CN109582459A - The method and device that the trustship process of application is migrated - Google Patents
The method and device that the trustship process of application is migrated Download PDFInfo
- Publication number
- CN109582459A CN109582459A CN201710908817.2A CN201710908817A CN109582459A CN 109582459 A CN109582459 A CN 109582459A CN 201710908817 A CN201710908817 A CN 201710908817A CN 109582459 A CN109582459 A CN 109582459A
- Authority
- CN
- China
- Prior art keywords
- trustship
- trustship process
- machine node
- migration
- resource
- 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
- 238000000034 method Methods 0.000 title claims abstract description 451
- 230000008569 process Effects 0.000 title claims abstract description 416
- 238000009826 distribution Methods 0.000 claims abstract description 30
- 230000005012 migration Effects 0.000 claims description 160
- 238000013508 migration Methods 0.000 claims description 158
- 238000003860 storage Methods 0.000 claims description 34
- 230000015654 memory Effects 0.000 claims description 29
- 230000001360 synchronised effect Effects 0.000 claims description 24
- 230000002159 abnormal effect Effects 0.000 claims description 20
- 238000004891 communication Methods 0.000 claims description 11
- 230000001960 triggered effect Effects 0.000 claims description 11
- 230000006378 damage Effects 0.000 claims description 6
- 238000012360 testing method Methods 0.000 claims description 5
- 238000007599 discharging Methods 0.000 claims description 4
- 238000007689 inspection Methods 0.000 claims description 4
- FFBHFFJDDLITSX-UHFFFAOYSA-N benzyl N-[2-hydroxy-4-(3-oxomorpholin-4-yl)phenyl]carbamate Chemical compound OC1=C(NC(=O)OCC2=CC=CC=C2)C=CC(=C1)N1CCOCC1=O FFBHFFJDDLITSX-UHFFFAOYSA-N 0.000 description 16
- 238000010586 diagram Methods 0.000 description 16
- 230000007704 transition Effects 0.000 description 7
- 238000012790 confirmation Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 238000001514 detection method Methods 0.000 description 3
- 230000009471 action Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 2
- 210000001320 hippocampus Anatomy 0.000 description 2
- 238000011897 real-time detection Methods 0.000 description 2
- 238000011084 recovery Methods 0.000 description 2
- 238000004064 recycling Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 238000005096 rolling process Methods 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 239000002699 waste material Substances 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000004883 computer application Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003032 molecular docking Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000011112 process operation Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5083—Techniques for rebalancing the load in a distributed system
- G06F9/5088—Techniques for rebalancing the load in a distributed system involving task migration
Abstract
A kind of method and device that the trustship process of application is migrated, comprising: be that the trustship process distributes new resources in purpose machine node when determination migrates the trustship process of the application of current machine node；In the target machine node, the new resources based on distribution start the trustship process；The trustship process discharges the resource that the trustship process is occupied on the current machine node after the target machine node starts successfully.The application at least can be avoided old resource and discharge and new resources applications the case where not getting off, and the stability of hosts applications can be improved.
Description
Technical field
The present invention relates to the methods that the trustship process of computer application technology more particularly to a kind of application is migrated
And device.
Background technique
One large-scale system of distributed resource scheduling usually manages thousands of machine nodes, each machine section
Point can be used as a resource node and provide physical resource.User can by system of distributed resource scheduling provide agreement or visitor
Family end is interacted with cluster, completes automatically dispose/operation work of resource bid and application.However since cluster is advised
Mould is huge, and the application type run on cluster is complicated, and quantity is various, it will usually due to resource isolation is not perfect, machine delay machine,
Hardware damage, network jitter and application self problem etc. reasons and cause using being operating abnormally, this is just needed the application
Trustship process rapidly move in cluster in other normal machines nodes.
Currently, the migration of trustship process is the resource release for first occupying trustship process, then new for the trustship process application
Resource simultaneously redeploys.Abnormal process has occupied after resource is released, not can guarantee current cluster have new resources for distribution and
It uses, it is likely that it will appear old resource and discharged and new resources the case where can not applying for various reasons, and one
There is such case in denier, is equivalent to user's application and is forced capacity reducing, not only increases the cost that user uses cluster, also will cause line
Upper service disruption.
In addition, the migration of trustship process is either user's triggering or is that resource scheduling system actively touches in the related technology
The case where hair, triggering mode is single, and certain applications are operating abnormally can not perceive in time, and the migration of trustship process is caused to lag, from
And cause application service unavailable for a long time, seriously affect user experience.
Summary of the invention
One of the technical issues of the application aims to solve at least in the related technology.
The application provides the method and device that a kind of trustship process of application is migrated, and at least can be avoided old resource
Release and new resources apply for the case where not getting off.
The application adopts the following technical scheme that
A kind of method that the trustship process of application is migrated, comprising:
It is the trustship in purpose machine node when determination migrates the trustship process of the application of current machine node
Process distributes new resources；
In the target machine node, the new resources based on distribution start the trustship process；
The trustship process discharges the trustship process in the current machine after the target machine node starts successfully
The resource occupied on device node.
Wherein, described before purpose machine node is trustship process distribution new resources, further includes: the current machine
When device nodal test is to own physical resource exception, the migration of the trustship process is triggered；Disappeared according to the migration from client
Breath triggers the migration of trustship process, the mark of the trustship process is carried in the migration message.
Wherein, the physical resource includes one or more following extremely:
Delay machine；
Network jitter；
Hardware damage.
Wherein, before the new resources based on distribution start the trustship process, further includes: by the current machine node
The local ephemeral data of the upper trustship process is synchronized to the target machine node.
Wherein, the local ephemeral data of trustship process described on the current machine node is synchronized to the target machine
Node, comprising: the local ephemeral data of trustship process described on the current machine node is copied to shared storage file system
System, then the local ephemeral data of the trustship process is copied to the target machine section from the shared storage file system
Point.
Wherein, after the resource that the release trustship process is occupied on current machine node, further includes: will be described
The local ephemeral data of the trustship process is deleted on current machine node.
Wherein, after the new resources starting trustship process based on distribution, further includes: the trustship process is in institute
When stating the starting failure of target machine node, automatic rollback.
Wherein, the automatic rollback includes one or more following:
The resource for keeping the trustship process to occupy on current machine node；
Discharge the new resources；
Delete data relevant to the trustship process on the target machine node.
A kind of device that the trustship process of application is migrated, comprising:
Distribution module, when being migrated for determining to the trustship process of the application of current machine node, in purpose machine
Node is that the trustship process distributes new resources；
Starting module, for starting the trustship process based on the new resources of distribution in the target machine node；
Release module, for after the target machine node starts successfully, discharging the trustship in the trustship process
The resource that process is occupied on the current machine node.
Wherein, further includes: trigger module, for when the current machine nodal test is to own physical resource exception,
Trigger the migration of the trustship process；And for being disappeared when receiving the migration message from client according to the migration
Breath triggers the migration of trustship process, and the mark of the trustship process is carried in the migration message.
Wherein, further includes: synchronization module, for by the local nonce of trustship process described on the current machine node
According to being synchronized to the target machine node.
Wherein, further includes: roll-back module is used in the trustship process when target machine node starts failure, automatically
Rollback.
