CN104850394A - Management method of distributed application program and distributed system - Google Patents

Abstract

The invention provides a management method of a distributed application program and a distributed system. The management method comprises the following steps: according to the logic relationship of each process in the distributed application program, establishing a process tree of the distributed application program; determining each parent process and the subprocess of each parent process in the process tree; adding a distribution strategy of the subprocess of each parent process into each parent process; according to the process tree, creating each parent process from top to bottom, and determining target equipment for distributing the subprocess of each parent process according to a distribution strategy of the subprocess of each parent process and the resource information of the distributed system; and creating the subprocess of each parent process in the target equipment for distributing the subprocess of each parent process. The management method of the distributed application program and the distributed system can improve the development efficiency of the distributed application program when the distributed application program can be automatically distributed by the distributed system.

Description

The management method of distributed application program and distributed system

Technical field

The present invention relates to field of computer technology, particularly relate to a kind of management method and distributed system of distributed application program.

Background technology

Distributed Application is by application program distribution on a different computer, the working method jointly accomplished a task by network.Because Distributed Application can improve the travelling speed of application program, and then can promote perception and the satisfaction of user, therefore, increasing application program has all been developed to distributed application program.

At present, have many general Distributed Architecture on the market, user can develop various types of distributed application program by arbitrary general Distributed Architecture, and runs distributed application program by the distributed system corresponding to general Distributed Architecture.In order to make distributed system when running distributed application program, on the computing machine that can automatically distributed application program be distributed in a network, user needs when development DDB application program, and the resource information in conjunction with distributed system sets the distributing position of distributed application program in distributed system.

But, adopt aforesaid way realize distributed system automatically distribute distributed application program time, user needs the time of at substantial to remove setting distributed application program distributing position in a distributed system.

Summary of the invention

The invention provides a kind of management method and distributed system of distributed application program, in order to solve in prior art to enable distributed system automatically distribute distributed application program, user, when using general Distributed Architecture development DDB application program, needs the time of at substantial to go to set the problem of distributed application program distributing position in a distributed system.

For achieving the above object, embodiments of the invention adopt following technical scheme:

First aspect, the invention provides a kind of management method of distributed application program, comprising:

According to the logical relation of process each in distributed application program, set up the process tree of described distributed application program;

Determine the subprocess of each parent process in described process tree and each described parent process;

The Distribution Strategy of the subprocess of described parent process is added in each described parent process; Wherein, the Distribution Strategy of the subprocess of described parent process is used to indicate the subprocess distribution mode in a distributed system of described parent process;

Each described parent process is created successively, and according to the Distribution Strategy of the subprocess of each described parent process and the resource information of described distributed system, the target device of the subprocess of each described parent process of determining to distribute according to described process tree order from top to bottom;

The subprocess of each described parent process is created in the target device of the subprocess of each described parent process of distribution.

In conjunction with first aspect, in the first possible embodiment of first aspect, described method also comprises:

The abnormality processing strategy of the subprocess of described parent process is added in each described parent process; Wherein, the abnormality processing strategy of the subprocess of described parent process is used to indicate when the subprocess of described parent process is abnormal, to the processing mode of the subprocess of described parent process.

In conjunction with the first possible embodiment of first aspect, in the embodiment that the second of first aspect is possible, the abnormality processing strategy of the subprocess of described parent process at least comprises following one: ignore abnormality processing strategy, and reboot process strategy, exits processing policy;

Wherein, ignore abnormality processing strategy and be, the processing policy of the unexpected message that the subprocess ignoring described parent process reports described in;

Described reboot process strategy is, by the processing policy that the subprocess of the described parent process reporting unexpected message is restarted;

The described processing policy that exits is, by processing policy out of service for the subprocess of the described parent process reporting unexpected message.

In conjunction with the first possible embodiment of first aspect or the possible embodiment of the second of first aspect, in the third possible embodiment of first aspect, the abnormality processing strategy of the subprocess of described parent process is also used to indicate when the subprocess of described parent process is abnormal, to the processing rule of other subprocesss of described parent process.

In conjunction with the third possible embodiment of first aspect, in the 4th kind of possible embodiment of first aspect, the abnormality processing strategy of the subprocess of described parent process also at least comprises following one: processing rule separately, whole processing rule;

Wherein, described independent processing rule is, to the processing rule that other subprocesss of described parent process do not process; Described whole processing rule is, whole subprocesss of described parent process is used to the processing rule of same abnormality processing strategy process.

In conjunction with first aspect to any one in the possible embodiment of the second of first aspect, in the 5th kind of possible embodiment of first aspect, the Distribution Strategy of the subprocess of described parent process at least comprises following one: local Distribution Strategy, integrated distribution strategy, Arbitrary distribution strategy;

Wherein, described local Distribution Strategy is, the subprocess of described parent process and described parent process is distributed in the Distribution Strategy in same target device;

Described integrated distribution strategy is, all subprocesss of described parent process are distributed in the Distribution Strategy in same target device;

Described Arbitrary distribution strategy is, the subprocess of described parent process and described parent process is according to the Distribution Strategy of the resource information Arbitrary distribution of described distributed system.

In conjunction with first aspect to any one in the possible embodiment of the second of first aspect, in the 6th kind of possible embodiment of first aspect, described method also comprises:

Monitoring strategies is added respectively in the subprocess of each described parent process and each described parent process; Described monitoring strategies is used to indicate the monitoring relation between the subprocess of each described parent process and each described parent process.

In conjunction with the 6th kind of possible embodiment of first aspect, in the 7th kind of possible embodiment of first aspect, when creating the subprocess of each described parent process in the target device of the subprocess of each described parent process that distributes, described method also comprises:

According to described monitoring strategies, set up each described parent process and be connected with the monitoring of the subprocess of each described parent process, be connected mutually transmission periodic heartbeat message with the subprocess of described parent process by described monitoring to make described parent process.

Second aspect, the invention provides a kind of distributed system, comprising: multiple resource apparatus and distributed platform; Wherein, described distributed platform for perform as arbitrary in claim 1-8 as described in the management method of distributed application program; Described multiple resource apparatus is for running described distributed application program.

