CN101996094A - Method and system for managing distributed resources - Google Patents
Method and system for managing distributed resources Download PDFInfo
- Publication number
- CN101996094A CN101996094A CN2009101656394A CN200910165639A CN101996094A CN 101996094 A CN101996094 A CN 101996094A CN 2009101656394 A CN2009101656394 A CN 2009101656394A CN 200910165639 A CN200910165639 A CN 200910165639A CN 101996094 A CN101996094 A CN 101996094A
- Authority
- CN
- China
- Prior art keywords
- resource
- management
- objects
- data
- control center
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Abstract
The invention relates to a method and system for managing distributed resources. The system comprises resource objects, data objects, management objects and a control center, wherein the control center is used for configuring the service functions and the backup modes of all resource objects in the system and allocating management tasks for all management objects; the management objects are used for taking charge of the resource management of managed recourse objects according to the management tasks allocated by the control center and providing a query and control service aiming at the managed recourse objects; the resource objects are used for completing corresponding logic functions under the control of the management objects; and the data objects are used for storing the resource information in the resource objects and the resource control information in the management objects. The invention increases the efficiency of resource utilization and simplifies the resource management process by adopting grading object management and effective fault-tolerant design and ensures fast recovery capability and mechanism under the condition that various errors occur in the resource.
Description
Technical field
The present invention relates to a kind of distributive resource managing method and system.
Background technology
In Distributed Computer System, different CPU veneer is realized different functions, and sharing out the work and helping one another between the veneer of, difference in functionality dissimilar by all externally embodies the unified and powerful processing power of total system.
Realization is to the management of different veneer different resources, make external function of system and processing power can adopt increase veneer quantity extended mode to realize very easily, certain processing veneer descends systemic-function unaffected unusually in the system, all is the problem that field of distributed type is paid close attention to.Below for sake of convenience, these are had the resource disposable plates that realize certain function abbreviate resource board as, and with these be used to monitor, a class veneer of distribution and management resource plate is called as the resource management plate.For resource management, the disposal route of industry has following several at present:
A kind of method is to adopt resource management and resource to use the mode of unification, and the resource of oneself is only managed and used to each veneer.The problem that this mode exists is that resource can not be shared, when the resource of certain veneer all can not have been worked, although other veneer has idle resource to apply for.Under this management mode, in order to guarantee that reliability need dispose more veneer and back up, the hardware cost of beginning increases, and also is unfavorable for later maintenance.
Another method is to adopt the centralized resource management mode, is that 96117990.2 the disclosed scheme of Chinese patent (publication number is CN1156362A) is exactly a representative of centralized resource management as number of patent application.This scheme is placed on all resources on the veneer carries out unified management, and other veneers need resource must arrive on this veneer to apply for.In case the problem that the method exists is the veneer of the unified centralized management resource machine of delaying, total system just will be the situation appearance of flop so, and will be abnormally dangerous.
Last a kind of method is to adopt the distributed resource management mode of dynamic self-adapting, is that the disclosed scheme of Chinese patent (publication number is CN101167307) of CN200580049520.0 is exactly a representative of distributed resource management as number of patent application.This scheme is shone upon all physical resources and logical resource, and the overall situation is shared on the veneer of all resource managements.The problem that the method exists is whole resource sharings and safeguards that the expense of its state is huge, veneer that need not all resource managements is all safeguarded the resource status table of a overall situation, the work born of resource management is comparatively complicated like this, and physical resource and logical resource configuration relation are fixed, if the later stage, system needed dilatation, need reconfigure, handle dumb.
Fig. 1 is the theory diagram of the distributed resource management system of dynamic self-adapting, and the corresponding resource of each resource management board management oneself is regularly mutual with all resource management plates on the resource management plate, obtains the information of overall available resources in the system.Resource has the resource management plate of respective home in this scheme, if the resource management plate lost efficacy, the resource of its management can't be used by others in system; In addition,,, must revise global configuration, can not set up new administrative relationships by dynamic increase resource management plate in case resource increases because the system resource management relation is fixed.And resource changes at any time in the system, and the resource management plate need constantly carry out broadcast synchronization, and along with resource management plate quantity increases, system's inter-sync communication flows is doubled and redoubled.Everything has all increased the cost of safeguarding greatly.
Summary of the invention
The technical problem to be solved in the present invention provides a kind of distributive resource managing method and system,
In order to address the above problem, the invention provides a kind of distributed resource management system, comprise resource object, data object, management object and control center, wherein:
Described control center is used for, the business function of all resource objects and backup mode in the configuration-system, and be all management object allocation manager tasks;
Described management object is used for, and according to the management role that described control center distributes, is responsible for the resource management of the resource object managed, and inquiry and control service at the resource object of being managed are provided;
Described resource object is used for, and finishes corresponding logic function under the control of management object;
Described data object is used for, resource information in the storage resources object and the resource control information in the management object.
