identification of device group determines that the designated group identifies corresponding server farm by comparison；

Registration subunit, for identifying the designated group for itself being registered to the determination in corresponding server farm.

The third aspect of the present invention provides a kind of group technology for distributed server cluster, the distributed clothes Business device cluster includes multiple server farms, which comprises

Receive the acquisition request that server to be registered is sent；

Preset grouping strategy is sent to the server to be registered according to the acquisition request, for the clothes to be registered Device be engaged in when getting distributed lock, itself is registered to by the server farm that can should be added according to the grouping strategy determination In corresponding server farm.

The fourth aspect of the present invention provides a kind of distributed server cluster, including multiple server farms, clothes to be registered Business device and a node；Wherein, the node is used to save the grouping strategy of the distributed server cluster, the node pair Answer the distributed server cluster；The server to be registered includes:

Tactful acquiring unit, for obtaining the grouping strategy of the distributed server cluster from the node；

The fifth aspect of the present invention provides a kind of computer readable storage medium, is stored thereon with computer program, the journey The method as described in any implementation of first aspect is able to carry out when sequence is executed by processor.

The sixth aspect of the present invention provides a kind of computer equipment for distributed server cluster, comprising:

At least one processor；And

At least one processor being connect with the processor communication, wherein

The memory is stored with the program instruction that can be executed by the processor, and the processor calls described program to refer to Enable the method for being able to carry out any implementation such as first aspect.

The technical solution provided in the embodiment of the present application, has at least the following technical effects or advantages:

The present invention provides a kind of group technology and device for distributed server cluster, the distributed server collection Group includes multiple server farms, which comprises obtains the grouping strategy of the distributed server cluster, the grouping Strategy is pre-stored under the corresponding node of the distributed server cluster；Distributed server cluster described in request Distributed lock；The distributed lock is such as got, then the server farm that should be added according to the grouping strategy determination, by itself It is registered in corresponding server farm；In this way, preset grouping strategy is stored under node, when cluster needs are grouped When, server to be registered obtains grouping strategy to node can be according to described point in the case where itself gets distributed lock The server farm that group policy determination should be added itself will be registered in corresponding server farm, can will be thus distributed Formula server cluster is divided into multiple server farms, and then can make each server zone by configuring each server farm Group provides different business services, to improve the efficiency of distributed server cluster.

The above description is only an overview of the technical scheme of the present invention, in order to better understand the technical means of the present invention, And it can be implemented in accordance with the contents of the specification, and in order to allow above and other objects of the present invention, feature and advantage can It is clearer and more comprehensible, the followings are specific embodiments of the present invention.

Detailed description of the invention

By reading the following detailed description of the preferred embodiment, various other advantages and benefits are common for this field Technical staff will become clear.The drawings are only for the purpose of illustrating a preferred embodiment, and is not considered as to the present invention Limitation.And throughout the drawings, the same reference numbers will be used to refer to the same parts.In the accompanying drawings:

The group technology process for distributed server cluster that Fig. 1 shows the offer of the embodiment of the present invention one is illustrated Figure；

The distributed lock provided Fig. 2 shows the embodiment of the present invention one is stored in the corresponding node of distributed server cluster Under schematic diagram；

The distributed lock that Fig. 3 shows the offer of the embodiment of the present invention one is stored in the node except distributed server cluster Under schematic diagram；

Fig. 4 shows the apparatus for grouping structural representation provided by Embodiment 2 of the present invention for distributed server cluster Figure；

Fig. 5 shows the structural schematic diagram of registering unit provided by Embodiment 2 of the present invention；

The group technology process for distributed server cluster that Fig. 6 shows the offer of the embodiment of the present invention three is illustrated Figure；

Fig. 7 shows the structural schematic diagram of the distributed server cluster of the offer of the embodiment of the present invention four.

Specific embodiment

Exemplary embodiments of the present disclosure are described in more detail below with reference to accompanying drawings.Although showing the disclosure in attached drawing Exemplary embodiment, it being understood, however, that may be realized in various forms the disclosure without should be by embodiments set forth here It is limited.On the contrary, these embodiments are provided to facilitate a more thoroughly understanding of the present invention, and can be by the scope of the present disclosure It is fully disclosed to those skilled in the art.

Embodiment one

The embodiment of the invention provides a kind of group technology for distributed server cluster, the method is from wait infuse Volume server is that the angle of executing subject is write, as shown in Figure 1, which comprises

S110, obtains the grouping strategy of the distributed server cluster, and the grouping strategy is pre-stored in described point Under the corresponding node of cloth server cluster；

When needing to be grouped distributed server cluster automatically, need to obtain the distributed server cluster Grouping strategy described in grouping strategy is pre-stored under the corresponding node of the distributed server cluster.The Distributed Services Device cluster includes multiple server farms, and each server farm is for providing different business services；Include under the node Multiple child nodes, each child node correspond to a server farm, again include multiple dynamic nodes, Mei Gedong under each child node Corresponding state node is (survival) server to have succeeded in registration.