A kind of resource manager in system of distributed resource scheduling, comprising:
It is stored with memory of the trustship process across machine migrator；
Processor is configured to read across the machine migrator of trustship process to execute operations described below:
It is the trustship in purpose machine node when determination migrates the trustship process of the application of current machine node
Process distributes new resources；
In the target machine node, the new resources based on distribution start the trustship process；
The trustship process discharges the trustship process in the current machine after the target machine node starts successfully
The resource occupied on device node.
Wherein, further includes: be configured to the first communication unit communicated with client；
The processor is additionally configured to read across the machine migrator of trustship process to execute operations described below:
According to the exception information that the current machine node reports, the migration of the trustship inspection, the abnormal letter are triggered
Breath indicates that the current machine node physical resource is abnormal；
The migration message from client is received by first communication unit, and phase is triggered according to the migration message
The migration of trustship process is answered, the mark of the trustship process is carried in the migration message.
Wherein, the processor is additionally configured to read across the machine migrator of trustship process to execute operations described below:
Before new resources based on distribution start the trustship process, the local of trustship process described on the current machine node is faced
When data be synchronized to the target machine node.
Wherein, the processor is additionally configured to read across the machine migrator of trustship process to execute operations described below: institute
State the automatic rollback when target machine node starts failure of trustship process.
A kind of method that the trustship process of application is migrated, comprising:
Migration message is sent to resource scheduling system according to user's operation, to trigger the migration of trustship process, the migration
The mark of the trustship process is carried in message.
A kind of client, comprising:
It is configured to the second communication unit communicated with resource scheduling system；
It is stored with memory of the trustship process across machine migrator；
Processor is configured to read across the machine migrator of trustship process to execute operations described below: according to user's operation
To resource scheduling system send migration message, to trigger the migration of trustship process, carried in the migration message trustship into
The mark of journey.
The application includes following advantages:
On the one hand, first to file new resources, and trustship is just discharged after trustship process starts on target machine node
The resource that process is occupied in current machine node can be avoided old resource and discharge and new resources applications the case where not getting off,
It solves the problems, such as to cause because of trustship process migration not only to can avoid the increase of user cost, Er Qieke using capacity reducing is forced
It effectively prevent online service to interrupt, to improve the stability of hosts applications, further promotes user experience.
On the other hand, the migration of resource scheduling system automatic trigger is provided and user triggers and migrates both modes, not only may be used
There are abnormal all scenes in reply application, and executes migration for abnormal hosts applications in time, effectively avoids migrating stagnant
Problem afterwards shortens the time of migration and service recovery, to promote user experience.
In another aspect, in across the machine migration failure of trustship process can automatic rollback, be restored to trustship process across machine and migrate it
Preceding state realizes transactional of the trustship process entirely across machine migration operation, effectively avoids because of adverse effect caused by migrating, together
When also can avoid because migration caused by resource scheduling system resource waste.
Another aspect, during across the machine migration of trustship process, by trustship process described on the current machine node
Local ephemeral data be synchronized to the target machine node, can guarantee trustship process local ephemeral data in transition process
Integrality avoids trustship process migration from generating the case where being operating abnormally because data are imperfect later.
Certainly, any product for implementing the application must be not necessarily required to reach all the above advantage simultaneously.
Detailed description of the invention
Fig. 1 is the schematic diagram of the exemplary performing environment of the application；
Fig. 2 is the schematic diagram of the another exemplary performing environment of the application；
Fig. 3 is the flow diagram that the trustship process that embodiment one is applied carries out moving method；
Fig. 4 is the exemplary structure schematic diagram for the device that the trustship process that embodiment two is applied is migrated；
Fig. 5 is the transactional execution flow diagram that 1 trustship process of example is migrated across machine；
Fig. 6 is that the schematic diagram of delay machine occurs for machine where trustship process in example 2；
Fig. 7 is the local nonce of synchronous trustship process after across the machine migration affairs for starting the trustship process in example 2
According to schematic diagram；
Fig. 8 is to dispose in target machine node in example 2 and start the schematic diagram after the trustship process；
Fig. 9 is schematic diagram of the trustship process after successfully starting up on machine where new resources in example 2；
Figure 10 is that the schematic diagram of delay machine occurs for machine where trustship process in example 3；
Figure 11 is the local nonce of synchronous trustship process after across the machine migration affairs for starting the trustship process in example 3
According to schematic diagram；
Figure 12 is to dispose in target machine node in example 3 and start the schematic diagram after the trustship process；
Figure 13 is schematic diagram of the trustship process after successfully starting up on machine where new resources in example 3；
Figure 14 is the schematic diagram of a scenario of 4 trustship process migration of example.
Specific embodiment
It is described in detail below in conjunction with technical solution of the accompanying drawings and embodiments to the application.
It should be noted that each feature in the embodiment of the present application and embodiment can be tied mutually if do not conflicted
It closes, within the scope of protection of this application.In addition, though logical order is shown in flow charts, but in certain situations
Under, it can be with the steps shown or described are performed in an order that is different from the one herein.
In a typical configuration, client or the calculating equipment of server may include one or more processors
(CPU), input/output interface, network interface and memory (memory).
Memory may include the non-volatile memory in computer-readable medium, random access memory (RAM) and/or
The forms such as Nonvolatile memory, such as read-only memory (ROM) or flash memory (flash RAM).Memory is computer-readable medium
Example.Memory may include module 1, module 2 ... ..., module N (N is the integer greater than 2).
Computer-readable medium includes permanent and non-permanent, removable and non-movable storage medium.Storage medium
It can be accomplished by any method or technique information storage.Information can be the mould of computer readable instructions, data structure, program
Block or other data.The example of the storage medium of computer includes, but are not limited to phase change memory (PRAM), static random-access is deposited
Reservoir (SRAM), dynamic random access memory (DRAM), other kinds of random access memory (RAM), read-only memory
(ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory techniques, CD-ROM are read-only
Memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, tape magnetic disk storage or
Other magnetic storage devices or any other non-transmission medium, can be used for storage can be accessed by a computing device information.According to
Herein defines, and computer-readable medium does not include non-temporary computer readable media (transitory media), such as modulates
Data-signal and carrier wave.
Currently, the migration of trustship process is generally realized by the following two kinds mode:
First, user perceives trustship process operation exception, and user's proactive notification resource scheduling system discharges abnormal trustship
The applied resource of process, user apply for new resources and redeploy respective application.
Second, resource scheduling system perceives trustship process crush or machine node delay machine, and resource scheduling system is actively released
The computing resource distributed by trustship process, and the computing resource new for corresponding trustship process application are put, and utilizes new meter
It calculates resource and redeploys respective application.
There are following two defects for the first above-mentioned implementation: 1) user's intervention is under any circumstance required, however
And not all application has the ability of real-time perception application operating status, even if user perceives service exception, will also result in length
The not available problem of Time Service；2) user does not ensure that collection in the case where actively discharging trustship process and having occupied resource
Group currently has new resources that can distribute, and has discharged to will appear old resource, and new resources can not be applied any more for various reasons
Situation about coming is ultimately caused using the serious consequences for being forced capacity reducing.