Further, described management object also comprises:
Copy management module is used to the main resource object and main with data object alternative back-up storage district of using, and carries out the object back-up storage; After finishing the object back-up storage, set up backup object and main with the mapping relations between object, and realize backup object with main with object synchronously;
The failure management module was used for when the master was lost efficacy with data object, adopted the preliminary data object that described master is carried out data with data object and recovered.
Further, described resource object also is used for, and after taking place unusual or moving back clothes, gives corresponding management object with status change notification;
Described management object also is used for, after receiving described status change notification, the state of this resource object and resource backup object thereof and Backup Data object in the global resource information data object is changed to down state, and, if all resource objects of same logic function all move back clothes, then close described logic function; And described state variation is distributed to other management objects and/or control center by incident.
Further, described resource object is used for, and reports resource information to management object after foundation, and the configuration of request logic function; And, after carrying out initialization, resource status is reported management object according to the logic function of management object configuration;
Described management object is used for, and inquires about the business function and the backup mode of this resource object according to the resource information that resource object reports to control center, and offers this resource object; And, after receiving the resource status that reports, the resource information of this resource object is write in the data object of global resource information, and resource increased be distributed to other management objects and/or control center by incident.
Further, described management object is used for, and disposes to control center's request management object after foundation;
Described control center is used for, and creates the record of this management object, and distributes logic function, the backup mode of the resource object that this management object manages and the maximum quantity of managing every type resource object.
Further, described management object also is used for, when the resource object of being managed is upgraded, the state of described resource object is changed to down state in the data object of global resource information, and resource closed by incident is distributed to other management objects and/or control center; And, after the upgrading of described resource object is finished, the state of described resource object is changed to normal operating condition in the data object of global resource information.
The present invention also provides a kind of distributive resource managing method, and system resource is divided into resource object, data object, management object and control center by function, and this method comprises:
By the business function and the backup mode of all resource objects in the described control center configuration-system, and be all management object allocation manager tasks;
The management role that described management object is distributed according to described control center, the resource management of the resource object that responsible quilt is managed, and provide the inquiry of the resource object of being managed at quilt and control to serve;
Described resource object is finished corresponding logic function under the control of management object;
Resource information and the resource control information in the management object in the resource object are stored in the described data object.
Further, described management object is main with resource object and main with data object alternative back-up storage district, carries out the object back-up storage; After finishing the object back-up storage, set up backup object and main with the mapping relations between object, and realize backup object with main with object synchronously;
When the master was lost efficacy with data object, described management object adopted the preliminary data object that described master is carried out the data recovery with data object.
Further, described method also comprises:
After described resource object takes place unusual or moves back clothes, give corresponding management object with status change notification;
After receiving described status change notification, described management object is changed to down state with the state of this resource object and resource backup object thereof and Backup Data object in the global resource information data object, and, if all resource objects of same logic function all move back clothes, then close described logic function; And described state variation is distributed to other management objects and/or control center by incident.
Further, when increasing resource object, described method comprises:
After setting up resource object, report resource information to described management object, and the configuration of request logic function;
Described management object is inquired about the business function and the backup mode of this resource object according to the resource information that resource object reports to control center, and offers this resource object;
Described resource object reports described management object with resource status after carrying out initialization according to the logic function of described management object configuration;
Described management object writes the resource information of this resource object in the data object of global resource information after receiving the described resource status that reports, and resource increased is distributed to other management objects and/or control center by incident.
Further, when increasing management object, described method comprises:
After setting up management object, dispose to control center's request management object;
The record of this management object is created by described control center, and distributes logic function, the backup mode of the resource object that this management object manages and the maximum quantity that can manage every type resource object.
Further, when described resource object was upgraded, described method comprised:
When the resource object of being managed was upgraded, management object was changed to down state with the state of described resource object in the data object of global resource information, and resource closed by incident was distributed to other management objects and/or control center;
After described resource object upgrading was finished, described management object was changed to normal operating condition with the state of described resource object in the data object of global resource information.
Description of drawings
Fig. 1 is the resource management architectures synoptic diagram of existing distributed resource management system;
Fig. 2 is the schematic network structure of distributed resource management system in the embodiment of the invention;
Fig. 3 is the synchronous schematic flow sheet of object data in the embodiment of the invention;
Fig. 4 is the schematic flow sheet that increases the resource management plate in the embodiment of the invention;
Fig. 5 is the schematic flow sheet that increases resource board in the embodiment of the invention;
Fig. 6 is the schematic flow sheet of resource board upgrading in the embodiment of the invention.