The management method of distributed application program provided by the invention and distributed system, by the multiple general Distribution Strategy preset, make user when using general Distributed Architecture to develop various types of distributed application program, all can according to the programmed logic of each distributed application program, the Distribution Strategy that the subprocess being incorporated as each parent process in each parent process in the process tree of each distributed application program is selected, make distributed system when creating each parent process of distributed application program successively according to the process tree of distributed application program order from top to bottom, can according to the resource information of the Distribution Strategy of the subprocess of the parent process in each parent process and distributed system, suitable target device is selected to create the subprocess of each parent process, distributed system is made to achieve the distributed application program that automatically distributes.With user in prior art when using general Distributed Architecture to develop each distributed application program, need the resource information in conjunction with distributed system, the time of at substantial goes to set each distributed application program distributing position in a distributed system, namely, the time of at substantial is needed to be that each distributed application program writes code, compare to realize distributed application program automatic distribution in a distributed system, the management method of the distributed application program that the embodiment of the present invention provides and distributed system, the time of User Exploitation distributed application program can be saved, improve the efficiency of distributed application program exploitation.

Accompanying drawing explanation

In order to be illustrated more clearly in the embodiment of the present invention or technical scheme of the prior art, be briefly described to the accompanying drawing used required in embodiment or description of the prior art below, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.

The process flow diagram of the management method embodiment one of the distributed application program that Fig. 1 provides for the embodiment of the present invention;

The structural representation of the process tree of the distributed application program that Fig. 2 provides for the embodiment of the present invention;

The process flow diagram of the management method embodiment two of the distributed application program that Fig. 3 provides for the embodiment of the present invention;

The process flow diagram of the management method embodiment three of the distributed application program that Fig. 4 provides for the embodiment of the present invention;

The process flow diagram of the management method embodiment four of the distributed application program that Fig. 5 provides for the embodiment of the present invention;

The subprocess of the parent process that Fig. 6 provides for the embodiment of the present invention and parent process adopts the distribution schematic diagram of local Distribution Strategy;

The subprocess of the parent process that Fig. 7 provides for the embodiment of the present invention and parent process adopts the distribution schematic diagram of integrated distribution strategy;

The subprocess of the parent process that Fig. 8 provides for the embodiment of the present invention and parent process adopts the distribution schematic diagram of Arbitrary distribution strategy;

The subprocess of the parent process that Fig. 9 provides for the embodiment of the present invention and parent process adopts integrated distribution strategy to add the distribution schematic diagram of lax Distribution Strategy;

The process flow diagram of the management method embodiment five of the distributed application program that Figure 10 provides for the embodiment of the present invention;

The process flow diagram of the management method embodiment six of the distributed application program that Figure 11 provides for the embodiment of the present invention;

The structural representation of the distributed system that Figure 12 provides for the embodiment of the present invention.

Embodiment

Below in conjunction with the accompanying drawing in the embodiment of the present invention, be clearly and completely described the technical scheme in the embodiment of the present invention, obviously, described embodiment is only the present invention's part embodiment, instead of whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art, not making the every other embodiment obtained under creative work prerequisite, belong to the scope of protection of the invention.

The embodiment of the present invention goes in the distributed system corresponding to general Distributed Architecture, and it can realize automatic management operating distributed application program in a distributed system.Therefore, the management method of the distributed application program that the embodiment of the present invention provides and distributed system, be intended to solve in prior art to enable distributed system automatically distribute distributed application program, user, when using general Distributed Architecture development DDB application program, needs the time of at substantial to go to set the problem of distributed application program distributing position in a distributed system.

It should be noted that, when distributed application program runs in a distributed system, process is distributed application program execution entity in a distributed system, and it can apply for and have the resource of distributed system.Because distributed application program is by the multiple resource apparatus executed in parallel in distributed system, so distributed application program can corresponding multiple process.Therefore, distributed application program multiple resource apparatus in a distributed system run just can be understood as distributed application program multiple processes multiple resource apparatus in a distributed system on run, connected by the network between multiple resource apparatus between process and communicate mutually.

It should be noted that, for convenience of description, in the following each embodiment of the present invention, " User Exploitation distributed application program " be acute pyogenic infection of finger tip " user uses general Distributed Architecture development DDB application program " all, and " distributed system " be acute pyogenic infection of finger tip " distributed system that general Distributed Architecture is corresponding " all.

With embodiment particularly, technical scheme of the present invention is described in detail below.These specific embodiments can be combined with each other below, may repeat no more for same or analogous concept or process in some embodiment.

The process flow diagram of the management method embodiment one of the distributed application program that Fig. 1 provides for the embodiment of the present invention.Wherein, the executive agent of the method can for having communication facilities or the communication module of management of process function in distributed system, as shown in Figure 1, the method can comprise the steps:

Step S101: according to the logical relation of process each in distributed application program, set up the process tree of distributed application program.

Concrete, in distributed application program, the logical relation of each process can be understood as each process and by the establishment relation between the process of this process creation, and be created relation between each process and the process creating this process, by above-mentioned establishment relation and be created the tree-like process creation relation that relation can set up out distributed application program, that is, the process tree of distributed application program.

The structural representation of the process tree of the distributed application program that Fig. 2 provides for the embodiment of the present invention, as shown in Figure 2, suppose that certain distributed application program has process A, B, C and D, wherein, process A creates process B and process C, process B creates process D, then, the process tree of this distributed application program can be as shown in Figure 2.

Step S102: the subprocess determining each parent process in process tree and each parent process.

Concrete, in the process tree of a distributed application program, each process is by the parent process of the process of this process creation, be then the subprocess of this parent process by the process of this process creation.Such as, continue with reference to Fig. 2, the process tree of the distributed application program according to Fig. 2, for process B and process D, because process B creates process D, then process B is parent process, and process D is the subprocess of parent process B.

In the process tree of a distributed application program, a process can be the parent process of a subprocess, it can also be the subprocess of another parent process, that is, according to the process tree order from top to bottom of distributed application program, outside the process (that is, the process of last one deck) of least significant end, all processes can be the parent process of some other processes or certain multiple process.Such as: continue with reference to Fig. 2, the process tree of the distributed application program according to Fig. 2, in the process tree of this distributed application program except the process C of end and process D, comprises two parent processes, be respectively: process A and process B.Process A is the parent process of subprocess B and subprocess C, and process B is the parent process of subprocess D; Wherein, process B is the parent process of subprocess D, and process B is also the subprocess of parent process A simultaneously.