Although above-mentioned second of implementation realizes automation, intervene without user, but also there are two defects: 1) nothing
Method guarantees the transactional of entire migration operation, i.e., trustship process has been occupied after resource is released, and not can guarantee current cluster has newly
Resource is for distributing and using, it is possible that old resource has discharged, and new resources can not be applied any more for various reasons
Situation about coming is forced capacity reducing to cause to apply；2) cluster can only perceive a small number of abnormal feelings such as machine delay machine, network jitter
Condition, however for the operation exception that application process itself occurs, cluster has no ability to automatic sensing, also just can not be in the case
The execution of triggering migration in time, will also result in and service not available problem for a long time.
In addition, above two scheme cannot be guaranteed the data integrity of migration front and back trustship process.
In summary, the trustship process migration of the relevant technologies is primarily present three following Key technique problems:
First, across machine migration triggering mode it is single: or using user triggering mode or be resource scheduling system
It actively triggers, the two can not be compatible with, so that certain applications can not perceive in time the case where operation exception, lead to trustship process
Migration lag, to cause application service unavailable for a long time.Even if the mode of user's triggering can be coped with abnormal using appearance
All scenes, but only service occur it is obvious abnormal or unavailable and by user's discovery, triggered again by user after can just execute and move
It moves, inherently there are problems that response lag, and if user does not have found or without actively triggering, necessarily will cause
Application service is unavailable for a long time.
Second, the release of old resource and the application of new resources misplace: trustship process in operating status can be all assigned
A resource moves to an other machine node if necessary, then needs including dimensions such as CPU, memory, network, disks
Old resource is discharged, while applying for a new resources again to resource scheduling system, if cannot in entire transition process
Guarantee transactional, just will appear old resource and discharged and new resources the case where can not applying for various reasons, and one
There is such case in denier, is equivalent to user's application and is forced capacity reducing, not only increases the cost that user uses cluster, also will cause line
Upper service disruption；
The local ephemeral data of third, trustship process can not synchronize: trustship process generally can all worked as in the process of running
Some ephemeral datas are written in the local disk of preceding machine node, after trustship process is migrated, if it cannot be guaranteed that these are interim
The synchronous migration of data runs appearance exception again after being then likely to result in trustship process migration.
For above three technical problem, the application proposes the following technical solution.
Firstly, involved term is explained as follows herein:
Distributed resource scheduling system: a physical cluster being made of thousands of machine nodes is responsible for cluster
In the management of physical resource (e.g., CPU, memory, network, disk etc.) of all machines, distribution, recycling and be deployed in cluster
Creation, management and destruction of middle hosts applications etc..
Hosts applications: user is deployed in the application of resource scheduling system.
Trustship process: the hosts applications being deployed in cluster generally comprise one or more functions module, each function mould
Block includes that one or more physics processes are constituted, these physics processes are known as trustship process, and a trustship process operates in one
On machine node, the multiple trustship processes for belonging to a hosts applications be may operate on different machines node.
Across the machine migration of trustship process: it is needed when hosts applications are operating abnormally by its trustship process from currently running machine
Device node moves to be run on other machines node, this process is known as across the machine migration of trustship process.
As shown in Figure 1, being the exemplary performing environment schematic diagram of technical scheme.Wherein, client can be by mutual
Networking or other kinds of mode are interacted with resource scheduling system, to complete the operations such as the resource bid of application, the deployment of application,
Resource scheduling system can be implemented as a physical cluster, include thousands of machine node (machine sections in the physical cluster
Point 1, machine node 2 ..., machine node n-1, machine node n, n is integer not less than 3).In addition, client and resource
The mode communicated between scheduling system is also possible to the other modes in addition to internet.In this regard, herein with no restriction.
As shown in Fig. 2, being the another exemplary performing environment schematic diagram of technical scheme.Wherein, client can lead to
It crosses internet or other kinds of mode is interacted with resource scheduling system, to complete resource bid, the deployment of application etc. of application
Operation, resource scheduling system logically may include resource manager and multiple machine nodes.Wherein, resource manager can be born
The management and distribution of all physical resources in entire physical cluster are blamed, and is responsible for management hosts applications and its trustship process.It is practical
In, the module that entire resource scheduling system is known as resource manager (ResourceManager) by one is responsible for management (should
Module is run in cluster in any one physical machine), which is responsible for all resource management and scheduling, every physical machine
On have and module that only only one is known as monitoring nodes device (NodeMonitor) is responsible for managing single machine physical resource and trustship
The life cycle of process.As an example it is assumed that client needs hosts applications 1 being deployed to resource scheduling system, then such as Fig. 2
Shown, client can be that hosts applications 1 distribute resource, and request scheduling of resource to resource scheduling system under the trigger of the user
The system deployment hosts applications 1.Resource scheduling system distributes resource under the request of client, for hosts applications 1, and by trustship
It is deployed on different machine nodes respectively using 1 multiple trustship processes (trustship process 1, trustship process 2, trustship process 3)
(trustship process 1 is deployed in machine node 1, trustship process 2 is deployed in machine node 2, trustship process 2 is deployed in machine node n),
Finally, multiple trustship processes (trustship process 1, trustship process 2, trustship process 3) of each machine node starting hosts applications 1 to
Realize the normal operation of hosts applications 1.
The technical solution of the application is described in detail below.
Embodiment one
A kind of method that the trustship process of application is migrated, as shown in figure 3, can include:
Step 301, when determination migrates the trustship process of the application of current machine node, it is in purpose machine node
The trustship process distributes new resources；
Step 302, in the target machine node, the new resources based on distribution start the trustship process；
Step 303, the trustship process discharges the trustship process in institute after the target machine node starts successfully
State the resource occupied on current machine node.
The method of the present embodiment, first to file new resources discharge old resource again, and in trustship process in target machine section
The resource that just release trustship process is occupied in current machine node after starting on point, can avoid old resource and has discharged and new resources
The case where application is not got off again solves the problems, such as to cause because of trustship process migration not only to can avoid using using capacity reducing is forced
The increase of family cost, and online service can be effectively prevent to interrupt, to improve the stability of hosts applications, is further promoted and used
Family experience.
The migration of various modes triggering trustship process can be used in method in the present embodiment.In a kind of implementation,
Purpose machine node is that the trustship process is distributed before new resources, can also include: the current machine nodal test to certainly
When body physical resource exception, the migration of the trustship process is triggered；According to triggering trustship when migration message from client into
The migration of journey carries the mark of the trustship process in the migration message.In this way, the present embodiment above method can prop up simultaneously
It holds the following two kinds mode: 1, resource scheduling system automatic trigger and executing the mode of trustship process migration；2, it is held after user's triggering
The mode of the corresponding trustship process migration of row.By the combination of both modes, across the machine migration of automation transactional and touching are realized
On the one hand the compatibility of across the machine two kinds of migration models of migration of hairdo transactional can be coped with using all scenes being operating abnormally, that is, exist
User, which perceives the unavailable rear user of service, can trigger the migration of corresponding trustship process, in the non-support such as machine delay machine or network jitter
Resource scheduling system automatic trigger in the case where itself exception of pipe process, on the other hand in the non-support such as machine delay machine or network jitter
In the case where itself exception of pipe process, resource scheduling system can actively to trustship process carry out fast transferring and restore without
User's intervention so as to triggering rapidly and efficiently and executes the migration of trustship process, effectively shortens migration and service recovery
Time, solve the problems, such as application service for a long time it is not available, promoted user experience.
