CN113821328A - Scheduling method and device for container cluster, electronic equipment and storage medium - Google Patents
Scheduling method and device for container cluster, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN113821328A CN113821328A CN202111393111.XA CN202111393111A CN113821328A CN 113821328 A CN113821328 A CN 113821328A CN 202111393111 A CN202111393111 A CN 202111393111A CN 113821328 A CN113821328 A CN 113821328A
- Authority
- CN
- China
- Prior art keywords
- scheduling
- preselection
- container
- working
- nodes
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
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
- G06F9/48—Program initiating; Program switching, e.g. by interrupt
- G06F9/4806—Task transfer initiation or dispatching
- G06F9/4843—Task transfer initiation or dispatching by program, e.g. task dispatcher, supervisor, operating system
-
- 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]
- G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
- G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
- G06F9/505—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals considering the load
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a scheduling method, a scheduling device, electronic equipment and a storage medium of a container cluster, wherein the method comprises the following steps: receiving a request to be scheduled, and starting a working node grouping device; pre-grouping a plurality of working nodes, and numbering and sequencing the working nodes in groups; preselecting according to groups according to the grouping numbers of the nodes, and judging whether the nodes meet the scheduling binding requirements of all containers; binding the container scheduling with a preselected successful working node, judging whether the container scheduling to the working node is normal, if so, recording a scheduling result to a statistical analyzer; sending an instruction to a preselection actuator, filtering the working nodes which fail in preselection, and recording the result to a statistical analyzer; checking the health running condition of the working node, executing a preselection actuator, and recording an execution result to a statistical analyzer; and judging whether the container scheduling of the current round is finished. The invention greatly improves the cloud native container cloud platform computing scheduling efficiency, shortens the scheduling time, and improves the accuracy rate and the success rate of preselection of the preselected nodes.
Description
Technical Field
The invention relates to the technical field of cloud computing, in particular to a scheduling method and device of a container cluster, electronic equipment and a storage medium.
Background
The global cloud computing technology has been developed for 20 years, and has evolved into the cloud native technology era to date through the virtualization era and the traditional cloud computing era. Resources in the "cloud" are available at any time, used on demand, paid for on use, and can be expanded indefinitely, a feature known as hydroelectric use of IT infrastructure.
The cloud native architecture is a set of architecture principles and design modes based on cloud native technology, and aims to maximally strip off non-service technical component parts in cloud application, so that cloud native facilities take over a large number of original non-functional characteristics (such as elasticity, safety, observability, gray level and the like) in the application, services are free from non-functional service interruption troubles, and the cloud native architecture has the characteristics of light weight, agility and high automation. The typical technology of the cloud native architecture is a container technology and a Kubernets arrangement and scheduling technology, and in the digital transformation process of an enterprise, the container technology and the Kubernets arrangement and scheduling technology also become the root of a novel PaaS platform computing infrastructure in the cloud native era.
Typical technology representatives of the cloud native architecture are Docker container technology and Kubernets orchestration and scheduling technology, wherein Docker provides lifecycle management of containers and Docker mirroring building of runtime containers. Its main advantage is to pack the settings and dependent items needed by the software/application program operation into a container, thus realizing the advantages of portability, etc. And Kubernetes is used to associate and orchestrate containers running on multiple hosts. Kubernetes is a brand-new distributed system support platform based on a container technology, provides a series of complete functions such as deployment and operation, resource scheduling, service discovery and dynamic expansion for containerized application on the basis of a Docker technology, and improves the convenience of large-scale container cluster management. And the default scheduler is used for selecting the working nodes for the newly established container and is responsible for resource scheduling of the cluster. The method comprises the steps that a most suitable working node is selected from a working node list for each container in a container list to be scheduled through a scheduling algorithm, the whole scheduling process is divided into a pre-selection stage, a preferred grading stage and a selection stage, and more than 30 types of pre-selection stage strategies and preferred grading stage functions exist. In a daily operation state of the cloud native architecture platform, a container scheduling condition often occurs, such as a Pod newly created through an API, a container Pod created by a control manager to complement a copy, when a container elastic scaling scene is triggered, a container Pod needs to be rescheduled and created when a working node crashes, and the like.