Embodiment
Core concept of the present invention is: a kind of tolerant system based on the classification Object Management group is provided, Resource Manager, resource entity and concrete resource information are carried out functional abstract, use modularization object mode to manage, and in each object, adopt fault-tolerant design, to improve the efficient that resource is used, simplify the resource management flow process, and guarantee under the situation that resource makes a mistake, to possess the ability and the mechanism of fast quick-recovery.
Among the present invention, be system to be carried out assembly according to management object, resource object and data object divide and functional abstract.The relation of resource and object is that resource is the concrete physical carrier of functional object, to liking the abstractdesription of resource or resource collection specific logical function.
Object is not only the abstract of specific logical function in the system, also is the fault-tolerant management unit that forms on distributed basis simultaneously, finishes by set up copy management module, failure management module, daily record and recovery module in basic communications platform and public service.Object is divided into primary and backup state, and the object that is in stand-by state in the system carries out data sync to main with object under copy management module control, make it to be consistent.If main lost efficacy with object, standby object can be realized the function replacement that can't harm.
Management object is in charge of state, configuration and the control of various resources in the system and all objects.Resource object is corresponding to abstract by unity logic function on the middle single one physical resource of system management object and resource object.Data object is the storage of the high availability that can be used by other object.Concrete function is divided into some control modules, and control module is the least unit of carrying concrete function, and control module is also with the mode unified management and the control of object.
Be provided with control center in the management object group that constitutes by management object set, be in charge of the distribution of each member's work of group of objects and collaborative.The management object group can be regarded a distributed system as, finishes the maintenance of whole configuration informations in the system and all resource statuss in the centralized management system.And each management object is used for the resource information of maintenance system as an agency of management object group, and resource information leaves in the data object of the overall situation.
Resource object is the abstractdesription to specific physical resource specific function in the system, it mainly acts on is to finish resource increase/deletion, resource backup copy, resource load sharing strategy, resource status control, functions such as resource status issue under management object control.
Data message in the resource object adopts the mode of data object to store.Data object is made up of the data storage of this locality establishment and the storage backup or the error checking and correction of strange land structure.Data object can provide reliable data storage function for upper layer application, guarantees the security of data access under various exception conditions.
As from the foregoing, the present invention makes upper layer application can be reduced to quoting an object or group objects set for the use of resource by the subject componentization with system.And the state variation of object can be announced the upper strata by the mode of incident issue.The incident issue refers to a kind of proactive notification mechanism of object.Whether upper layer application can change by oneself needing to pay close attention to the own object of being concerned about when using certain object, if object will change sometime in future, can in time notify this application by message event.
Based on above thinking, the invention provides a kind of technical scheme of the distributed resource management based on the classification object, be applied to comprise the distributed system of a plurality of resource boards and a plurality of resource management plates, this scheme comprises following content:
A. the control center in the resource management group of system disposes the use-pattern of all resource boards and resource management plate, comprises business function that every type veneer can provide and backup mode etc., and is handed down to each resource board and resource management plate;
Wherein, business function refers to the function service-number of unified planning, and use to resource only need number just can be realized by function service in the upper strata, and does not need to be concerned about the concrete physical message of resource; Backup mode refers to that the veneer of the type supports the concrete scope of N and M in N+M (N master uses resource, M the resource backup) backup mode.
B. the Task Distribution of all resource management plates of the unified control of control center and collaborative is determined control center earlier after the resource management plate powers on, and management object is obtained the resource object list information that this veneer need be responsible for bearing from control center;
C. after resource board powers on, report its other resources information to the management object group, the management object group is inquired about the function service-number and the backup mode of this resource board according to the resource self information, offers this resource board, gives certain management object with the resource management agent Task Distribution then;
D. on this management object, set up the resource object control module, be responsible for the resource management of this resource board, and inquiry and control service at this resource board are provided;
Wherein, the resource object control module is responsible for collecting each resource running status of managed object, and reports control center, when the resource object of being managed takes place unusual or moves back clothes, the state variation incident is distributed to the upper layer application of care.
E. under management object control, resource object on the resource board and data object can be realized the management of other resource board of the same type by quick copy, also can realize single board backup by keeping backup object and main data consistency with object fast synchronously;
Wherein, be carried on a series of object (management objects as the whole logic functions of the veneer of functional entity, resource object and data object) in the combination, object itself has location independence and redundancy backup feature, so the quick copy of object can be provided, under the situation that certain object lost efficacy, can use the object of quick copy to realize original function replacement.