In a kind of implementation, the process of detection physics resource exception may is that each machine section in resource scheduling system
The state of point difference real-time detection own physical resource, when physical resource exception occurs in itself, into resource scheduling system
Resource manager reports corresponding exception information (for example, trustship process accidentally quit information), so that resource manager is receiving
After the exception information on automatic trigger corresponding machine node trustship process migration operation.Here, the class of physical resource exception
Type can include but is not limited to as follows: delay machine, network jitter, hardware damage, the data exception of run process.In this regard, can also
With using there is other implementations, the application is not limited.
Here, trustship process exception can be divided into two big dimensions: the trustship process exception of user and the object of machine node
Manage resource exception.Wherein, the trustship process exception of user can behave as: trustship process exception exits, trustship proceeding internal memory overflows
(OOM, Out Of Memory).The physical resource of machine node can behave as extremely: machine loading (Load) exception, network flow
Amount is abnormal, disk read-write is abnormal, machine memory abnormal and other machines resource exception are (for example, file handle number exception, machine
Device Thread Count is abnormal).The mode of above-mentioned detection physics resource exception is applicable to the physical resource abnormal conditions of this diversified forms.
Machine node, which is unable to complete, when due to delay machine and obstructed network reports the work, resource manager can monitor trustship
Process exception accidentally quit but do not receive corresponding machine node transmission unexpected message when, be determined as corresponding machine node delay machine or
Network is obstructed.
In a kind of implementation, reception may is that according to the migration of the migration message triggering trustship process from client
To after the migration message of client, need to migrate according to the trustship process identification (PID) confirmation carried in the migration message
Trustship process, then trigger the migration operation of the trustship process.Such as some websites, some trustship of the website is migrated
It at least may include application identities and process identification (PID) in migration message when process, wherein application identities are used for unique identification one
A application, process identification (PID) apply an affiliated trustship process for unique identification.It in addition to this, can be in the migration message
One or more following: migration failure number of retries, migration, which unsuccessfully retry time interval, whether needs to do local ephemeral data moves
It moves.Here, for the trustship process of different application, may be used also other than carrying trustship process identification (PID) in corresponding migration message
To include other content.In this regard, the present embodiment is with no restriction.
In the present embodiment, resource tune may is that in the process that purpose machine node is trustship process distribution new resources
The resource manager of degree system inquires the available resources situation of all machine nodes in addition to current machine node, according to trustship
(the resource specification can be to be reported or resource manager is according to the trustship resource specification needed for process by current machine node
The specification of process occupied resource on current machine node determines, can with the rule of occupied resource on current machine node
Lattice are identical), a machine is selected from all machine nodes in addition to current machine node by scheduled resource scheduling algorithm
Device node, and the resource of dimension is separated to the trustship process from the resource of selected machine node.Here, resource tune
Degree algorithm can be pre-configured with or specify according to actual needs.For example, resource scheduling algorithm may is that selection available resources are not small
The resource specification needed for the trustship process and the smallest machine node of load.In other embodiments, other moneys also can be used
Source dispatching algorithm is distributed new resources for trustship process, in this regard, herein with no restriction.
Some trustship processes could be operated normally dependent on its local ephemeral data, some trustship processes are independent of local
Ephemeral data can be run.For the trustship process dependent on local ephemeral data operation, it is necessary to by its local ephemeral data
It is synchronous, the problem of being unable to operate normally after trustship process migration is caused to avoid because its local ephemeral data is imperfect.In view of
This, in the present embodiment, under disk and the available situation of network, can also during migrating trustship process by the trustship into
The relevant local ephemeral data of journey synchronizes.In a kind of implementation, the support can be started in the new resources based on distribution
Before pipe process, the local ephemeral data of trustship process described on the current machine node is synchronized to the target machine section
Point.In this way, in its local ephemeral data of the preamble of hosts applications starting, it is ensured that trustship process is locally interim in transition process
The integrality of data, avoid due to data are imperfect and influence migration after trustship process normal operation.
The local ephemeral data of trustship process described on the current machine node is synchronized to the target machine node
Specific implementation can there are many.In a kind of implementation, may is that by the trustship on the current machine node into
The local ephemeral data of journey is copied to shared storage file system, then from the shared storage file system by the trustship into
The local ephemeral data of journey is copied to the target machine node.In this way, can be during disposing trustship process from shared
When downloading the program bag of the trustship process in storage file system while reading trustship process described in shared storage file system
Local ephemeral data, execution efficiency can be accelerated.Certainly, in other embodiments, can also be held using other implementations
The synchronization of row trustship process local ephemeral data, in this regard, herein with no restriction.
It, can be same by the local ephemeral data of trustship process described on the current machine node in a kind of implementation
After step to the target machine node, the local ephemeral data of trustship process described on the current machine node is deleted.
In this way, the occupied old resource of trustship process can be discharged to greatest extent, unnecessary data redundancy is avoided, saves storage money
Source.
It can also include: the trustship after the new resources based on distribution start the trustship process in the present embodiment
Process is when target machine node starts failure, automatic rollback.In this way, trustship process migration can be restored to by automatic rollback
State before realizes transactional of the trustship process entirely across machine migration operation, thus bad caused by effectively avoiding because of migration
It influences, while also can avoid the waste because of resource scheduling system resource caused by migrating.
In a kind of implementation, the automatic rollback is including but not limited to as follows: 1) keeping the trustship process current
The resource occupied on machine node；2) new resources are released to；3) delete on the target machine node with the trustship into
The relevant data of journey.In addition to this, automatic rollback can also include: by the sheet of trustship process described in shared storage file system
Ground ephemeral data is deleted and related data of the trustship process on current machine node, file or other retain.When
So, in other embodiments, the operation of automatic rollback can also include other content, in this regard, herein with no restriction.
In a kind of implementation, the present embodiment can also be arranged the operation of trustship process migration and execute for circulation.Specifically,
If migration failure for the first time when trustship process executes migration, automatically begin to execute secondary migration behaviour after automatic rollback
Make, until the success of trustship process migration, or until the number of migration operation has reached the preset upper limit.Here, trustship process is moved
The circulation execution for moving operation can be user setting, be also possible to resource scheduling system default setting.If it is user setting
Words, user can be arranged relevant parameter by client, be sent relevant parameter to resource scheduling system, scheduling of resource by client
The hosts applications or trustship process that system is specified for user configure, need to specify these trustship of hosts applications into
Journey or specified trustship process recycled when migration operation according to the configuration.
In the present embodiment, the resource of each trustship process be can include but is not limited to: CPU, memory, I/O resource, network flow
Amount etc..
It is relatively high and be necessary to ensure that service can be normal in real time to be particularly applicable to requirement of real-time for the method for the present embodiment
A kind of application (for example, electric business website, in line computation etc.) of operation, during its trustship process migration can strict guarantee answer
Unaffected with itself, not only can ensure that apply will not service exception because capacity reducing is forced, but also can guarantee that its service is real
Shi Keyong.Certainly, the method for the present embodiment also can be applied to the not high application of requirement of real-time.