In the process tree of a distributed application program, the quantity of the subprocess of the parent process that different parent processes is corresponding can be different.Such as, continue with reference to Fig. 2, the process tree of the distributed application program according to Fig. 2, the subprocess of parent process A has two, is respectively: process B and process C; The subprocess of parent process B only has one, i.e. process D.

Step S103: the Distribution Strategy adding the subprocess of parent process in each parent process; Wherein, the Distribution Strategy of the subprocess of parent process is used to indicate the subprocess distribution mode in a distributed system of parent process.

Concrete, the Distribution Strategy of the subprocess of the parent process described in the embodiment of the present invention can have multiple, the subprocess distribution mode in a distributed system of the corresponding a kind of parent process of each Distribution Strategy, the plurality of Distribution Strategy is general Distribution Strategy, that is, the subprocess of each parent process in the process tree of the distributed application program of arbitrary type, all can find a Distribution Strategy being applicable to oneself distribution mode in a distributed system in the plurality of Distribution Strategy.During specific implementation, user can when development DDB application program, according to the programmed logic of distributed application program, namely, communication frequency between the subprocess of each parent process in the process tree of distributed application program and other subprocesss of parent process and parent process, adds the Distribution Strategy of the subprocess of applicable parent process in each parent process.Such as, if the subprocess of parent process communicates with parent process frequently, then a kind of subprocess by parent process and parent process can be selected to be distributed in Distribution Strategy on same resource apparatus for the subprocess of this parent process, to improve the operational efficiency of distributed application program, reduce the communication overhead in distributed system.

The representation of the embodiment of the present invention to the Distribution Strategy of the subprocess of above-mentioned parent process does not limit, the parameter that such as can be able to identify for distributed system, code block corresponding to parameter is arranged in a distributed system, during specific implementation, user is when developing dissimilar distributed application program, suitable Distribution Strategy can be selected for the subprocess of each parent process in the process tree of each distributed application program, and parameter corresponding for the Distribution Strategy of the subprocess of each parent process is added each parent process, and do not need again the resource information of time in conjunction with distributed system of at substantial, for each distributed application program setting distributing position in a distributed system.

Optionally, when a certain parent process in the process tree of distributed application program has multiple subprocess, the Distribution Strategy of each subprocess of this parent process can be identical, also can be different.Such as: continue with reference to Fig. 2, the subprocess of the parent process A in the process tree of this distributed application program has two, be respectively: process B and process C, then the Distribution Strategy of subprocess B and subprocess C can be same Distribution Strategy, also can be different Distribution Strategy.

Step S104: create each parent process successively according to process tree order from top to bottom, and according to the Distribution Strategy of the subprocess of each parent process and the resource information of distributed system, the target device of the subprocess of each parent process of determining to distribute.

Concrete, the resource information of above-mentioned distributed system can be the active volume, available bandwidth, CPU occupation rate etc. of each resource apparatus in distributed system.Above-mentioned target device is the resource apparatus running distributed application program in distributed system, such as: computing machine, server, virtual machine etc.After above-mentioned distributed application program is submitted to distributed system, owing to adding the Distribution Strategy of the subprocess of parent process in each parent process in the process tree of distributed application program, then distributed system can according to the process tree of distributed application program order from top to bottom, automatic establishment first parent process started in the process tree of distributed application program, an i.e. process of most top layer in process tree, and while this parent process of establishment, according to the Distribution Strategy of the subprocess of this parent process added in this parent process, in conjunction with the resource information of distributed system, automatically to determine to distribute the target device of subprocess of this parent process.In this manner, the target device of the subprocess of all parent processes in the process tree of distributed application program can be determined.Like this, distributed system, when various types of distributed application program that run user is developed, all by the way, can determine the target device of the subprocess of each parent process in the process tree of each distributed application program automatically.

For the parameter that the Distribution Strategy of the subprocess of above-mentioned parent process can identify for distributed system, wherein, code block corresponding to parameter is arranged in a distributed system, then after above-mentioned distributed application program is submitted to distributed system, distributed system is when reading or get Distribution Strategy (i.e. the parameter) of subprocess of the parent process added in each parent process, just can call the code block of this Distribution Strategy (i.e. parameter) correspondence stored in a distributed system, and in conjunction with the resource information of distributed system, determine the target device of the subprocess of all parent processes in the process tree of distributed application program.

Optionally, when a certain parent process in the process tree of distributed application program has multiple subprocess, if the Distribution Strategy of the plurality of subprocess is identical, then the target device of the plurality of subprocess that distributes can for same.

Step S105: the subprocess creating each parent process in the target device of the subprocess of each parent process of distribution.

Concrete, when above-mentioned distributed system is according to the Distribution Strategy of subprocess of each parent process in the process tree of distributed application program and the resource information of distributed system, determine to distribute each parent process subprocess target device after, distributed system can create the subprocess of parent process automatically in the target device of correspondence, and in this manner, until all processes in process tree are created.So far, distributed system achieves the distributed application program that automatically distributes.

For the distributed application program using universal distributed Development of Framework in prior art, in order to make distributed system when running distributed application program, can automatically distributed application program be distributed on the resource apparatus of distributed system, user needs when development DDB application program, resource information in conjunction with distributed system sets the distributing position of distributed application program in distributed system, cause user when using general Distributed Architecture to develop each distributed application program, the time of at substantial is all needed to go to set each distributed application program distributing position in a distributed system, namely, the time of at substantial is needed to be that each distributed application program writes code, to realize distributed application program automatic distribution in a distributed system.And in the embodiment of the present invention, user is when using general framework development DDB application program, can according to the programmed logic of distributed application program, subprocess for each parent process in the process tree of distributed application program selects one to be applicable to oneself Distribution Strategy in general Distribution Strategy, and adds in parent process.Like this, when distributed system creates each parent process according to the order from top to bottom of the process tree of distributed application program, just can according to the resource information of the Distribution Strategy of the subprocess of the parent process in each parent process and distributed system, the resource apparatus establishment subprocess that automatic selection is suitable, makes distributed system achieve the distributed application program that automatically distributes.The embodiment of the present invention compared with prior art, user realize distributed system automatically distribute distributed application program time, do not need again the time of at substantial to go to write code, save the time of user, improve distributed application program exploitation efficiency.