Specifically, when distributed server cluster starts, server to be registered sends request point to the node Grouping strategy is issued in server to be registered by the first solicited message of group policy, the node according to the solicited message.

S111, the distributed lock of distributed server cluster described in request；

Here, the grouping strategy and distributed lock can be stored under the same node, can also be stored in different Under node；The i.e. described distributed lock can be stored under the corresponding node of distributed server cluster, can also be stored in distribution Under node except formula server cluster.

, there are two server farms, respectively average mark under A node Group 1 and average packet 2；As can be seen that distributed lock lock and average packet 1 and average packet 2 are sane level relationships, therefore it is distributed Formula lock lock is stored under the corresponding node of distributed server cluster.

Likewise, referring to Fig. 3, such as server cluster corresponding A node, there are two server farms under A node, respectively Average packet 1 and average packet 2；As can be seen that distributed lock lockA and A node are sane level relationships, therefore distributed lock LockA is stored under the node except distributed server cluster.

When distributed lock is stored under the node outside distributed server cluster, the distributed server collection is obtained The distributed lock of group, which just refers to, obtains the corresponding distributed lock of distributed server cluster.Wherein, the distributed lock be can be with Distributed lock based on ZK is also possible to the distributed lock based on redis database.

It should be noted that father node can be first nodes, or multistage node, certain distributed server collection Father node can also not be needed in group, does not have father node in the distributed server cluster in the present embodiment.

It is illustrated for saving in the same node, server to be registered Distributed Services described in request It is specific as follows when the distributed lock of device cluster:

After server to be registered gets grouping strategy, it can continue to send request distributed lock to the node Second solicited message.Requesting the specific implementation of distributed lock, there are many modes, for example, the node is believed according to second request Server to be registered is added the first of the distributed lock and obtains queue by breath, is had in second solicited message described wait infuse The precedence information of volume server.

After server to be registered is added into the first acquisition queue, server to be registered is to described in the node request First obtains the priority of the Servers-all in queue in addition to itself；Whether judgement itself the priority is described the One obtains the highest priority in queue, if the highest priority in the first acquisition queue, then is defaulted as directly acquiring The distributed lock.

When the priority for determining itself is not the highest priority in the first acquisition queue, further includes: call monitoring The mechanism monitors adjacent priority corresponding server higher than own priority；Determining the corresponding service of the adjacent priority In the case that device discharges the distributed lock, the distributed lock is directly acquired.

As a kind of optional embodiment, server to be registered is added to first obtain in queue after, node gets the One obtains the priority of Servers-all in queue, when node determines that the priority of the server to be registered has been in queue Highest priority server when, distributed lock is distributed to server to be registered.

After server to be registered gets distribution formula lock, the server transmission point in waiting in queue is obtained to first Cloth locks occupied information.

As a kind of optional embodiment, the distribution of server to be registered distributed server cluster described in request Formula lock, comprising:

After server to be registered gets grouping strategy, it can continue to send request distributed lock to the node Server to be registered is added the second of the distributed lock according to second solicited message by third solicited message, the node Obtain queue.

After server to be registered is added into the second acquisition queue, server to be registered is to second described in node request It obtains the Servers-all in queue in addition to itself and obtains the serial number in queue second；Whether the serial number for judging itself is institute The smallest serial number in the second acquisition queue is stated, if the smallest serial number, is then considered as and directly acquires the distributed lock.

Similarly, when the serial number for determining itself is not the smallest serial number in the second acquisition queue, further includes: call prison Depending on the mechanism monitors adjacent server smaller than itself serial number；Determining that the adjacent server discharges the distributed lock In the case of, directly acquire the distributed lock.

As a kind of optional embodiment, server to be registered is added to second obtain in queue after, node gets the One obtain queue in Servers-all serial number, when determine the server to be registered be second queue in the smallest clothes of serial number When business device, distributed lock is distributed to server to be registered.

Similarly, after server to be registered gets distribution formula lock, the service in waiting in queue is obtained to second Device sends distributed lock occupied information.

It should be understood that the sequence of step S110 and S111 can be interchanged, i.e., server to be registered can be asked first Ask acquisition grouping strategy, rear request distributed lock；Can also first request distributed lock, rear request is grouped plan Slightly, herein with no restrictions.

S112 such as gets the distributed lock, then the server farm that should be added according to the grouping strategy determination, will It itself is registered in corresponding server farm.

In this step, the distributed lock is such as got, then the server zone that should be added according to the grouping strategy determination Group itself will be registered in corresponding server farm.