F. provide the smooth upgrade function under management object control, it is upgrade status that veneer is set in the resource object control module, the version objects after the upgrading is carried out the upper strata isolate, and after upgrading was finished, cancellation veneer isolation reentered service.
Among the present invention, resource entity can be regarded the set of a group objects as, and the veneer smooth upgrade can be reduced to the function upgrading of each object self in the set; Upgrade for object version, can in the management object control module, upgrade status be set, related resource is in non-service state, veneer is in segregate state in system, but underlying resource control keeps unimpeded, in case upgrading is finished, control center removes the control that the upper strata is isolated to management object control module release upgrade success incident.
Further, above-mentioned distribution management method based on the classification object also has following characteristics:
1. internal resource can self-adaptation expands: for the increase of resource, and the resource physical message Resources allocation object that initiatively reports according to resource entity, and the relevant management object control module of application does resource status control, provides the upper strata relevant logic function; For the deletion of resource, corresponding management object control unit is recovered in the idle control module pond, and resource object is changed to unavailable, closes relevant logic function;
2. management object mainly is to distinguish normal and abnormality for the State Control of managed object, and the upgrading isolation.Normal condition be resource in working order; Abnormality refers to that resource object detects unusually, needs self-healing state, needs under this state to keep isolating fully with other object; The upgrading isolation is that object is in upgrade status, but because new and old edition may be incompatible, keep this part object to be in part and isolate, and promptly upper layer application is isolated, and the state that bottom control is unimpeded in case upgrading is finished, is removed the part isolated controlling.
3. data object overall situation unified management, data object can provide the function of distributed storage.Create the storage of the local retention data object of veneer of data object, realize quick access; Data object need back up or verification on other processor, guarantees that the data when local data object lost efficacy are recovered.Adopt the regularly way of synchronous and real-time locking phase combination, guarantee the requirement of user for data storing reliability, fault-tolerance and availability.Specifically be divided into two types of static data backup and dynamic data backups, static data mainly refers to user's configuration data and system journal and partial service data, and its characteristics are that data volume is big, upgrades to change slowly the reliability requirement height; Dynamic data mainly is the data of professional some intermediatenesses in operational process, and characteristics are that data volume is relatively little, upgrades to change soon, and reliability requirement is relatively low.At two kinds of data types, all use the data fault-tolerant assembly management, user applies data of different types object directly writes down increase at data object, deletion, operations such as inquiry, fault-tolerant assembly is responsible for duplicating of data object, failure management, daily record, and wrong the recovery.Difference is, static data mainly is placed in the storage medium, and dynamic data is retained in the internal memory of single board.Can adopt the mode of backup or error check that storage is backed up, and corresponding fast quick-recovery.
4. the user of system resource management and system resource separates, and management object group membership manages liking loose coupling for its management.Management object is responsible for collecting each resource running status of managed object, and reports control center's service, when the managed object resource takes place unusual or moves back clothes, status change notification is given the upper layer application of being concerned about.Resource Manager is indifferent to concrete resource user, and resource user is to obtain by the mode of incident issue for the variation of resource status.Change when Resource Manager perceives resource status, be distributed to the resource user that all were subscribed to by incident, to trigger the timely response of relevant upper layer application at once.
As from the foregoing, maximum characteristics of the present invention are by the classification Object Management group, abstract management object, resource object and the data object of being divided into of systemic-function, each object adopts fault-tolerant design, realizes quick copy and function replacement, internal resource backup and the veneer smooth upgrade of the expansion of veneer self-adaptation, object.Improve the efficient that resource is used, simplify the resource management flow process, and guarantee under the situation that resource makes a mistake, to possess the ability and the mechanism of fast quick-recovery.
Another characteristics of the present invention are that the rule of resource distribution (rather than traditional concrete resource distribution relation) is joined in the system resource management, draw up concrete configuration relation by the resource management responsibility system, can dynamically adjust configuration with the actual quantity of resource management plate and resource board, thus the situation of the uneven and various system exceptions of each member's veneer division of labor in the answering system flexibly.
Be further described below in conjunction with the enforcement of the drawings and specific embodiments technical solution of the present invention.
The embodiment of the invention based on the network structure of the distributed management system of classification object as shown in Figure 2, with abstract management object, resource object and the data object of being divided into of systemic-function, wherein:
Management object is in charge of the state of various resources in the system and all objects, configuration and control.A distributed system be can regard as by the management object group that the management object set constitutes, the maintenance of whole configuration informations in the system and all resource statuss in the centralized management system finished.Comprise control center in the management object group, control center is in charge of the distribution of each member's work of group of objects and works in coordination with.Management object member in the management object group bears main upper-layer service, and realizes the management of (comprising resource object and data object) of other object.