The management method of the distributed application program that the embodiment of the present invention provides, by the multiple general Distribution Strategy preset, make user when using general Distributed Architecture to develop various types of distributed application program, all can according to the programmed logic of each distributed application program, the Distribution Strategy that the subprocess being incorporated as each parent process in each parent process in the process tree of each distributed application program is selected, make distributed system when creating each parent process of distributed application program successively according to the process tree of distributed application program order from top to bottom, can according to the resource information of the Distribution Strategy of the subprocess of the parent process in each parent process and distributed system, suitable target device is selected to create the subprocess of each parent process, distributed system is made to achieve the distributed application program that automatically distributes.With user in prior art when using general Distributed Architecture to develop each distributed application program, need the resource information in conjunction with distributed system, the time of at substantial goes to set each distributed application program distributing position in a distributed system, namely, the time of at substantial is needed to be that each distributed application program writes code, compare to realize distributed application program automatic distribution in a distributed system, the management method of the distributed application program that the embodiment of the present invention provides, the time of User Exploitation distributed application program can be saved, improve the efficiency of distributed application program exploitation.

Further, on above-mentioned basis embodiment illustrated in fig. 1, the process flow diagram of the management method embodiment two of the distributed application program that Fig. 3 provides for the embodiment of the present invention.Wherein, the executive agent of the method can for having communication facilities or the communication module of management of process function in distributed system, as shown in Figure 3, after step s 102 and before step S104, the method can also comprise the steps:

Step S103A: the abnormality processing strategy adding the subprocess of parent process in each parent process; Wherein, the abnormality processing strategy of the subprocess of parent process is used to indicate when the subprocess of parent process is abnormal, to the processing mode of the subprocess of parent process.

Concrete, the abnormality processing strategy of the subprocess of the parent process described in the embodiment of the present invention can have multiple, the processing mode of the corresponding a kind of subprocess to parent process of each abnormality processing strategy, the plurality of abnormality processing strategy is general Distribution Strategy, that is, to the processing mode of the exception of the subprocess of each parent process in the process tree of the distributed application program of arbitrary type, one all can be found in the plurality of abnormality processing strategy to be applicable to oneself abnormality processing strategy.During specific implementation, user can when development DDB application program, according to the programmed logic of distributed application program, namely, communication frequency between the subprocess of each parent process in the process tree of distributed application program and other subprocesss of parent process and parent process, adds the abnormality processing strategy of the subprocess of applicable parent process in each parent process.Such as, if separate operation between the subprocess of parent process and parent process, then when can be a kind of appearance of the subprocess when parent process of subprocess selection exception of this parent process, the abnormality processing strategy that the subprocess of parent process is restarted.

The representation of the embodiment of the present invention to the abnormality processing strategy of the subprocess of above-mentioned parent process does not limit, the parameter that such as can be able to identify for distributed system, code block corresponding to parameter is arranged in a distributed system, during specific implementation, user is when using general Distributed Architecture to develop dissimilar distributed application program, suitable abnormality processing strategy can be selected for the subprocess of each parent process in the process tree of each distributed application program, and parameter corresponding for the abnormality processing strategy of the subprocess of each parent process is added each parent process, and do not need to be separately each process setting abnormality processing mode of each distributed application program, namely, the time of at substantial is not needed to be the code that each process of each distributed application program writes process anomalous mode, to realize the exception of the automatic process distributed application program in distributed system.Like this, when distributed application program runs in a distributed system, if a certain subprocess occurs abnormal, subprocess reports unexpected message to the parent process of subprocess automatically, by the abnormality processing strategy of parent process according to the subprocess of the parent process added in parent process, determine the processing mode of the subprocess to parent process, and processed according to the subprocess of this processing mode to parent process by distributed system, make distributed system automatically can process the exception of distributed application program.

Further, when parent process has multiple subprocess, then the abnormality processing strategy of the subprocess of the parent process described in step S103A is also used to indicate when the subprocess of parent process is abnormal, to the processing rule of other subprocesss of parent process.

Concrete, user can when development DDB application program, according to the programmed logic of distributed application program, namely, communication frequency between the subprocess of each parent process in the process tree of distributed application program and other subprocesss of parent process, sets the processing rule of other subprocesss of parent process in the abnormality processing strategy of the subprocess of each parent process.Such as, if separate operation between other processes of the subprocess of parent process and parent process, the subprocess of parent process is processed to the operation of other subprocesss not affecting parent process, then can select when the subprocess of this parent process is abnormal in the exception of the subprocess of parent process is tactful, only the subprocess of this parent process is processed, not to the processing rule that other subprocesss of parent process process.Like this, when distributed application program runs in a distributed system, if when the exception of subprocess is reported the parent process of subprocess by subprocess, parent process then can according to the processing rule to other subprocesss of parent process in the abnormality processing strategy of the subprocess of the parent process added in parent process, while determining the processing mode to the subprocess of parent process, determine the processing mode of other subprocesss to parent process, and processed according to this processing mode other subprocesss to parent process by distributed system, make distributed system automatically can process the exception of distributed application program, distributed application program is normally run.

Same, the representation of above-mentioned processing rule can with reference to the description of the above-mentioned representation to Distribution Strategy and abnormality processing strategy, and the embodiment of the present invention does not repeat them here.

It should be noted that, the execution sequence of the embodiment of the present invention to step S103 and step S103A does not limit, and can first perform step S103 and perform step S103A again, also can first perform step S103A and perform step S103 again.

The management method of the distributed application program that the embodiment of the present invention provides, by the multiple general abnormality processing strategy preset, make user when using general Distributed Architecture to develop various types of distributed application program, all can according to the programmed logic of each distributed application program, the abnormality processing strategy that the subprocess being incorporated as each parent process in each parent process in the process tree of each distributed application program is selected, make distributed system when running distributed application program, when the subprocess of distributed application program occurs abnormal, the parent process of this subprocess can according to the abnormality processing strategy of the subprocess of parent process, determine the processing rule of the processing mode of the subprocess of parent process and other subprocesss of parent process, and transfer to distributed system automatically to process.With user in prior art when using general Distributed Architecture to develop each distributed application program, in order to the exception making distributed system automatically can process distributed application program, the time of at substantial is needed to be that the code that each process of each distributed application program writes abnormality processing is compared, the management method of the distributed application program that the embodiment of the present invention provides, the time of User Exploitation distributed application program can be saved further, improve the efficiency of further distributed application program exploitation.