Specifically, as a kind of optional embodiment, the grouping strategy includes: average packet rule and the distribution Formula server cluster is divided into the information of M server farm；The server zone that should be added according to the grouping strategy determination Group itself will be registered in corresponding server farm, comprising:

Know the registered server total quantity that each server farm includes in current M server farm, passes through ratio Relatively determine the current registered least server farm of server total quantity；

For example, M is 2, i.e. distributed server cluster includes two groups of server farms, is known in first group of server farm Current registered server total quantity is 3, and current registered server total quantity is 4 in second group of server farm, then logical It crosses and compares, server determination to be registered should be registered in second group of server farm, then itself will be registered to first group of server In group.

As a kind of optional embodiment, the grouping strategy may include: weight rule of classification, the Distributed Services Device cluster is divided into the information of M server farm and the weighted value of each server farm；It is described according to the grouping strategy The server farm that determination should be added itself will be registered in corresponding server farm, comprising:

Know the registered server total quantity that each server farm includes in current M server farm, by with Corresponding preset quantity is compared, and determines that current registered server total quantity is also not up to the service of this group of preset quantity Device group；

For example M is 2, i.e., distributed server cluster includes two groups of server farms, the weight of first group of server farm Value is 1/3, and the weighted value of second group of server farm is 2/3, the destination server total quantity of the distributed server cluster It is 9, then first group of preset number of servers of server farm is 3, second group of preset number of servers of server farm It is 6.If knowing, chartered number of servers is 3 to first group of server farm, and second group of server farm has been registered Number of servers be 5, then determining registered clothes in second group of server farm after server to be registered is compared Device quantity of being engaged in also is not up to the preset quantity of second group of server farm, then itself will be registered in second group of server farm.

As a kind of optional embodiment, the grouping strategy includes: load rule of classification and the Distributed Services Device cluster is divided into the information of M server farm；The server farm that should be added according to the grouping strategy determination, will be certainly Body is registered in corresponding server farm, comprising:

The operating load for knowing each server farm in current M server farm is determined currently to transport by comparing The highest server farm of row load；

For example, M is 2, i.e. distributed server cluster includes two groups of server farms, knows first group of server farm Operating load is 80%, and the operating load in second group of server farm is 80%, then by comparing, determining first group of clothes The operating load highest of Wu Qi group, server determination to be registered should be registered in first group of server farm, then infuse itself Volume is into first group of server farm.

As a kind of optional embodiment, the grouping strategy includes: designated packet rule, the distributed server collection Group be divided into M server farm information and in each server farm group identification；It is described according to the grouping strategy The server farm that determination should be added itself will be registered in corresponding server farm, comprising:

Know the group identification of each server farm in designated group mark and current M server farm；

It determines that the designated group identifies corresponding server farm by comparison, itself will be registered to the determination Designated group identifies in corresponding server farm.

For example, M is 2, i.e., distributed server cluster includes two groups of server farms, the group of first group of server farm It is identified as A, the group identification of second group of server farm is B, if designated group is identified as A, server to be registered passes through comparison It determines that the corresponding server farm of designated group mark A should be first group of server farm, then itself is registered to first In group server farm.

Server each in this way itself can will be registered in corresponding server farm automatically according to grouping strategy, and When there is server online or offline in server farm, the requirement being grouped automatically still can satisfy；And due to different Server farm also improves the efficiency of distributed server cluster for providing different business service.

Here, the server farm that should be added according to the grouping strategy determination, itself will be registered to corresponding clothes After in Wu Qi group, further includes:

Obtain the redundant configuration file in the grouping strategy；Based on the redundant configuration file to current server itself Carry out redundant configuration；In this way when the case where delay machine occurs in server to be registered, has redundancy machine and take over server to be registered Work, not will lead to loss of data.

And obtain resource distribution file in the grouping strategy；Based on the resource distribution file to target service Data reserve corresponding resource pool；The target service data include: user's Added Business or user's confidential transactions.It in this way can be with Service clearance is provided with reserved resource pool for the Added Business and confidential transactions of user.

After distributed type assemblies establish, need the module of certain cluster service that can check by WATCH function point The corresponding node of cloth server cluster checks specific server in server farm quantity and each server farm Information.

Embodiment two

Corresponding to embodiment one, the present embodiment provides a kind of apparatus for grouping for distributed server cluster, such as Fig. 4 institute Show, described device includes: tactful acquiring unit 21, distributed lock acquiring unit 22 and registering unit 23；Wherein,

Tactful acquiring unit 21 is used to obtain the grouping strategy of the distributed server cluster, and the grouping strategy is preparatory It is stored under the corresponding node of the distributed server cluster, the distributed server cluster includes multiple server zones Group, each server farm provide different business services；It include multiple child nodes, each child node corresponding one i.e. under the node A server farm includes again multiple dynamic nodes under each child node, and corresponding each dynamic node is to have succeeded in registration (survival) server.

Specifically, when distributed server cluster starts, tactful acquiring unit 21 sends request to the node Grouping strategy is issued to server to be registered according to the solicited message by the first solicited message of grouping strategy, the node In.