Embodiment two
A kind of device that the trustship process of application is migrated, as shown in figure 4, may include:
Distribution module 41, when being migrated for determining to the trustship process of the application of current machine node, in purpose machine
Device node is that the trustship process distributes new resources；
Starting module 42, for starting the trustship process based on the new resources of distribution in the target machine node；
Release module 43, for after the target machine node starts successfully, discharging the support in the trustship process
The resource that pipe process is occupied on the current machine node.
In a kind of implementation, above-mentioned apparatus can also include: trigger module 44, for examining in the current machine node
When measuring own physical resource exception, the migration of the trustship process is triggered；And for receiving moving from client
The migration of trustship process is triggered when moving message according to the migration message, the mark of the trustship process is carried in the migration message
Know.
In a kind of implementation, above-mentioned apparatus can also include: synchronization module 45, for by the current machine node
The local ephemeral data of the trustship process is synchronized to the target machine node.
In a kind of implementation, above-mentioned apparatus can also include: roll-back module 46, in the trustship process in target
When the starting failure of machine node, automatic rollback.
In above-mentioned apparatus in the present embodiment, distribution module 41 is responsible for the application of resource, can be software, hardware or both
Combination.
In above-mentioned apparatus in the present embodiment, starting module 42 is responsible for deployment of the trustship process on some machine node,
It can be the combination of software, hardware or both.
In above-mentioned apparatus in the present embodiment, release module 43 is responsible for the release of resource, can be software, hardware or both
Combination.
In above-mentioned apparatus in the present embodiment, trigger module 44 is responsible for triggering some or the migration of certain trustship processes is grasped
Make, can be the combination of software, hardware or both.
In above-mentioned apparatus in the present embodiment, synchronization module 45 is responsible for the synchronization of trustship process local ephemeral data, can be with
It is the combination of software, hardware or both.
In above-mentioned apparatus in the present embodiment, roll-back module 46 is responsible for executing rolling back action when migrating failure, can be
The combination of software, hardware or both.
The above-mentioned apparatus of the present embodiment can be set in the resource manager in resource scheduling system, also can be implemented as
The resource manager.
The other details of the present embodiment can refer to embodiment one.
Embodiment three
A kind of resource manager in system of distributed resource scheduling, comprising:
It is stored with memory of the trustship process across machine migrator；
Processor is configured to read across the machine migrator of trustship process to execute operations described below: determine to current machine
It is that the trustship process distributes new resources in purpose machine node when the trustship process of the application of device node is migrated；Institute
Target machine node is stated, the new resources based on distribution start the trustship process；The trustship process is in the target machine section
After point starts successfully, the resource that the trustship process is occupied on the current machine node is discharged.
In a kind of implementation, above-mentioned resource manager can also include: be configured to be communicated with client first
Communication unit；The processor may be additionally configured to read across the machine migrator of trustship process to execute operations described below: according to
The exception information that the current machine node reports triggers the migration of the trustship inspection, works as described in the exception information instruction
Preceding machine node physical resource is abnormal；The migration message from client is received by first communication unit, and according to institute
The migration of migration message triggering trustship process is stated, the mark of the trustship process is carried in the migration message.
In a kind of implementation, the processor may be additionally configured to read across the machine migrator of trustship process to hold
Row operations described below: before the new resources based on distribution start the trustship process, by the support on the current machine node
The local ephemeral data of pipe process is synchronized to the target machine node.
In a kind of implementation, the processor, be also configured as reading across the machine migrator of trustship process with
Execute operations described below: the trustship process automatic rollback when target machine node starts failure.
The other details of the present embodiment can refer to embodiment one.
Example IV
A kind of method that the trustship process of application is migrated is applied to client；The method includes at least: according to
Family operates to resource scheduling system and sends migration message, to trigger the migration of trustship process, in the migration message described in carrying
The mark of trustship process.
In a kind of implementation, client passes through the Software Development Kit that can call the provided interface of resource scheduling system
(SDK, Software Development Kit) realizes the communication between resource scheduling system.In practical application, it will specify
SDK be mounted in the client of user, client can by load SDK come call system of distributed resource scheduling provide
Interface sends message to resource scheduling system by internet.The client of user can be the console of user, such as
For the application of website, client is responsible for the console of website operation management.
Embodiment five
A kind of client, comprising:
It is configured to the second communication unit communicated with resource scheduling system；
It is stored with memory of the trustship process across machine migrator；
Processor is configured to read across the machine migrator of trustship process to execute operations described below: according to user's operation
To resource scheduling system send migration message, to trigger the migration of trustship process, carried in the migration message trustship into
The mark of journey.
In practical application, specified SDK is mounted in the client of user, client can be held by load SDK
Row across the machine migrator of trustship process, calls distributed resource scheduling during executing across the machine migrator of trustship inspection
The interface of system offer sends message to resource scheduling system by internet.
The other details of the present embodiment can refer to example IV.
Embodiment six
A kind of computer readable storage medium is stored with trustship process migration journey on the computer readable storage medium
Sequence realizes what the trustship process applied described in embodiment one was migrated when the trustship process migration program is executed by processor
The step of method.
Embodiment seven
A kind of computer readable storage medium is stored with trustship process migration journey on the computer readable storage medium
Sequence realizes what the trustship process applied described in example IV was migrated when the trustship process migration program is executed by processor
The step of method.
The example implementations of the various embodiments described above are described in detail below.It should be noted that hereafter each reality
Example can be combined with each other.Also, each process, implementation procedure etc. can also be adjusted according to the needs of practical application in Examples below
It is whole.In addition, in practical applications, the various embodiments described above can also have other implementations, in this regard, herein with no restriction.
Example 1
The exemplary execution process of the application is described in detail in this example.
As shown in figure 5, the transactional that trustship process is migrated across machine in this example executes process may include:
Step 501, detect that the service of trustship process is abnormal；
Step 502, verify whether to cause extremely for cluster physical resource? if it is, 504 are gone to step, if not
It is then to continue step 503；
In a kind of implementation, the physics for being determined as loading the machines nodes such as height, delay machine, hardware damage, network jitter is provided
Active layer face is abnormal, is considered as the extremely caused trustship process service exception of cluster physical resource, can jump directly to step at this time
504, the trustship process being directed to automatically on failed machines node by resource scheduling system initiates across the machine migration process of transactional, this money
The mode that system automatic trigger is dispatched in source occurs to the whole process time-consuming of trustship process resumption service to touch with respect to user from failure
The mode of hair is shorter.
In another implementation, for the abnormal scene of trustship process service that cluster can not automatically detect, such as into
Journey hang is lived without response, be can be considered that non-clustered physical resource causes extremely, can be continued step 103, is triggered and is started by user
The migration process of entire trustship process, the mode of the mode and resource scheduling system automatic trigger complement one another, can be for such as
The clusters such as the operation exception that application process itself occurs have no ability to the trustship process perceived service exception and carry out at migration
Reason.