Unity logic function abstract on the single one physical resource in the corresponding system of resource object.Having one or more resource objects on the same physical resource, all is to manage by the control module on the management object.For for simplicity, the situation of having only a resource object on the physical resource has only been described among Fig. 2.
Data object is the storage of the high availability that can be used by other object.It may be noted that data object is a global administration, does not belong to some veneers.But the veneer of creating data object need be realized quick access in the storage of local retention data object.Data object also need back up on other processor or have verification, thereby guarantees the data recovery when local data object lost efficacy.
Based on the distributed resource management system shown in Fig. 2, before system's operation, need the function service and the backup mode of all types of veneers of configuration earlier.The invention provides the function service-number of unified planning, use to resource board only need number just can be realized by function service in the upper strata; Backup mode refers to that the veneer of the type supports the concrete scope of N and M in N+M (N master uses resource, M the resource backup) backup mode.
Based on the system of above distributed resource management, its management to object is finished realization by set up copy management module, failure management module, daily record and recovery module in basic communications platform and public service.Fig. 3 shows the operation steps of following when creating object:
Step 301 when applying for object on veneer, at first need be set up the object storage district in veneer this locality;
Step 302, by the replication manager alternative back-up storage district of object, its effect is to select suitable veneer to carry out the object back-up storage between the veneer of same type;
The object tools manager generally resides on the resource management plate, and only works when establishment/deletion backup object and object synchronization failure, and normal object synchronization does not then need to participate in.
Step 303, the object tools manager object tools by other veneer of the same type is set up the mapping relations of backup object and native object, and after finishing back-up storage and finishing, returns the object storage address to veneer;
Step 304, the object storage district returns the object number that takies;
Step 305 after writing data and receive return results in object, is carried out data sync from the trend backup object;
Step 306, if the object synchronization success, then flow process finishes; If object synchronization is unsuccessful, the notify object replication manager is initiated object synchronization again.
When the master was lost efficacy with object, standby object can provide service in a period of time, and assisted main rebuliding with object.After the stable state, the master need guarantee maximum access efficiency to the management object of creating it or the migration of resource object place veneer with data object.Standby object lost efficacy, and was then selected other backup object by the decision-making of object tools manager, and guaranteed on new backup object and main data sync with object.
Fig. 4 is based on the process of the increase resource management plate of above distributed resource management system, mainly may further comprise the steps:
Step 401 when increasing the resource management plate, mainly is to set up management object;
Step 402, newly-built management object is to control center's request management object configuration;
Step 403, control center sets up the management object record, allocation manager resource board type and quantity;
Step 404, management role is assigned to management object by control center, comprises resource board type, backup mode that this management object need be born and maximum quantity itself that can manage every type of resource board etc.;
Step 405, the management object initialization, management object is as an agency of management object group, can independently finish the maintenance of a class in the system or configuration of a few class resource board and follow-up related resource state.
In addition, above-mentioned distributed resource management system is when deletion resource management plate, and at first control center can be set to whole resource objects of this resource management board management unavailable in the data object of global resource information.The resource object that the resource management plate is managed is not having under the management object control module control situation, can initiate initialization (logic function request) again, with the administrative relationships of foundation and other resource management plate.
Fig. 5 is based on the process that above distributed resource management system increases resource board.Mainly may further comprise the steps:
Step 501, resource board is set up resource object, obtains the physical message of local resource;
Step 502, resource board is to the logic function configuration of management object group request resource object;
Step 503, the management object of bearing the type resource board responds this logic function request, and sets up control module, is responsible for the management and the control of this resource object, is the resource board dispatch logic function;
It should be noted that, the management object group has the qualification that a plurality of management objects possess this resource board of management, but the data object control of record global resource information writes, if previous management object writes, then other management object only possesses the ability of reading to this resource object.Therefore, actual management and the control of having only first to write global resource management of information object imputability resource object.
Step 504, resource object reports management object control module with its other resources state according to the logic function initialization of configuration after initialization is finished;
Step 505, the management object control module writes resource information in the data object of global resource information, and the resource increase is distributed to other management object by incident;
Step 506, other management objects increase the top service that message informing is concerned about with resource, can write down the daily record that resource increases operation such as control center; Upper-layer service also can begin as required to use of resource etc.