Distributed lock of the distributed lock acquiring unit 22 for distributed server cluster described in request；Here, institute Stating grouping strategy and distributed lock can be stored under the same node, can also be stored under different nodes, i.e., as implemented Described in example one, distributed lock can be stored under the corresponding node of distributed server cluster, can also be stored in distributed clothes Under node except business device cluster.

It is illustrated for being stored under the same node, distributed lock acquiring unit 22 is distributed described in request It is specific as follows when the distributed lock of formula server cluster:

After tactful acquiring unit 21 gets grouping strategy, distributed lock acquiring unit 22 can continue to send out to the node Send the second solicited message of request distributed lock.Requesting the specific implementation of distributed lock, there are many modes, for example, the section Point obtains queue according to the distributed lock is added in server to be registered by second solicited message first, and described second asks Seek the precedence information that the server to be registered is had in information.

After server to be registered is added into the first acquisition queue, distributed lock acquiring unit 22 is obtained to node request Take the priority of the Servers-all in the first acquisition queue in addition to itself；Judgement it is described itself priority whether be Described first obtains the highest priority in queue, if the highest priority in the first acquisition queue, is then defaulted as straight It obtains and takes the distributed lock.

When the priority for determining itself is not the highest priority in the first acquisition queue, distributed lock acquiring unit 22 are also used to: monitoring mechanism being called to monitor the adjacent priority corresponding server higher than own priority；Determining the phase In the case where the adjacent priority corresponding server release distributed lock, the distributed lock is directly acquired.

After getting distribution formula lock, distributed lock acquiring unit 22 obtains the service in waiting in queue to first Device sends distributed lock occupied information.

As a kind of optional embodiment, the distributed server cluster described in request of distributed lock acquiring unit 22 Distributed lock, comprising:

After tactful acquiring unit 21 gets grouping strategy, distributed lock acquiring unit 22 can continue to send out to the node The third solicited message of request distributed lock is sent, the node adds server to be registered according to second solicited message Enter the distributed lock second obtains queue.

Server to be registered is added to second obtain in queue after, distributed lock acquiring unit 22 is to node request institute It states the Servers-all in the second acquisition queue in addition to itself and obtains the serial number in queue second；Judging the serial number of itself is The no the smallest serial number in the second acquisition queue is then considered as if the smallest serial number and directly acquires the distribution Formula lock.

Similarly, when the serial number for determining itself is not the smallest serial number in the second acquisition queue, distributed lock is obtained Unit 22 is also used to: calling the adjacent server that monitoring mechanism monitoring is smaller than itself serial number；Determining the adjacent service In the case that device discharges the distributed lock, the distributed lock is directly acquired.

Similarly, after getting distribution formula lock, distributed lock acquiring unit 22 is used to obtain the place in queue to second Distributed lock occupied information is sent in the server of waiting.

It should be understood that tactful acquiring unit 21 can first request grouping strategy, then distributed lock obtains single 22 request distributed lock again of member；It is also possible to lock the first request distributed lock of acquiring unit 22, tactful acquiring unit 21 Request grouping strategy again, herein with no restrictions.

Registering unit 23 is used for when determining that the distributed lock acquiring unit gets the distributed lock, according to described The server farm that grouping strategy determination should be added itself will be registered in corresponding server farm.

Specifically, referring to Fig. 5, the registering unit 23 include: first determine subelement 31, second determine subelement 32 and Registration subunit 33.

As a kind of optional embodiment, the grouping strategy includes: average packet rule and the Distributed Services Device cluster is divided into the information of M server farm；First determines that subelement 31 is used to determine certainly according to the average packet rule Body should be registered in the current registered least server farm of server total quantity；

Second determines that subelement 32 is used to know that each server farm in current M server farm to include registered Server total quantity determines the current registered least server farm of server total quantity by comparing；

Registration subunit 33 is used to that the current registered least clothes of server total quantity of the determination itself will to be registered to In Wu Qi group.

For example, M is 2, i.e., distributed server cluster includes two groups of server farms, and second determines that subelement 32 knows the Current registered server total quantity is 3 in one group of server farm, current registered server in second group of server farm Total quantity is 4, then by comparing, determination should be registered in second group of server farm, and registration subunit 33 then infuses itself Volume is into first group of server farm.

As a kind of optional embodiment, the grouping strategy may include: weight rule of classification, the Distributed Services Device cluster is divided into the information of M server farm and the weighted value of each server farm；So described first determines that son is single Member 31 is for determining that itself should be registered to current registered server total quantity is not up to the group according to the weight rule of classification In the server farm of preset quantity, the preset quantity is according to the weighted value and the distributed server collection of server farm The destination server total quantity of group determines；

Second determines that subelement 32 is used to know that each server farm in current M server farm to include registered Server total quantity determines that current registered server total quantity does not reach also by being compared with corresponding preset quantity To the server farm of this group of preset quantity；

Registration subunit 33 is for the current registered server total quantity for itself being registered to the determination to be also not up to In the server farm of this group of preset quantity.