Step 503, user triggers the migration affairs of the trustship process, then gos to step 505；
Step 504, start the migration affairs of the trustship process automatically；
Step 505, start affairs, distribute new resources for trustship process；
Step 506, whether resource bid succeeds, and if it is continues step 508, otherwise continues step 507；
Step 507, transaction rollback, migration failure, gos to step 513；
Step 508, it attempts the local ephemeral data of the trustship process is synchronous；
Step 509, in target machine node deployment and start trustship process；
Does step 510, trustship process start success? if it is, step 511, otherwise continues step 512；
Step 511, the resource that release/recycling trustship process occupies on current machine node, migrates successfully, jumps
To step 513；
Step 512, transaction rollback discharges new resources, migration failure；
Step 513, migration affairs terminate.
This example can automatic trigger trustship process migration, can also user trigger trustship process migration, migration fail
When can also automatic rollback, realize the transactional that trustship process is migrated across machine, can be avoided old resource and discharged and new resources
Apply for the case where not getting off again, solves the problems, such as to cause to apply because of trustship process migration to be forced capacity reducing, and can be efficiently fast
The trustship process to operation exception of speed carries out across machine migration.
Example 2
The exemplary realization process that this example migrates trustship process under resource scheduling system automatic trigger mode across machine into
Row is described in detail.
Following entity or logic main body involved in this example:
Resource manager (Resource Manager): a component in distributed resource scheduling system is responsible for whole
The management and distribution of all physical resources in a cluster, and it is responsible for all hosts applications and its trustship process in management cluster.
Node manager (Node Manager): a component in distributed resource scheduling system is responsible for management institute
Physical resource on machine node, and it is responsible for the life cycle of trustship process of the management operating on the machine node.
Client (Client): it is responsible for interacting with resource scheduling system under the trigger of the user, completes the money of application
Source application, deployment of application etc. operate.In the application, which can be realized to install broker program on a user device,
User equipment realizes above-mentioned function by loading the broker program.
Shared storage file system (i.e. distributed file system (Distributed Storage)): being a kind of permission text
The file system that part is shared on multiple host by network can allow the user on more machines to share file and store empty
Between.
In this example, across the machine transition process of trustship process transactional under resource scheduling system automatic trigger mode includes such as
Under:
1) as shown in fig. 6, physical resource exception, such as machine delay machine occur for machine where trustship process.
2) above-mentioned physical resource is perceived by resource scheduling system extremely, as shown in fig. 7, resource scheduling system actively starts
Across the machine migration affairs of the trustship process: being first trustship process other machines application portion same size in the cluster
New resources the local ephemeral data of the trustship process is synchronized to target machine node if applied successfully.If Shen
Please be unsuccessful, the migration of current trustship contact terminates.
In a kind of implementation, apply for that the process of resource may is that Resource Scheduler is being provided according to the specification of required resource
Suitable machine node is found in source scheduling system (for example, meeting the resource specification demands of the new resources and loading the smallest
Machine node), and corresponding resource is distributed on the machine node to dispose the trustship process.
In a kind of implementation, the mode that the local ephemeral data of trustship process is synchronized to target machine node can be with
It is: the local ephemeral data of the trustship process is copied to shared storage file system, then by it from shared storage file system
Target machine node is copied in system.
In a kind of implementation, data can not be restored by being likely to result in due to machine delay machine etc., but not blocking process
Migration.At this point, can also abandon counting if being unable to complete the local ephemeral data copy of the trustship process in above process
According to synchronization, subsequent migration work is continued to execute.
3) as shown in figure 8, being disposed in target machine node and starting the trustship process.
In a kind of implementation, disposed in target machine node and start the trustship process process may is that from
The program bag of the trustship process is downloaded in shared storage file system to the target machine node, the target machine node
It parses described program packet and runs to start the trustship process.
4) as shown in figure 9, if trustship process is successfully starting up on machine where new resources, shared storage file is deleted
The local ephemeral data (if previous step copy data success) of trustship process described in system, at the same discharge the trustship into
The resource that journey is occupied in current machine node, the entire affairs that migrate are completed, and the corresponding service applied of the trustship process is extensive
It is multiple.If the trustship process starts failure on target machine node, rollback process is executed, that is, discharge and is recycled is described new
Resource, the resource for keeping the trustship process to be occupied in current machine node is constant, removes the migration because of the trustship process
And generated on target machine node file, data or other etc., to return to the state before the trustship process migration.
It, in automatic mode can be when the trustship process migration fails according to default configuration weight in a kind of implementation
It is multiple that above-mentioned migration affairs are executed again, to ensure that the trustship process being capable of Successful migration.
In a kind of implementation, the synchronization of local ephemeral data can also be disposed on target machine node the trustship into
It is carried out during journey.Specifically, shared storage file system first can be copied to the local ephemeral data of the trustship process
In system, simultaneously by the trustship when program bag of the trustship process is downloaded from shared storage file system during deployment
The local ephemeral data of process is copied, and is parsed again later and is run the described program packet starting trustship process.
Example 3
This example triggers mode subiculum pipe process to user and is described in detail across the exemplary realization process that machine migrates.
In this example, user triggers across the machine transition process of mode subiculum pipe process transactional and includes the following:
1) as shown in Figure 10, trustship process is abnormal because occurring service the reason of certain non-physical resource level, such as
Process itself crush (i.e. process accidentally quit) or process hang lives (i.e. process can not be executed normally).
2) as shown in figure 11, across the machine migration stream of the client hair message informing resource scheduling system starting transactional of user
Journey, resource scheduling system start across the machine migration process of transactional of corresponding trustship process in the case where user triggers.Specifically, resource
Scheduling system is trustship process Shen on the other machines node in addition to current machine node where the trustship process
Please the new resources of a same size the local ephemeral data of the trustship process is synchronized to target if applied successfully
Machine node.If application is unsuccessful, the migration of current trustship contact terminates.
In a kind of implementation, the client of user can load for docking with resource scheduling system or for calling money
The broker program (for example, SDK) of the provided interface of system is dispatched in source, calls resource scheduling system institute by loading the broker program
The interface of offer sends message to resource scheduling system, across the machine migration stream of the transactional to start process corresponding to respective application
Journey.
In a kind of implementation, the mode that the local ephemeral data of trustship process is synchronized to target machine node can be with
It is: as shown in figure 11, the local ephemeral data of the trustship process is copied to shared storage file system, then by it from shared
Target machine node is copied in storage file system.
It, can in above process if the local ephemeral data of the trustship process copies failure in a kind of implementation
It is synchronous to abandon corresponding data, continue subsequent migration operation.
3) it as shown in figure 12, is redeployed in target machine node and starts trustship process.
4) as shown in figure 13, if the trustship process deletes shared deposit in success pull-up on machine where new resources
The local ephemeral data (if the success of previous step copy data) of trustship process described in file system is stored up, while described in release
The resource that trustship process is occupied in current machine node, the entire affairs that migrate are completed, the clothes of the corresponding application of the trustship process
Business is successfully recovered.If the trustship process carries out rolling back action, that is, discharges simultaneously in pull-up failure on machine where new resources
Recycle the new resources, the resource that keeps the trustship process to be occupied in current machine node is constant, removes because of the trustship
The migration of process and generated on target machine node file, data or other etc., with return to the trustship process migration it
Preceding state.
Example 4
This example combines exemplary application scene that the transition process of the application trustship process is described in detail.
As shown in figure 14, for this hosts applications of electric business website, it is assumed that there are four trustship processes for hosts applications tool: support
Pipe process 1, trustship process 2, trustship process 3, trustship process 4 are deployed in four machine nodes of resource scheduling system: machine respectively
Device node 1, machine node 2, machine node n-1, machine node n.