In addition, above-mentioned distributed resource management system is when the deletion resource board, for example resource board moves back clothes, be unavailable at first with related resource object record in the global resource information data object, if all resource objects with logic function withdraw from, also need to close relevant logic function, promptly system no longer provides the interrelated logic function.Then the resource deletion is distributed to the upper layer application that registration is concerned about by incident.At last, corresponding management object control module is changed to backed up data object and resource object unavailable, and disabled object will unified deletion after postponing a period of time.Thereby finish the process of resource board deletion.If resource board promptly recovers normally in very short time, at this moment, can attempt the object reparation earlier because the data on the backup object are not removed immediately, the maintenance system is relatively stable.
Fig. 6 is based on the process of the resource board upgrading of above distributed resource management system, mainly may further comprise the steps:
Step 601, the resource board edition upgrading begins, and the management object control module is changed to the related resource object unavailable in global resource information data object;
Step 602, control module is assigned the upgrading task to resource board;
Step 603, control module are closed resource the upper layer application that is distributed to care by incident;
Step 604 after the resource board upgrading is finished, is carried out initialization again, and this process is identical with aforementioned resource board increase process, comprises to flow processs such as the configuration of management object group request logic function, initialization, resource status report;
Step 605, management object control module are changed to latest edition with the related resource object and identify in global resource information data object;
Step 606, the management object control module message of issuing a notice is carried the upgrade status functional configuration;
Step 607, management object control module waiting system latest edition is switched, and waits for if do not switch then continue; If system switches, then notify resource object to remove the upgrading isolation;
Step 608, resource object according to the notification message in the step 606 check up-to-date functional configuration whether with this locality keep consistent, as inconsistent, then resource object also needs to reinitialize;
Step 609 after the resource board function is normal, reports resource status to the management object control module;
Step 610, the management object control module is changed to normal condition with the related resource object in global resource information data object;
Step 611, management object control module normally are distributed to resource the upper layer application of care by incident.
So far, finish the smooth upgrade process of resource board.
In sum, system and method for the present invention is owing to adopt the classification Object Management group and effectively fault-tolerant design, can realize easily that the veneer self-adaptation expands, the quick copy of function and substitute, internal resource backup and veneer smooth upgrade; And improved the efficient that resource is used, simplified the resource management flow process, and guaranteed under the situation of various mistakes takes place resource, to possess ability and the mechanism of fast quick-recovery.Compare with the traditional distributed resource management mode, have clear in structure, use characteristics simple, flexible configuration.
Claims (12)
1. a distributed resource management system is characterized in that, comprises resource object, data object, management object and control center, wherein:
Described control center is used for, the business function of all resource objects and backup mode in the configuration-system, and be all management object allocation manager tasks;
Described management object is used for, and according to the management role that described control center distributes, is responsible for the resource management of the resource object managed, and inquiry and control service at the resource object of being managed are provided;
Described resource object is used for, and finishes corresponding logic function under the control of management object;
Described data object is used for, resource information in the storage resources object and the resource control information in the management object.
2. the system as claimed in claim 1 is characterized in that, described management object also comprises:
Copy management module is used to the main resource object and main with data object alternative back-up storage district of using, and carries out the object back-up storage; After finishing the object back-up storage, set up backup object and main with the mapping relations between object, and realize backup object with main with object synchronously;
The failure management module was used for when the master was lost efficacy with data object, adopted the preliminary data object that described master is carried out data with data object and recovered.
3. system as claimed in claim 2 is characterized in that,
Described resource object also is used for, and after taking place unusual or moving back clothes, gives corresponding management object with status change notification;
Described management object also is used for, after receiving described status change notification, the state of this resource object and resource backup object thereof and Backup Data object in the global resource information data object is changed to down state, and, if all resource objects of same logic function all move back clothes, then close described logic function; And described state variation is distributed to other management objects and/or control center by incident.
4. as claim 1,2 or 3 described systems, it is characterized in that,
Described resource object is used for, and reports resource information to management object after foundation, and the configuration of request logic function; And, after carrying out initialization, resource status is reported management object according to the logic function of management object configuration;
Described management object is used for, and inquires about the business function and the backup mode of this resource object according to the resource information that resource object reports to control center, and offers this resource object; And, after receiving the resource status that reports, the resource information of this resource object is write in the data object of global resource information, and resource increased be distributed to other management objects and/or control center by incident.
5. as claim 1,2 or 3 described systems, it is characterized in that,
Described management object is used for, and disposes to control center's request management object after foundation;
Described control center is used for, and creates the record of this management object, and distributes logic function, the backup mode of the resource object that this management object manages and the maximum quantity of managing every type resource object.