For example M is 2, i.e., distributed server cluster includes two groups of server farms, the weight of first group of server farm Value is 1/3, and the weighted value of second group of server farm is 2/3, the destination server total quantity of the distributed server cluster It is 9, then first group of preset number of servers of server farm is 3, second group of preset number of servers of server farm It is 6.If second determines that subelement 32 knows that chartered number of servers is 3 to first group of server farm, second group of clothes Chartered number of servers is 5 for Wu Qi group, then determining registered in second group of server farm after being compared Number of servers also be not up to second group of server farm preset quantity, then itself is registered to second by registration subunit 33 In group server farm.

As a kind of optional embodiment, the grouping strategy includes: load rule of classification and the Distributed Services Device cluster is divided into the information of M server farm；So first determines subelement 31 for true according to the load rule of classification It is fixed itself to be registered in the highest server farm of current operating load；

Second determines that subelement 32 knows the operating load of each server farm in current M server farm, passes through Compare and determines the highest server farm of current operating load；

Registration subunit 33 is for itself will be registered in the highest server farm of current operating load of the determination.

For example, M is 2, i.e., distributed server cluster includes two groups of server farms, and second determines that subelement 32 knows the The operating load of one group of server farm is 80%, and the operating load in second group of server farm is 80%, then passing through ratio Compared with, determine the operating load highest of first group of server farm, when determination should be registered in first group of server farm, registration Unit 33 itself will be then registered in first group of server farm.

As a kind of optional embodiment, the grouping strategy includes: designated packet rule, the distributed server collection Group be divided into M server farm information and in each server farm group identification；So first determines subelement 31 For determining that itself should be registered to designated group identifies in corresponding server farm according to the designated packet rule；

Second determines subelement 32 for knowing each server zone in designated group mark and current M server farm The group identification of group determines that the designated group identifies corresponding server farm by comparison；

Registration subunit 33 is used to identify the designated group for itself being registered to the determination in corresponding server farm.

For example, M is 2, i.e., distributed server cluster includes two groups of server farms, the group of first group of server farm It is identified as A, the group identification of second group of server farm is B, if designated group is identified as A, second determines that subelement 32 passes through Comparison determines that the corresponding server farm of designated group mark A should be first group of server farm, and registration subunit 33 is then It itself will be registered in first group of server farm.

Here, the server to be registered further includes configuration unit 24, and registration subunit 33 itself will be registered to accordingly After in server farm, configuration unit 24 is used for:

Embodiment three

Present embodiments provide a kind of group technology for distributed server cluster, the distributed server cluster Including multiple server farms, each server farm is for providing different business services.The area of the present embodiment and embodiment 1 It is not, embodiment is first is that be the angle introduction of executing subject from server to be registered, embodiment is third is that from distributed server Management node (usually host node is known as node to be consistent with embodiment one in the present embodiment) in cluster is to hold The angle of row main body is write.

As shown in Figure 6, which comprises

S410 receives the acquisition request that server to be registered is sent；

When needing to be grouped distributed server cluster automatically, when distributed server cluster starting, node The acquisition request that server to be registered is sent is received, the node has pre-saved grouping strategy, and the node corresponds to described point Cloth server cluster.

S411 sends preset grouping strategy to the server to be registered according to the acquisition request, for it is described to Registrar is when getting distributed lock, the server farm that can should be added according to the grouping strategy determination, by itself It is registered in corresponding server farm.

After receiving acquisition request, preset grouping plan is sent to the server to be registered according to the acquisition request Slightly, so that the server to be registered is when getting distributed lock, the service that can should be added according to the grouping strategy determination Device group itself will be registered in corresponding server farm.

Here, the server to be registered needs to send distributed lock acquisition to node and asks after getting grouping strategy It asks, distributed lock can be got.

It should be noted that the grouping strategy and distributed lock can be stored under the same node, can also save Under different nodes, i.e., as described in embodiment one, distributed lock can be stored in the corresponding node of distributed server cluster Under, it can also be stored under the node except distributed server cluster.Server to be registered can first send grouping strategy Acquisition request can also first send the acquisition request of distributed lock, herein with no restrictions.

It is illustrated for being stored under the same node, server to be registered Distributed Services described in request It is specific as follows when the distributed lock of device cluster:

After server to be registered gets grouping strategy, it can continue to send request distributed lock to the node Request, requesting the specific implementation of distributed lock, there are many modes, for example, the node is according to the request for obtaining distributed lock Server to be registered is added the first of the distributed lock and obtains queue, is had in the request for obtaining distributed lock described The precedence information of server to be registered.