In the present embodiment, resource scheduling system and client pass through 10,000,000,000 network communications.Each machine in resource scheduling system
Device node and shared storage file system etc. preferably use solid state hard disk (SSD, Solid State Drives), it can be achieved that fast
Fast efficient trustship process migration.
The physical resource situation of each machine node real-time detection itself, it is assumed that the NodeManager detection in machine node
To trustship process accidentally quit, the information for then reporting trustship process surprisingly to exit extremely is to resource manager, resource manager
After receiving the information that the trustship process that machine node 1 reports surprisingly exits extremely, confirmation needs to run on machine node 1
Trustship process 1 migrated, then start execute trustship process 1 across machine migration, it is as follows: resource manager confirmation trustship process
1 needs the resource of following specification: after 4 CPU, 4 memories, I/O resource 3, apply for the new resources of dimension for trustship process 1,
Resource manager determines that machine node 3 is capable of providing the resource of dimension by resource scheduling algorithm, then from machine node 3
The new resources of dimension are distributed to trustship process 1, and confirm new resources application success.After this, resource manager is by machine
The local ephemeral data of trustship process 1 is synchronized to machine node 3 on node 1, and disposes on machine node 3 and start the support
Pipe process 1, after confirmation trustship process 1 successfully starts up on machine node 3, release trustship process 1 occupies on machine node 1
Resource and by machine node 1 trustship process 1 local ephemeral data delete, in this way, trustship process 1 is just moved from machine node
Machine node 3 is moved on to.
When user perceives the operation exception of trustship process 3, triggering client sends migration message to resource manager, described
The mark of migration message carrying trustship process 3.After resource manager receives the migration message, according to the trustship wherein carried
The mark of process 3 executes across the machine migration of trustship process 3.After across the machine migration for executing trustship process 3, by it from machine section
Point n-1 has moved to machine node 4.Here, across the machine transition process of trustship process 3 is similar with above-mentioned trustship process 1, no longer superfluous
It states.
It should be noted that above-mentioned Figure 14 is merely illustrative, it is not intended to limit this application.In other practical application scenes
Under, the trustship process migration of the application can also be realized by other means.
Those of ordinary skill in the art will appreciate that all or part of the steps in the above method can be instructed by program
Related hardware is completed, and described program can store in computer readable storage medium, such as read-only memory, disk or CD
Deng.Optionally, one or more integrated circuits can be used also to realize in all or part of the steps of above-described embodiment.Accordingly
Ground, each module/unit in above-described embodiment can take the form of hardware realization, can also use the shape of software function module
Formula is realized.The application is not limited to the combination of the hardware and software of any particular form.
Certainly, the application can also have other various embodiments, ripe without departing substantially from the application spirit and its essence
Various corresponding changes and modifications, but these corresponding changes and change ought can be made according to the application by knowing those skilled in the art
Shape all should belong to the protection scope of claims hereof.
Claims (18)
1. the method that a kind of trustship process of application is migrated, comprising:
It is the trustship process in purpose machine node when determination migrates the trustship process of the application of current machine node
Distribute new resources；
In the target machine node, the new resources based on distribution start the trustship process；
The trustship process discharges the trustship process in the current machine section after the target machine node starts successfully
The resource occupied on point.
2. the method according to claim 1, wherein it is described purpose machine node be the trustship process distribution
Before new resources, further includes:
When the current machine nodal test is to own physical resource exception, the migration of the trustship process is triggered；
According to the migration message from client, trigger the migration of trustship process, carried in the migration message trustship into
The mark of journey.
3. according to the method described in claim 2, it is characterized in that, the physical resource includes one or more following extremely:
Delay machine；
Network jitter；
Hardware damage.
4. the method according to claim 1, wherein the new resources based on distribution start the trustship process it
Before, further includes:
The local ephemeral data of trustship process described on the current machine node is synchronized to the target machine node.
5. according to the method described in claim 4, it is characterized in that, by the sheet of trustship process described on the current machine node
Ground ephemeral data is synchronized to the target machine node, comprising:
The local ephemeral data of trustship process described on the current machine node is copied to shared storage file system, then from
The local ephemeral data of the trustship process is copied to the target machine node in the shared storage file system.
6. according to claim 1, method described in 4 or 5, which is characterized in that the release trustship process is in current machine
After the resource occupied on node, further includes:
The local ephemeral data of trustship process described on the current machine node is deleted.
7. according to claim 1, method described in 4 or 5, which is characterized in that the new resources based on distribution start the support
After pipe process, further includes:
The trustship process is when the target machine node starts failure, automatic rollback.
8. the method according to the description of claim 7 is characterized in that the automatic rollback includes one or more following:
The resource for keeping the trustship process to occupy on current machine node；
Discharge the new resources；
Delete data relevant to the trustship process on the target machine node.
9. the device that a kind of trustship process of application is migrated, comprising:
Distribution module, when being migrated for determining to the trustship process of the application of current machine node, in purpose machine node
New resources are distributed for the trustship process；
Starting module, for starting the trustship process based on the new resources of distribution in the target machine node；
Release module, for after the target machine node starts successfully, discharging the trustship process in the trustship process
The resource occupied on the current machine node.
10. device according to claim 9, which is characterized in that further include:
Trigger module, for triggering the trustship process when the current machine nodal test is to own physical resource exception
Migration；And for triggering trustship process according to the migration message when receiving the migration message from client
It migrates, the mark of the trustship process is carried in the migration message.
11. device according to claim 9, which is characterized in that further include:
Synchronization module, for the local ephemeral data of trustship process described on the current machine node to be synchronized to the target
Machine node.
12. device according to claim 9, which is characterized in that further include:
Roll-back module is used for the automatic rollback in the trustship process when target machine node starts failure.
13. the resource manager in a kind of system of distributed resource scheduling, comprising:
It is stored with memory of the trustship process across machine migrator；
Processor is configured to read across the machine migrator of trustship process to execute operations described below:
It is the trustship process in purpose machine node when determination migrates the trustship process of the application of current machine node
Distribute new resources；
In the target machine node, the new resources based on distribution start the trustship process；
The trustship process discharges the trustship process in the current machine section after the target machine node starts successfully
The resource occupied on point.
14. resource manager according to claim 13, which is characterized in that
Further include: it is configured to the first communication unit communicated with client；
The processor is additionally configured to read across the machine migrator of trustship process to execute operations described below:
According to the exception information that the current machine node reports, the migration of the trustship inspection is triggered, the exception information refers to
Show that the current machine node physical resource is abnormal；
The migration message from client is received by first communication unit, and corresponding support is triggered according to the migration message
The migration of pipe process carries the mark of the trustship process in the migration message.
15. resource manager according to claim 13, which is characterized in that the processor is additionally configured to described in reading
Across the machine migrator of trustship process is to execute operations described below:
Before the new resources based on distribution start the trustship process, by trustship process described on the current machine node
Local ephemeral data is synchronized to the target machine node.
16. resource manager according to claim 13, which is characterized in that the processor is additionally configured to described in reading
Across the machine migrator of trustship process is to execute operations described below:
The trustship process automatic rollback when target machine node starts failure.