6. as claim 1,2 or 3 described systems, it is characterized in that,
Described management object also is used for, when the resource object of being managed is upgraded, the state of described resource object is changed to down state in the data object of global resource information, and resource closed by incident is distributed to other management objects and/or control center; And, after the upgrading of described resource object is finished, the state of described resource object is changed to normal operating condition in the data object of global resource information.
7. a distributive resource managing method is characterized in that, system resource is divided into resource object, data object, management object and control center by function, and this method comprises:
By the business function and the backup mode of all resource objects in the described control center configuration-system, and be all management object allocation manager tasks;
The management role that described management object is distributed according to described control center, the resource management of the resource object that responsible quilt is managed, and provide the inquiry of the resource object of being managed at quilt and control to serve;
Described resource object is finished corresponding logic function under the control of management object;
Resource information and the resource control information in the management object in the resource object are stored in the described data object.
8. method as claimed in claim 7 is characterized in that, described method also comprises:
Described management object is main with resource object and main with data object alternative back-up storage district, carries out the object back-up storage; After finishing the object back-up storage, set up backup object and main with the mapping relations between object, and realize backup object with main with object synchronously;
When the master was lost efficacy with data object, described management object adopted the preliminary data object that described master is carried out the data recovery with data object.
9. method as claimed in claim 8 is characterized in that, described method also comprises:
After described resource object takes place unusual or moves back clothes, give corresponding management object with status change notification;
After receiving described status change notification, described management object is changed to down state with the state of this resource object and resource backup object thereof and Backup Data object in the global resource information data object, and, if all resource objects of same logic function all move back clothes, then close described logic function; And described state variation is distributed to other management objects and/or control center by incident.
10. as claim 7,8 or 9 described methods, it is characterized in that when increasing resource object, described method comprises:
After setting up resource object, report resource information to described management object, and the configuration of request logic function;
Described management object is inquired about the business function and the backup mode of this resource object according to the resource information that resource object reports to control center, and offers this resource object;
Described resource object reports described management object with resource status after carrying out initialization according to the logic function of described management object configuration;
Described management object writes the resource information of this resource object in the data object of global resource information after receiving the described resource status that reports, and resource increased is distributed to other management objects and/or control center by incident.
11., it is characterized in that when increasing management object, described method comprises as claim 7,8 or 9 described methods:
After setting up management object, dispose to control center's request management object;
The record of this management object is created by described control center, and distributes logic function, the backup mode of the resource object that this management object manages and the maximum quantity that can manage every type resource object.
12., it is characterized in that when described resource object was upgraded, described method comprised as claim 7,8 or 9 described methods:
When the resource object of being managed was upgraded, management object was changed to down state with the state of described resource object in the data object of global resource information, and resource closed by incident was distributed to other management objects and/or control center;
After described resource object upgrading was finished, described management object was changed to normal operating condition with the state of described resource object in the data object of global resource information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009101656394A CN101996094A (en) | 2009-08-12 | 2009-08-12 | Method and system for managing distributed resources |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009101656394A CN101996094A (en) | 2009-08-12 | 2009-08-12 | Method and system for managing distributed resources |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101996094A true CN101996094A (en) | 2011-03-30 |
Family
ID=43786279
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2009101656394A Pending CN101996094A (en) | 2009-08-12 | 2009-08-12 | Method and system for managing distributed resources |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101996094A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102185710A (en) * | 2011-04-26 | 2011-09-14 | 中兴通讯股份有限公司 | Method and system for managing veneer log |