Further, on above-mentioned basis embodiment illustrated in fig. 3, the process flow diagram of the management method embodiment three of the distributed application program that Fig. 4 provides for the embodiment of the present invention.Wherein, the executive agent of the method can for having communication facilities or the communication module of management of process function in distributed system, as shown in Figure 4, after step s 102 and before step S104, the method can also comprise the steps:

Step S103B: add monitoring strategies respectively in the subprocess of each parent process and each parent process; Monitoring strategies is used to indicate the monitoring relation between the subprocess of each parent process and each parent process.

Concrete, when distributed application program a certain process residing for resource apparatus occur abnormal time, such as: resource apparatus power down etc., then process can be automatically out of service, in order to this exception making the parent process of distributed application program can find and process the subprocess of parent process fast, the embodiment of the present invention is when development DDB application program, monitoring strategies is added in each parent process in the process tree of distributed application program and the subprocess of each parent process, to make distribution application system when running this distributed application program, monitoring relation between the subprocess just can determining each parent process and each parent process according to this monitoring strategies.It should be noted that, the mode of " when development DDB application program; add monitoring strategies in each parent process in the process tree of distributed application program and the subprocess of each parent process " can adopt the implementation that in prior art, those skilled in the art adopt, and the embodiment of the present invention does not limit this.

Then further, after above-mentioned steps S105, the method can also comprise the steps:

Step S106: according to monitoring strategies, sets up each parent process and is connected with the monitoring of the subprocess of each parent process, is connected the periodic heartbeat message of transmission mutually to make parent process with the subprocess of parent process by monitoring.

Concrete, when distributed application program is run in a distributed system, when distributed system creates the subprocess of each parent process in the target device of the subprocess of each parent process that distributes, according to the monitoring strategies added in parent process and subprocess, automatically the monitoring between the subprocess setting up each parent process and each parent process connects, and after foundation monitoring connects, notify that parent process and subprocess are connected by monitoring and send periodic heartbeat message mutually.When parent process because when extremely exiting, the subprocess of parent process also can pass through the periodic heartbeat message not receiving parent process transmission in Preset Time, after determining that parent process exits, out of service voluntarily.When the resource apparatus residing for subprocess occurs extremely to cause subprocess to exit, by not receiving the periodic heartbeat information of the subprocess transmission of parent process in Preset Time, parent process can initiatively find that the subprocess of parent process exits.When the subprocess of parent process determination parent process exits, parent process can according to the abnormality processing strategy of the subprocess of the parent process added in parent process, determine the processing mode of the subprocess to parent process, and processed according to the subprocess of this processing mode to parent process by distributed system.Above-mentioned monitoring connects and can connect for TCP, and the mode that the subprocess of above-mentioned parent process and parent process sends mutually periodic heartbeat message can be send heartbeat message mutually by the IP address of the subprocess of parent process and parent process.The transmission cycle of above-mentioned heartbeat message and above-mentioned Preset Time all can according to the requirements set of user.

Optionally, first parent process in the process tree of above-mentioned distributed application program, namely in process tree, a process of most top layer can set and set up with distributed system the monitoring strategies monitored and be connected, make distributed system can monitor the exception of first parent process, thus the arbitrary process that ensure that in the process tree of distributed application program occur any abnormal time, can be found rapidly by parent process, and the abnormality processing strategy of subprocess according to the parent process added in parent process, determine the subprocess processing mode of parent process, and processed according to the subprocess of this processing mode to parent process by distributed system.

By the way, make each parent process in the process tree of distributed application program can find the exception of the subprocess of parent process fast, and the abnormality processing strategy of subprocess according to the parent process added in parent process, determine the processing mode of the subprocess to parent process, and processed according to the subprocess of this processing mode to parent process by distributed system, make distributed system automatically and timely can process the exception of distributed application program, ensure that the normal operation of distributed application program.

It should be noted that, the execution sequence of the embodiment of the present invention to step S103, step S103A and step S103B does not limit.

The management method of the distributed application program that the embodiment of the present invention provides, can in the time saving User Exploitation distributed application program further, improve on the basis of the efficiency of further distributed application program exploitation, make distributed system automatically and timely can process the exception of distributed application program, ensure that the normal operation of distributed application program, the accuracy of the result of calculation that improve distributed application program operation or the service being supplied to client.

Adopt several specific embodiment below, the technical scheme of embodiment of the method shown in Fig. 1, Fig. 3 and Fig. 4 is described in detail.

The process flow diagram of the management method embodiment four of the distributed application program that Fig. 5 provides for the embodiment of the present invention.Wherein, the executive agent of the method can for having communication facilities or the communication module of management of process function in distributed system.As shown in Figure 5, the method can comprise the steps:

Step S501: according to the logical relation of process each in distributed application program, set up the process tree of distributed application program.

Wherein, the associated description of this step see the description of above-mentioned steps S101, can not repeat them here.

Step S502: the subprocess determining each parent process in the process tree of distributed application program and each parent process.

Wherein, the associated description of this step see the description of above-mentioned steps S102, can not repeat them here.

Step S503: the Distribution Strategy adding the subprocess of parent process in each parent process in the process tree of distributed application program.

As the above-mentioned embodiment of the present invention described by the management method embodiment one of distributed application program that provides, the Distribution Strategy of the subprocess of parent process can have multiple, the subprocess distribution mode in a distributed system of the corresponding a kind of parent process of each Distribution Strategy.Optionally, the Distribution Strategy of the subprocess of above-mentioned parent process can at least comprise following one: local Distribution Strategy, integrated distribution strategy, Arbitrary distribution strategy.

Wherein, the subprocess of the parent process that Fig. 6 provides for the embodiment of the present invention and parent process adopts the distribution schematic diagram of local Distribution Strategy, and as shown in Figure 6, local Distribution Strategy is the Distribution Strategy that the subprocess of parent process and parent process is distributed in same target device.When communicating frequent between parent process and the subprocess of parent process, this Distribution Strategy can be selected, to make the operational efficiency of distributed application program higher for the subprocess of this parent process.Exemplary, Web service application program is in a distributed manner example, due to such distributed application program for user provides dissimilar service, data dependence is there is (namely between process due to these dissimilar service correspondences, process needs the data using another process to produce in operational process), then can select local Distribution Strategy for the subprocess of each parent process in the process tree of such distributed application program.