17. the method that a kind of trustship process of application is migrated, comprising:
Migration message is sent to resource scheduling system according to user's operation, to trigger the migration of trustship process, the migration message
The middle mark for carrying the trustship process.
18. a kind of client, comprising:
It is configured to the second communication unit communicated with resource scheduling system；
It is stored with memory of the trustship process across machine migrator；
Processor is configured to read across the machine migrator of trustship process to execute operations described below: according to user's operation to money
Source scheduling system sends migration message and carries the trustship process in the migration message to trigger the migration of trustship process
Mark.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710908817.2A CN109582459A (en) | 2017-09-29 | 2017-09-29 | The method and device that the trustship process of application is migrated |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710908817.2A CN109582459A (en) | 2017-09-29 | 2017-09-29 | The method and device that the trustship process of application is migrated |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109582459A true CN109582459A (en) | 2019-04-05 |
Family
ID=65914324
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710908817.2A Pending CN109582459A (en) | 2017-09-29 | 2017-09-29 | The method and device that the trustship process of application is migrated |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109582459A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111240833A (en) * | 2019-12-31 | 2020-06-05 | 厦门网宿有限公司 | Resource migration method and device |
| CN111240848A (en) * | 2020-02-14 | 2020-06-05 | 腾讯科技（深圳）有限公司 | Task allocation processing method and system |
| CN111949380A (en) * | 2020-09-18 | 2020-11-17 | 成都精灵云科技有限公司 | System and method for realizing application migration based on sentinel model |
| CN113961319A (en) * | 2021-08-13 | 2022-01-21 | 北京字节跳动网络技术有限公司 | Method and device for job thermal migration, electronic equipment and storage medium |
| CN115480893A (en) * | 2022-09-06 | 2022-12-16 | 江苏安超云软件有限公司 | Application migration method, system and cluster among heterogeneous nodes |
| CN117493024A (en) * | 2023-12-28 | 2024-02-02 | 北京趋动智能科技有限公司 | Multi-process heterogeneous program migration method, storage medium and electronic equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104468759A (en) * | 2014-11-27 | 2015-03-25 | 中国联合网络通信集团有限公司 | Method and device for achieving application migration in PaaS platform |
| US20150331721A1 (en) * | 2013-01-28 | 2015-11-19 | Fujitsu Limited | Process migration method, computer system and computer program |
| JP2016099972A (en) * | 2014-11-26 | 2016-05-30 | 日本電信電話株式会社 | Process migration method and cluster system |
| CN106657173A (en) * | 2015-10-29 | 2017-05-10 | 华为技术有限公司 | Business transfer method and device in software upgrading under NFV configuration, and server |
| CN106663034A (en) * | 2014-05-09 | 2017-05-10 | 亚马逊技术股份有限公司 | Migration of applications between an enterprise-based network and a multi-tenant network |
-
2017
- 2017-09-29 CN CN201710908817.2A patent/CN109582459A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150331721A1 (en) * | 2013-01-28 | 2015-11-19 | Fujitsu Limited | Process migration method, computer system and computer program |
| CN106663034A (en) * | 2014-05-09 | 2017-05-10 | 亚马逊技术股份有限公司 | Migration of applications between an enterprise-based network and a multi-tenant network |
| JP2016099972A (en) * | 2014-11-26 | 2016-05-30 | 日本電信電話株式会社 | Process migration method and cluster system |
| CN104468759A (en) * | 2014-11-27 | 2015-03-25 | 中国联合网络通信集团有限公司 | Method and device for achieving application migration in PaaS platform |
| CN106657173A (en) * | 2015-10-29 | 2017-05-10 | 华为技术有限公司 | Business transfer method and device in software upgrading under NFV configuration, and server |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111240833A (en) * | 2019-12-31 | 2020-06-05 | 厦门网宿有限公司 | Resource migration method and device |
| CN111240833B (en) * | 2019-12-31 | 2023-03-17 | 厦门网宿有限公司 | Resource migration method and device |
| CN111240848A (en) * | 2020-02-14 | 2020-06-05 | 腾讯科技（深圳）有限公司 | Task allocation processing method and system |
| CN111949380A (en) * | 2020-09-18 | 2020-11-17 | 成都精灵云科技有限公司 | System and method for realizing application migration based on sentinel model |
| CN113961319A (en) * | 2021-08-13 | 2022-01-21 | 北京字节跳动网络技术有限公司 | Method and device for job thermal migration, electronic equipment and storage medium |
| CN113961319B (en) * | 2021-08-13 | 2023-11-07 | 抖音视界有限公司 | Method and device for job hot migration, electronic equipment and storage medium |
| CN115480893A (en) * | 2022-09-06 | 2022-12-16 | 江苏安超云软件有限公司 | Application migration method, system and cluster among heterogeneous nodes |
| CN117493024A (en) * | 2023-12-28 | 2024-02-02 | 北京趋动智能科技有限公司 | Multi-process heterogeneous program migration method, storage medium and electronic equipment |
| CN117493024B (en) * | 2023-12-28 | 2024-04-19 | 北京趋动智能科技有限公司 | Multi-process heterogeneous program migration method, storage medium and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109582459A (en) | The method and device that the trustship process of application is migrated | |
| EP3518110B1 (en) | Designation of a standby node | |
| CN111290834B (en) | Method, device and equipment for realizing high service availability based on cloud management platform | |
| CN110377395B (en) | Pod migration method in Kubernetes cluster | |
| CN109683826B (en) | Capacity expansion method and device for distributed storage system | |
| CN111338774B (en) | Distributed timing task scheduling system and computing device | |
| CN111338773B (en) | Distributed timing task scheduling method, scheduling system and server cluster | |
| US20190288914A1 (en) | Allocating VNFC Instances with Anti Affinity Rule to Hosts | |
| CN109656742B (en) | Node exception handling method and device and storage medium | |
| US8843581B2 (en) | Live object pattern for use with a distributed cache | |
| CN101256533A (en) | System and method for performing memory management in a computing environment | |
| US20220283846A1 (en) | Pod deployment method and apparatus | |
| CN106331081B (en) | Information synchronization method and device | |
| CN113064744A (en) | Task processing method and device, computer readable medium and electronic equipment | |
| WO2020063550A1 (en) | Policy decision method, apparatus and system, and storage medium, policy decision unit and cluster | |
| CN105357042A (en) | High-availability cluster system, master node and slave node | |
| CN114185558A (en) | Native application master selection method and device based on K8s and storage medium | |
| CN111078119A (en) | Data reconstruction method, system, device and computer readable storage medium | |
| EP4006725A1 (en) | Virtual machine migration processing and strategy generation method, apparatus and device, and storage medium | |
| CN103500126B (en) | A kind of automatization fault-tolerant configuration method of cloud computing platform | |
| CN116095145B (en) | Data control method and system of VPC cluster | |
| CN108243205A (en) | A kind of method, equipment and system for being used to control cloud platform resource allocation | |
| CN105743696A (en) | Cloud computing platform management method | |
| CN104468674B (en) | Data migration method and device | |
| CN115794306A (en) | Resource allocation method and device based on preemption instance, electronic equipment and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190405 |
|
| RJ01 | Rejection of invention patent application after publication |