After server to be registered gets grouping strategy, it can continue to send request distributed lock to the node Request, the node are obtained according to the distributed lock is added in server to be registered by the request for obtaining distributed lock second Take queue.

If server to be registered gets the distributed lock, then the server that should be added according to the grouping strategy determination Group itself will be registered in corresponding server farm.

Here, after itself being registered in corresponding server farm, further includes:

Example IV

The present embodiment additionally provides a kind of distributed server cluster, and the present embodiment is write using entire cluster as angle It writes, referring to Fig. 7, the distributed server cluster includes multiple server farms, for example is M group (1~M)；Clothes to be registered Business device 51 and a node 52；Wherein, the node 52 is used to save the grouping strategy of the distributed server cluster, described The corresponding distributed server cluster of node 52；With continued reference to Fig. 2, the server 51 to be registered includes: that strategy obtains list Member 21, distributed lock acquiring unit 22 and registering unit 23；Wherein,

Tactful acquiring unit 21 is used to obtain the grouping strategy of the distributed server cluster, and the grouping strategy is preparatory It is stored under the corresponding node 52 of the distributed server cluster.

Specifically, referring to Fig. 3, the registering unit 23 include: first determine subelement 31, second determine subelement 32 and Registration subunit.

The embodiment of the invention also discloses a kind of computer readable storage mediums, are stored thereon with computer program, the journey The step of above method embodiment is realized when sequence is executed by processor.

The embodiment of the invention also discloses a kind of computer equipments, including memory, processor and storage are on a memory And the computer program that can be run on a processor, the processor realize the step of above method embodiment when executing described program Suddenly.

A kind of group technology and device for distributed server cluster provided by the invention, the distributed server Cluster includes multiple server farms, which comprises obtain the grouping strategy of the distributed server cluster, described point Group policy is pre-stored under the corresponding node of the distributed server cluster；Distributed server cluster described in request Distributed lock；The distributed lock is such as got, then the server farm that should be added according to the grouping strategy determination will be certainly Body is registered in corresponding server farm；In this way, preset grouping strategy is stored under node, when cluster needs to carry out certainly When dynamic grouping, server to be registered obtains grouping strategy to node can basis in the case where itself gets distributed lock The server farm that the grouping strategy determination should be added itself will be registered in corresponding server farm, thus can be with Distributed server cluster is divided into multiple server farms, and then each clothes can be made by configuring each server farm Wu Qi group provides different business services, to improve the efficiency of distributed server cluster；And the clothes in server farm Device be engaged in when online or offline, automatic grouping can be still met by grouping strategy and is required, this assures distributions to take The operational efficiency of business device cluster.

Algorithm and display are not inherently related to any particular computer, virtual system, or other device provided herein. Various general-purpose systems can also be used together with teachings based herein.As described above, it constructs required by this kind of system Structure be obvious.In addition, the present invention is also not directed to any particular programming language.It should be understood that can use various Programming language realizes summary of the invention described herein, and the description done above to language-specific is to disclose this hair Bright preferred forms.

In the instructions provided here, numerous specific details are set forth.It is to be appreciated, however, that implementation of the invention Example can be practiced without these specific details.In some instances, well known method, structure is not been shown in detail And technology, so as not to obscure the understanding of this specification.

Similarly, it should be understood that in order to simplify the disclosure and help to understand one or more of the various inventive aspects, Above in the description of exemplary embodiment of the present invention, each feature of the invention is grouped together into single implementation sometimes In example, figure or descriptions thereof.However, the disclosed method should not be interpreted as reflecting the following intention: i.e. required to protect Shield the present invention claims features more more than feature expressly recited in each claim.More precisely, as following Claims reflect as, inventive aspect is all features less than single embodiment disclosed above.Therefore, Thus the claims for following specific embodiment are expressly incorporated in the specific embodiment, wherein each claim itself All as a separate embodiment of the present invention.

Those skilled in the art will understand that can be carried out adaptively to the module in the equipment in embodiment Change and they are arranged in one or more devices different from this embodiment.It can be the module or list in embodiment Member or component are combined into a module or unit or component, and furthermore they can be divided into multiple submodule or subelement or Sub-component.Other than such feature and/or at least some of process or unit exclude each other, it can use any Combination is to all features disclosed in this specification (including adjoint claim, abstract and attached drawing) and so disclosed All process or units of what method or apparatus are combined.Unless expressly stated otherwise, this specification is (including adjoint power Benefit require, abstract and attached drawing) disclosed in each feature can carry out generation with an alternative feature that provides the same, equivalent, or similar purpose It replaces.

In addition, it will be appreciated by those of skill in the art that although some embodiments in this include institute in other embodiments Including certain features rather than other feature, but the combination of the feature of different embodiment means in the scope of the present invention Within and form different embodiments.For example, in the following claims, embodiment claimed it is any it One can in any combination mode come using.