The subprocess of the parent process that Fig. 7 provides for the embodiment of the present invention and parent process adopts the distribution schematic diagram of integrated distribution strategy, and as shown in Figure 7, integrated distribution strategy is, all subprocesss of parent process are distributed in the Distribution Strategy in same target device.When communication between all subprocesss of parent process is frequent, this Distribution Strategy can be selected, to make the operational efficiency of distributed application program higher for the subprocess of this parent process.Exemplary, web game program is example in a distributed manner, because such distributed application program has multiple for different user service processes in the Same Scene of web game, and between this multiple process, communication is frequent, then can be Distribution Strategy in the subprocess choice set of each parent process in the process tree of such distributed application program.

The subprocess of the parent process that Fig. 8 provides for the embodiment of the present invention and parent process adopts the distribution schematic diagram of Arbitrary distribution strategy, as shown in Figure 8, Arbitrary distribution strategy is, the subprocess of parent process and parent process is according to the Distribution Strategy of the resource information Arbitrary distribution of distributed system.When parent process and operation between the subprocess of parent process separate, when communicating less, and operation between each subprocess of parent process is also separate, when communicating less, this Distribution Strategy can be selected, to make the utilization that the resource of distributed system can be balanced for the subprocess of this parent process.Exemplary, data processor is example in a distributed manner, because such distributed application program needs the process utilizing a large amount of parallel running to carry out numerical evaluation simultaneously, to improve the computing velocity of distributed application program, wherein, the equal independent operating of each process, it can be then the subprocess selection Arbitrary distribution strategy of each parent process in the process tree of such distributed application program, to improve the utilization factor of the computational resource of distributed system substantially, shorten the working time of distributed application program.

When the Distribution Strategy that the subprocess for parent process is selected be local Distribution Strategy or integrated distribution strategy time, further alternative, lax Distribution Strategy can also be added on the basis of local Distribution Strategy or integrated distribution strategy.When this slackness Distribution Strategy is used to indicate the target device overload when the subprocess of the parent process that distributes, can the process also do not created on the target device be distributed on other resource apparatus, with the load of balancing resource equipment.Exemplary, lax Distribution Strategy is added for integrated distribution strategy, the subprocess of the parent process that Fig. 9 provides for the embodiment of the present invention and parent process adopts integrated distribution strategy to add the distribution schematic diagram of lax Distribution Strategy, as shown in Figure 9, when all subprocesss should be distributed in same resource apparatus, if this resource apparatus is load too high during the above-mentioned subprocess of establishment, then the subprocess be not also created can be distributed on other resource apparatus of same frame; If all resource apparatus in the frame that this resource apparatus is corresponding all transship, then the subprocess be not also created can be distributed in the resource apparatus in other frames.

Further alternative, the representation of the embodiment of the present invention to the Distribution Strategy of the subprocess of above-mentioned parent process does not limit, the parameter that such as can be able to identify for distributed system, and the code block corresponding to parameter is arranged in a distributed system.When the parameter that Distribution Strategy adopts distributed system to identify, specifically can with reference to following example, wherein, parameter corresponding to local Distribution Strategy is: Strategy is local distribution, the parameter that integrated distribution strategy is corresponding is: Strategy is integrated distribution, the parameter that Arbitrary distribution strategy is corresponding is: Strategy is Arbitrary distribution, then code block corresponding to the parameter arranged in distributed system specifically can be as follows:

Step S504: the abnormality processing strategy adding the subprocess of parent process in each parent process in the process tree of distributed application program.

As the above-mentioned embodiment of the present invention described by the management method embodiment two of distributed application program that provides, the abnormality processing strategy of the subprocess of parent process can have multiple, the processing mode of the corresponding a kind of subprocess to parent process of each abnormality processing strategy.Optionally, the abnormality processing strategy of the subprocess of above-mentioned parent process can at least comprise following one: ignore abnormality processing strategy, reboot process strategy, exits processing policy.

Wherein, ignoring abnormality processing strategy is, the processing policy of the unexpected message that the subprocess ignoring parent process reports.When the exception of the subprocess of parent process is negligible on the impact of the operation of the subprocess of parent process, this abnormality processing strategy can be selected for the subprocess of this parent process.Exemplary, to calculate extremely (such as: divisor is for zero, computing spilling etc.) be example, owing to calculating, exception is usual to be caused by distributed application program itself, also cannot avoid occurring calculating extremely even if restart the subprocess of parent process, and this calculating can't cause the subprocess of parent process to run extremely, then can select to ignore abnormality processing strategy for the subprocess of parent process.

Reboot process strategy is, by the processing policy that the subprocess of the parent process reporting unexpected message is restarted.When the abnormal of subprocess of parent process has impact to the operation of the subprocess of this parent process, but during by the operation of restarting the subprocess that can recover this parent process, this abnormality processing strategy can be selected for the subprocess of this parent process.Wherein, above-mentioned subprocess of restarting parent process is that distributed system is again according to the Distribution Strategy of the subprocess of the parent process added in parent process, in conjunction with the resource information of distributed system, redefine the target device of the subprocess of distribution parent process, and on the target device redefined, create the subprocess of this parent process.Exemplary, occur that hardware fault (such as with the resource apparatus residing for the subprocess of parent process, the low memory of resource apparatus, the disk failures etc. of resource apparatus), because the exception of the subprocess of the parent process hardware fault that is resource apparatus causes, can select to re-create this subprocess on new resource apparatus when can avoid again occurring this exception, then can be that the subprocess of parent process selects reboot process strategy.

Exiting processing policy is, by processing policy out of service for the subprocess of the parent process reporting unexpected message.When the abnormal impact caused the subprocess of parent process of the subprocess of parent process, when also cannot be recovered by the subprocess of restarting parent process, this abnormality processing strategy can be selected for the subprocess of this parent process.Exemplary, the compulsory withdrawal of formula system is distributed for the subprocess of parent process, cause for distributed system because such is abnormal, even if restart the subprocess of this parent process, also cannot avoid again being distributed the compulsory withdrawal of formula system, then can select to exit processing policy for the subprocess of this parent process.

Further alternative, when parent process cannot process the unexpected message of subprocess active reporting of parent process, parent process can operation suspension, and this unexpected message is reported the parent process of parent process, is processed by the parent process of this parent process.

When parent process has multiple subprocess, then further alternative, the abnormality processing strategy of the subprocess of above-mentioned parent process also at least comprises following one: processing rule separately, whole processing rule.