| CN102263665A (en) * | 2011-08-15 | 2011-11-30 | 中兴通讯股份有限公司 | Method and device for processing resources, single board and equipment |
| CN102932177A (en) * | 2012-10-30 | 2013-02-13 | 东南大学 | Inter-domain distributed resource control method |
| CN103810174A (en) * | 2012-11-06 | 2014-05-21 | 阿里巴巴集团控股有限公司 | Database switching method and database switching system |
| CN103902646A (en) * | 2013-12-27 | 2014-07-02 | 北京天融信软件有限公司 | Distributed task managing system and method |
| CN104217309A (en) * | 2014-09-25 | 2014-12-17 | 中国人民解放军信息工程大学 | Method and device for managing information objects of system resources |
| CN108701132A (en) * | 2015-11-27 | 2018-10-23 | 华为技术有限公司 | Resource management system and method |
| CN108833331A (en) * | 2018-04-09 | 2018-11-16 | 海信集团有限公司 | A kind of creation of resource object, resource object state reporting method and device |
| CN109308230A (en) * | 2017-07-27 | 2019-02-05 | 华为技术有限公司 | A kind of backup method and device |
| CN110543315A (en) * | 2019-09-06 | 2019-12-06 | 程延辉 | distributed operating system of kbroker, storage medium and electronic equipment |
-
2009
- 2009-08-12 CN CN2009101656394A patent/CN101996094A/en active Pending
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102185710A (en) * | 2011-04-26 | 2011-09-14 | 中兴通讯股份有限公司 | Method and system for managing veneer log |
| CN102263665A (en) * | 2011-08-15 | 2011-11-30 | 中兴通讯股份有限公司 | Method and device for processing resources, single board and equipment |
| CN102932177A (en) * | 2012-10-30 | 2013-02-13 | 东南大学 | Inter-domain distributed resource control method |
| CN103810174A (en) * | 2012-11-06 | 2014-05-21 | 阿里巴巴集团控股有限公司 | Database switching method and database switching system |
| CN103810174B (en) * | 2012-11-06 | 2017-04-12 | 阿里巴巴集团控股有限公司 | Database switching method and database switching system |
| CN103902646A (en) * | 2013-12-27 | 2014-07-02 | 北京天融信软件有限公司 | Distributed task managing system and method |
| CN103902646B (en) * | 2013-12-27 | 2017-04-19 | 北京天融信软件有限公司 | Distributed task managing system and method |
| CN104217309A (en) * | 2014-09-25 | 2014-12-17 | 中国人民解放军信息工程大学 | Method and device for managing information objects of system resources |
| CN108701132A (en) * | 2015-11-27 | 2018-10-23 | 华为技术有限公司 | Resource management system and method |
| US11467874B2 (en) | 2015-11-27 | 2022-10-11 | Huawei Cloud Computing Technologies Co., Ltd. | System and method for resource management |
| CN109308230A (en) * | 2017-07-27 | 2019-02-05 | 华为技术有限公司 | A kind of backup method and device |
| CN109308230B (en) * | 2017-07-27 | 2020-11-17 | 华为技术有限公司 | Backup method and device |
| CN108833331A (en) * | 2018-04-09 | 2018-11-16 | 海信集团有限公司 | A kind of creation of resource object, resource object state reporting method and device |
| CN108833331B (en) * | 2018-04-09 | 2020-12-08 | 海信集团有限公司 | Resource object creating method and device and resource object state reporting method and device |
| CN110543315A (en) * | 2019-09-06 | 2019-12-06 | 程延辉 | distributed operating system of kbroker, storage medium and electronic equipment |
| CN110543315B (en) * | 2019-09-06 | 2021-08-31 | 程延辉 | Distributed operating system of kbroker, storage medium and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101996094A (en) | Method and system for managing distributed resources | |
| EP3694148B1 (en) | Configuration modification method for storage cluster, storage cluster and computer system | |
| CN101401073B (en) | Consistent updates across storage subsystems | |
| CN100412810C (en) | Method and system for automatically managing the state of replicated data of a computing environment | |
| RU2208834C2 (en) | Method and system for recovery of database integrity in system of bitslice databases without resource sharing using shared virtual discs and automated data medium for them | |
| US7197632B2 (en) | Storage system and cluster maintenance | |
| CN101667181B (en) | Method, device and system for data disaster tolerance | |
| CN103414712B (en) | A kind of distributed virtual desktop management system and method | |
| CN103647849A (en) | Method and device for migrating businesses and disaster recovery system | |
| EP2643771B1 (en) | Real time database system | |
| EP2281240A1 (en) | Maintaining data integrity in data servers across data centers | |
| CN102938705B (en) | Method for managing and switching high availability multi-machine backup routing table | |
| DE112006002531T5 (en) | Application of virtual servers for high availability solutions and disaster recovery solutions | |
| CN102710752B (en) | Calamity is for storage system | |
| CN101515296A (en) | Data updating method and device | |
| CN103345470A (en) | Database disaster tolerance method, database disaster tolerance system and server | |
| CN101980192A (en) | Object-based cluster file system management method and cluster file system | |
| CN1329829C (en) | Back-up restoring method for user database | |
| CN110727709A (en) | Cluster database system | |
| CN101114232A (en) | Method of distributing disk image in computer system | |
| CN106850269A (en) | A kind of management system of cloud platform | |
| CN102262591A (en) | Garbage collection method and system for memory copy system | |
| US9148430B2 (en) | Method of managing usage rights in a share group of servers | |
| JP2001218241A (en) | Electric exchange | |
| CN103428288B (en) | Based on the copies synchronized method of subregion state table and coordinator node |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20110330 |