Various component embodiments of the invention can be implemented in hardware, or to run on one or more processors Software module realize, or be implemented in a combination thereof.It will be understood by those of skill in the art that can be used in practice Microprocessor or digital signal processor (DSP) realize gateway according to an embodiment of the present invention, proxy server, in system Some or all components some or all functions.The present invention is also implemented as executing side as described herein Some or all device or device programs (for example, computer program and computer program product) of method.It is such It realizes that program of the invention can store on a computer-readable medium, or can have the shape of one or more signal Formula.Such signal can be downloaded from an internet website to obtain, and perhaps be provided on the carrier signal or with any other shape Formula provides.

It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and ability Field technique personnel can be designed alternative embodiment without departing from the scope of the appended claims.In the claims, Any reference symbol between parentheses should not be configured to limitations on claims.Word "comprising" does not exclude the presence of not Element or step listed in the claims.Word "a" or "an" located in front of the element does not exclude the presence of multiple such Element.The present invention can be by means of including the hardware of several different elements and being come by means of properly programmed computer real It is existing.In the unit claims listing several devices, several in these devices can be through the same hardware branch To embody.The use of word first, second, and third does not indicate any sequence.These words can be explained and be run after fame Claim.

The invention discloses, A1, a kind of group technology for distributed server cluster, the distributed server collection Group includes multiple server farms, which comprises

The distributed lock of distributed server cluster described in request；

A2, method as described in a1, the grouping strategy include: average packet rule and the distributed server collection Group is divided into the information of M server farm；The server farm that should be added according to the grouping strategy determination, itself is infused Volume is into corresponding server farm, comprising:

A3, method as described in a1, the grouping strategy include: weight rule of classification, the distributed server cluster It is divided into the information of M server farm and the weighted value of each server farm；It is described to be answered according to grouping strategy determination The server farm of addition itself will be registered in corresponding server farm, comprising:

A4, method as described in a1, the grouping strategy include: load rule of classification and the distributed server collection Group is divided into the information of M server farm；The server farm that should be added according to the grouping strategy determination, itself is infused Volume is into corresponding server farm, comprising:

A5, method as described in a1, the grouping strategy include: designated packet rule, the distributed server cluster Be divided into M server farm information and in each server farm group identification；It is described true according to the grouping strategy Surely the server farm that should be added itself will be registered in corresponding server farm, comprising:

The group identification for knowing each server farm in designated group mark and current M server farm, by right Than determining that the designated group identifies corresponding server farm；

A6, the method as described in any one of A1 to A5, each server farm in the distributed server cluster For providing different business services.

B7, a kind of apparatus for grouping for distributed server cluster, the distributed server cluster include multiple clothes Wu Qi group, described device include:

B8, device as described in b7, the grouping strategy include: average packet rule and the distributed server collection Group is divided into the information of M server farm；The registering unit includes:

Second determines subelement, for knowing that each server farm in current M server farm includes registered Server total quantity determines the current registered least server farm of server total quantity by comparing；

B9, device as described in b7, the grouping strategy include: weight rule of classification, the distributed server cluster It is divided into the information of M server farm and the weighted value of each server farm；The registering unit includes:

Second determines subelement, for knowing that each server farm in current M server farm includes registered Server total quantity determines that current registered server total quantity does not reach also by being compared with corresponding preset quantity To the server farm of this group of preset quantity；

B10, device as described in b7, the grouping strategy include: load rule of classification and the distributed server Cluster is divided into the information of M server farm；The registering unit includes:

Second determines subelement, for knowing the operating load of each server farm in current M server farm, leads to It crosses to compare and determines the highest server farm of current operating load；

B11, device as described in b7, the grouping strategy include: designated packet rule, the distributed server cluster Be divided into M server farm information and in each server farm group identification；The registering unit includes:

Second determines subelement, for knowing each server zone in designated group mark and current M server farm The group identification of group determines that the designated group identifies corresponding server farm by comparison；

B12, the device as described in any one of B7 to B11, each server zone in the distributed server cluster Group is for providing different business services.

C13, a kind of group technology for distributed server cluster, the distributed server cluster include multiple clothes Wu Qi group, which comprises

Receive the acquisition request that server to be registered is sent；

D14, a kind of distributed server cluster, including multiple server farms, server to be registered and a node；Its In, the node is used to save the grouping strategy of the distributed server cluster, and the node corresponds to the Distributed Services Device cluster；The server to be registered includes:

D15, the cluster as described in D14, the grouping strategy include: average packet rule and the distributed server Cluster is divided into the information of M server farm；The registering unit includes:

D16, the cluster as described in D14, the grouping strategy include: weight rule of classification, the distributed server collection Group is divided into the information of M server farm and the weighted value of each server farm；The registering unit includes:

D17, the cluster as described in D14, the grouping strategy include: load rule of classification and the distributed server Cluster is divided into the information of M server farm；The registering unit includes:

D18, the cluster as described in D14, the grouping strategy include: designated packet rule, the distributed server collection Group be divided into M server farm information and in each server farm group identification；The registering unit includes:

D19, the cluster as described in any one of D14 to D18, each server zone in the distributed server cluster Group is for providing different business services.