Wherein, independent processing rule is, to the processing rule that other subprocesss of parent process do not process.When processing some subprocesss that exception occurs of parent process, when not affecting the operation of other subprocesss of parent process, this processing rule can be selected for other subprocesss of parent process.

Whole processing rule is, whole subprocesss of parent process is used to the processing rule of same abnormality processing strategy process.When abnormal subprocess occurring parent process some and processing, during all cannot normally the running of other subprocesss of parent process, this processing rule can be selected for other subprocesss of parent process.

Step S505: add monitoring strategies respectively in each parent process in the process tree of distributed application program and the subprocess of each parent process.

Wherein, the associated description of this step see the description of above-mentioned steps S103B, can not repeat them here.

It should be noted that, the execution sequence of the embodiment of the present invention to step S503, step S504 and step S505 does not limit.

Step S506: distributed system, according to the process tree of distributed application program order from top to bottom, creates first parent process.

Wherein, the associated description of this step see the description of above-mentioned steps S104, can not repeat them here.

Step S507: the Distribution Strategy of distributed system according to the subprocess of first parent process and the resource information of distributed system, determines the target device of the subprocess of distribution first parent process.

Wherein, the associated description of this step see the description of above-mentioned steps S104, can not repeat them here.

Step S508: distributed system creates the subprocess of first parent process in the target device of the subprocess of distribution first parent process.

Wherein, the associated description of this step see the description of above-mentioned steps S105, can not repeat them here.

Step S509: distributed system is set up first parent process and is connected with the monitoring of the subprocess of first parent process.

Wherein, the associated description of this step see the description of above-mentioned steps S106, can not repeat them here.

Step S510: distributed system notifies that first parent process is connected the periodic heartbeat message of transmission mutually with the subprocess of first parent process by monitoring.

Wherein, the associated description of this step see the description of above-mentioned steps S106, can not repeat them here.

Step S511: distributed system is according to the process tree of distributed application program order from top to bottom, continue to use the mode of step S507 to step S510, create the subprocess of each parent process in the process tree of distributed application program and each parent process successively, until all create complete in a distributed system by all processes in the process tree of distributed application program.

So far, distributed system achieves the distributed application program that automatically distributes.

The management method of the distributed application program that the embodiment of the present invention provides, by the multiple general Distribution Strategy preset, make user when using general Distributed Architecture to develop various types of distributed application program, all can according to the programmed logic of each distributed application program, the Distribution Strategy that the subprocess being incorporated as each parent process in each parent process in the process tree of each distributed application program is selected, make distributed system when creating each parent process of distributed application program successively according to the process tree of distributed application program order from top to bottom, can according to the resource information of the Distribution Strategy of the subprocess of the parent process in each parent process and distributed system, suitable target device is selected to create the subprocess of each parent process, distributed system is made to achieve the distributed application program that automatically distributes.With user in prior art when using general Distributed Architecture to develop each distributed application program, need the resource information in conjunction with distributed system, the time of at substantial goes to set each distributed application program distributing position in a distributed system, namely, the time of at substantial is needed to be that each distributed application program writes code, compare to realize distributed application program automatic distribution in a distributed system, the management method of the distributed application program that the embodiment of the present invention provides, the time of User Exploitation distributed application program can be saved, improve the efficiency of distributed application program exploitation.

Further, on above-mentioned basis embodiment illustrated in fig. 5, the process flow diagram of the management method embodiment five of the distributed application program that Figure 10 provides for the embodiment of the present invention, as shown in Figure 10, the application scenarios of the method is that the subprocess of a certain parent process in the process tree of a certain distributed application program run in distributed system occurs abnormal, active reporting is to the management method of distributed application program during parent process, wherein, the executive agent of the method can for having communication facilities or the communication module of management of process function in distributed system, then can also comprise the steps: after step S512

Step S1001: subprocess is connected initiatively to the parent process transmission unexpected message of this subprocess by monitoring.

Step S1002: parent process is after receiving the unexpected message of subprocess transmission of parent process, according to the abnormality processing strategy of the subprocess of the parent process added in parent process, determine the processing mode of the subprocess to parent process, and the processing rule of other subprocesss to parent process.

Step S1003: the processing mode of the subprocess to parent process that parent process will be determined, and distributed system is sent to the processing rule of other subprocesss of parent process.

Step S1004: distributed system processes according to the subprocess of processing mode to parent process of the subprocess of the parent process received, processes according to processing rule other subprocesss to parent process of other subprocesss of the parent process received simultaneously.

So far, distributed system achieves any exception of process distributed application program active reporting automatically.

The management method of the distributed application program that the embodiment of the present invention provides, by the multiple general abnormality processing strategy preset, make user when using general Distributed Architecture to develop various types of distributed application program, all can according to the programmed logic of each distributed application program, the abnormality processing strategy that the subprocess being incorporated as each parent process in each parent process in the process tree of each distributed application program is selected, make distributed system when running distributed application program, when the subprocess of distributed application program occurs abnormal, the parent process of this subprocess can according to the abnormality processing strategy of the subprocess of parent process, determine the processing rule of the processing mode of the subprocess of parent process and other subprocesss of parent process, and transfer to distributed system automatically to process.With user in prior art when using general Distributed Architecture to develop each distributed application program, in order to the exception making distributed system automatically can process distributed application program, the time of at substantial is needed to be that the code that each process of each distributed application program writes abnormality processing is compared, the management method of the distributed application program that the embodiment of the present invention provides, the time of User Exploitation distributed application program can be saved further, improve the efficiency of further distributed application program exploitation.

Further, on above-mentioned Fig. 5 and basis embodiment illustrated in fig. 10, the process flow diagram of the management method embodiment six of the distributed application program that Figure 11 provides for the embodiment of the present invention, as shown in figure 11, the application scenarios of the method be subprocess that a certain parent process in the process tree of a certain distributed application program run in distributed system does not receive this parent process in Preset Time send periodic heartbeat message time the management method of distributed application program, wherein, the executive agent of the method can for having communication facilities or the communication module of management of process function in distributed system, then can also comprise the steps: after step S512

Step S1101: the periodic heartbeat message that parent process is sent by the subprocess not receiving parent process in Preset Time, confirms that the subprocess of parent process exits.