E20, a kind of computer readable storage medium, are stored thereon with computer program, when which is executed by processor It performs the steps of

The grouping strategy of the distributed server cluster is obtained, the grouping strategy is pre-stored in the distributed clothes It is engaged under the corresponding node of device cluster；The distributed server cluster includes multiple server farms；

The distributed lock of distributed server cluster described in request；

F21, a kind of computer equipment for distributed server cluster, comprising:

At least one processor；And

At least one processor being connect with the processor communication, wherein

The memory is stored with the program instruction that can be executed by the processor, and the processor calls described program to refer to Enable the method being able to carry out as described in A1 to A6 or C13 is any.

Claims

1. a kind of group technology for distributed server cluster, which is characterized in that the distributed server cluster includes Multiple server farms, which comprises

The grouping strategy of the distributed server cluster is obtained, the grouping strategy is pre-stored in the distributed server Under the corresponding node of cluster；

The distributed lock of distributed server cluster described in request；

The distributed lock is such as got, then the server farm that should be added according to the grouping strategy determination registers itself Into corresponding server farm.

2. the method as described in claim 1, which is characterized in that the grouping strategy includes: that average packet is regular and described Distributed server cluster is divided into the information of M server farm；The service that should be added according to the grouping strategy determination Device group itself will be registered in corresponding server farm, comprising:

Being determined according to the average packet rule itself should be registered to the current registered least server zone of server total quantity In group；

Know the registered server total quantity that each server farm includes in presently described M server farm, passes through ratio Relatively determine the current registered least server farm of server total quantity；

3. the method as described in claim 1, which is characterized in that the grouping strategy includes: weight rule of classification, the distribution Formula server cluster is divided into the information of M server farm and the weighted value of each server farm；It is described to divide according to The server farm that group policy determination should be added itself will be registered in corresponding server farm, comprising:

Being determined according to the weight rule of classification itself should be registered to current registered server total quantity to be not up to the group default In the server farm of quantity, the preset quantity is according to the weighted value of server farm and the distributed server cluster Destination server total quantity determines；

Know the registered server total quantity that each server farm includes in presently described M server farm, by with Corresponding preset quantity is compared, and determines that current registered server total quantity is also not up to the service of this group of preset quantity Device group；

The current registered server total quantity for itself being registered to the determination is also not up to the server of this group of preset quantity In group.

4. the method as described in claim 1, which is characterized in that the grouping strategy includes: load rule of classification and described Distributed server cluster is divided into the information of M server farm；The service that should be added according to the grouping strategy determination Device group itself will be registered in corresponding server farm, comprising:

The operating load for knowing each server farm in presently described M server farm is determined currently to transport by comparing The highest server farm of row load；

5. the method as described in claim 1, which is characterized in that the grouping strategy includes: designated packet rule, the distribution Formula server cluster be divided into M server farm information and in each server farm group identification；It is described according to institute The server farm that grouping strategy determination should be added is stated, itself will be registered in corresponding server farm, comprising:

The group identification for knowing each server farm in designated group mark and presently described M server farm, by right Than determining that the designated group identifies corresponding server farm；

6. a kind of apparatus for grouping for distributed server cluster, which is characterized in that the distributed server cluster includes Multiple server farms, described device include:

Tactful acquiring unit, for obtaining the grouping strategy of the distributed server cluster, the grouping strategy is pre-saved Under the corresponding node of the distributed server cluster；

Registering unit, for when determining that the distributed lock acquiring unit gets the distributed lock, according to the grouping The server farm that tactful determination should be added itself will be registered in corresponding server farm.

7. a kind of group technology for distributed server cluster, which is characterized in that the distributed server cluster includes Multiple server farms, which comprises

Receive the acquisition request that server to be registered is sent；

Preset grouping strategy is sent to the server to be registered according to the acquisition request, for the server to be registered When getting distributed lock, the server farm that can should be added according to the grouping strategy determination itself will be registered to accordingly Server farm in.

8. a kind of distributed server cluster, which is characterized in that including multiple server farms, server to be registered and a section Point；Wherein, the node is used to save the grouping strategy of the distributed server cluster, and the node corresponds to the distribution Server cluster；The server to be registered includes:

9. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is held by processor Be able to carry out when row claim 1 to 5 or claim 7 it is any as described in method.

10. a kind of computer equipment for distributed server cluster characterized by comprising

At least one processor；And

At least one processor being connect with the processor communication, wherein

The memory is stored with the program instruction that can be executed by the processor, and the processor calls described program to instruct energy It is enough execute claim 1 to 5 or claim 7 it is any as described in method.