Step S1102: parent process, after confirming that the subprocess of parent process exits, according to the abnormality processing strategy of the subprocess of the parent process added in parent process, determines the processing mode of the subprocess to parent process, and the processing rule of other subprocesss to parent process.

Step S1103: the processing mode of the subprocess to parent process that parent process will be determined, and distributed system is sent to the processing rule of other subprocesss of parent process.

Step S1104: distributed system processes according to the subprocess of processing mode to parent process of the subprocess of the parent process received, processes according to processing rule other subprocesss to parent process of other subprocesss of the parent process received simultaneously.

So far, distributed system achieves any exception of process distributed application program automatically.

The management method of the distributed application program that the embodiment of the present invention provides, can in the time saving User Exploitation distributed application program further, improve on the basis of the efficiency of further distributed application program exploitation, make distributed system automatically and timely can process the exception of distributed application program, ensure that the normal operation of distributed application program, the accuracy of the result of calculation that improve distributed application program operation or the service being supplied to client.

One of ordinary skill in the art will appreciate that: all or part of step realizing above-mentioned each embodiment of the method can have been come by the hardware that programmed instruction is relevant.Aforesaid program can be stored in a computer read/write memory medium.This program, when performing, performs the step comprising above-mentioned each embodiment of the method; And aforesaid storage medium comprises: ROM, RAM, magnetic disc or CD etc. various can be program code stored medium.

The structural representation of the distributed system that Figure 12 provides for the embodiment of the present invention.As shown in figure 12, this distributed system comprises: multiple resource apparatus 1201 and distributed platform 1202.

Wherein, distributed platform 1202 for perform as arbitrary in claim 1-8 as described in the management method of distributed application program; Multiple resource apparatus 1201 is for running distributed application program.

Concrete, above-mentioned distributed platform can for having communication facilities or the communication module of management of process function in distributed system, above-mentioned resource apparatus can be computing machine, server, virtual machine etc.The distributed system that the embodiment of the present invention provides, may be used for the technical scheme performing either method embodiment shown in Fig. 1, Fig. 3, Fig. 4, Fig. 5, Figure 10 and Figure 11, it realizes principle and technique effect is similar, repeats no more herein.

Last it is noted that above each embodiment is only in order to illustrate technical scheme of the present invention, be not intended to limit; Although with reference to foregoing embodiments to invention has been detailed description, those of ordinary skill in the art is to be understood that: it still can be modified to the technical scheme described in foregoing embodiments, or carries out equivalent replacement to wherein some or all of technical characteristic; And these amendments or replacement, do not make the essence of appropriate technical solution depart from the scope of various embodiments of the present invention technical scheme.

Claims (9)

1. a management method for distributed application program, is characterized in that, comprising:

According to the logical relation of process each in distributed application program, set up the process tree of described distributed application program;

Determine the subprocess of each parent process in described process tree and each described parent process;

The Distribution Strategy of the subprocess of described parent process is added in each described parent process; Wherein, the Distribution Strategy of the subprocess of described parent process is used to indicate the subprocess distribution mode in a distributed system of described parent process;

Each described parent process is created successively, and according to the Distribution Strategy of the subprocess of each described parent process and the resource information of described distributed system, the target device of the subprocess of each described parent process of determining to distribute according to described process tree order from top to bottom;

The subprocess of each described parent process is created in the target device of the subprocess of each described parent process of distribution.

2. the management method of distributed application program according to claim 1, is characterized in that, described method also comprises:

The abnormality processing strategy of the subprocess of described parent process is added in each described parent process; Wherein, the abnormality processing strategy of the subprocess of described parent process is used to indicate when the subprocess of described parent process is abnormal, to the processing mode of the subprocess of described parent process.

3. the management method of distributed application program according to claim 2, is characterized in that, the abnormality processing strategy of the subprocess of described parent process at least comprises following one: ignore abnormality processing strategy, and reboot process strategy, exits processing policy;

Wherein, ignore abnormality processing strategy and be, the processing policy of the unexpected message that the subprocess ignoring described parent process reports described in;

Described reboot process strategy is, by the processing policy that the subprocess of the described parent process reporting unexpected message is restarted;

The described processing policy that exits is, by processing policy out of service for the subprocess of the described parent process reporting unexpected message.

4. the management method of the distributed application program according to Claims 2 or 3, it is characterized in that, the abnormality processing strategy of the subprocess of described parent process is also used to indicate when the subprocess of described parent process is abnormal, to the processing rule of other subprocesss of described parent process.

5. the management method of distributed application program according to claim 4, is characterized in that, the abnormality processing strategy of the subprocess of described parent process also at least comprises following one: processing rule separately, whole processing rule;

Wherein, described independent processing rule is, to the processing rule that other subprocesss of described parent process do not process; Described whole processing rule is, whole subprocesss of described parent process is used to the processing rule of same abnormality processing strategy process.

6. the management method of the distributed application program according to any one of claim 1-3, is characterized in that, the Distribution Strategy of the subprocess of described parent process at least comprises following one: local Distribution Strategy, integrated distribution strategy, Arbitrary distribution strategy;

Wherein, described local Distribution Strategy is, the subprocess of described parent process and described parent process is distributed in the Distribution Strategy in same target device;

Described integrated distribution strategy is, all subprocesss of described parent process are distributed in the Distribution Strategy in same target device;

Described Arbitrary distribution strategy is, the subprocess of described parent process and described parent process is according to the Distribution Strategy of the resource information Arbitrary distribution of described distributed system.

7. the management method of the distributed application program according to any one of claim 1-3, is characterized in that, described method also comprises:

Monitoring strategies is added respectively in the subprocess of each described parent process and each described parent process; Described monitoring strategies is used to indicate the monitoring relation between the subprocess of each described parent process and each described parent process.

8. the management method of distributed application program according to claim 7, is characterized in that, when creating the subprocess of each described parent process in the target device of the subprocess of each described parent process that distributes, described method also comprises:

According to described monitoring strategies, set up each described parent process and be connected with the monitoring of the subprocess of each described parent process, be connected mutually transmission periodic heartbeat message with the subprocess of described parent process by described monitoring to make described parent process.

9. a distributed system, is characterized in that, comprising: multiple resource apparatus and distributed platform; Wherein, described distributed platform for perform as arbitrary in claim 1-8 as described in the management method of distributed application program; Described multiple resource apparatus is for running described distributed application program.