Under the high-frequency operation scheduling mechanism, the default scheduler working mechanism traverses more than 30 kinds of preselection stage strategies and preference scoring stages every time the scheduling operation is triggered, and the scheduler continues to judge whether the subsequent preselection filtering conditions are met even if the current working node is known not to meet a certain preselection filtering condition in the preselection stage. Under the condition of large-scale container number, service number and working node number, the judgment logics waste much calculation time, and the time consumed for scheduling containers at one time is more than about several seconds. With the continuous enlargement of the cluster scale, when the number of running containers, the number of services and the number of working nodes are continuously increased, the problems of performance and stability are gradually highlighted, and due to the defects of low throughput of the scheduler and default scheduling strategies of the scheduler, the service capacity expansion overtime fails. On large-scale clusters, the scheduling of one container takes a long time. Because the Kubernetes scheduler is a queued scheduler model, once the number of containers waiting for a capacity expansion peak is too large, the subsequent container expansion queuing is overtime, which causes container scheduling failure, container expansion failure, service system program response overtime and error reporting, poor user experience, recovery of failed containers after the capacity expansion peak, and high service cost.
Disclosure of Invention
In view of the foregoing problems, the present invention provides a scheduling method and apparatus for a container cluster, an electronic device, and a storage medium.
In order to solve the technical problems, the invention adopts the technical scheme that:
in a first aspect, the present invention discloses a method for scheduling a container cluster, which includes the following steps: s1, receiving a request to be scheduled, and starting a working node grouping device; s2, pre-grouping a plurality of working nodes in the container cluster, and carrying out grouping numbering and numbering sequencing; s3, preselecting according to groups according to the grouping numbers of the working nodes, judging whether the working nodes in the preselected node groups meet the scheduling binding requirements of all containers, if so, preselecting successfully, executing a step S4, and if not, preselecting fails, and executing a step S5; s4, binding the container scheduling with the working nodes which are successfully preselected, judging whether the container scheduling to the working nodes is normal, if so, recording a scheduling result to a statistical analyzer, and if not, executing a step S6; s5, sending an instruction to a preselection actuator, filtering the working nodes which fail to preselection, and recording the result to a statistical analyzer; s6, checking the health running condition of the working node, executing a preselection actuator, and recording the execution result to a statistical analyzer; and S7, judging whether the container scheduling of the current round is finished, if so, carrying out preheating operation before the next round of scheduling, and if not, continuing to preselect other operation nodes.
As a preferred scheme, the pre-grouping a plurality of working nodes in a container cluster includes: and sequencing the plurality of working nodes in the container cluster according to the workload quantity of each working node, and grouping the working nodes into a group according to 30-50 working nodes.
Preferably, in step S3, if the working node in the node group checks that there is one of the container scheduling binding requirements that is not satisfied, the preselection fails, and the remaining container binding requirements are not checked.
Preferably, the container scheduling binding requirement includes: whether the name of the working node is matched with the NodeName expected to be dispatched by the Pod is checked, whether a port in a port list used by the preselected Pod is occupied in the preselected working node or not is judged, whether a requested storage volume is available in the current working node or not is judged, and whether the resource availability on the working node meets the operation requirement of the Pod object or not is checked.
Preferably, before the pre-selection actuator is executed, the method further comprises the steps of checking whether a pre-selection controller and the pre-selection actuator are on-line or not, executing the pre-selection actuator if the pre-selection controller and the pre-selection actuator are on-line, and starting the daemon process and then executing the pre-selection actuator if the pre-selection controller and the pre-selection actuator are not on-line.
Preferably, in step S7, the preheating operation includes: and clearing the queue cache of the previous scheduling of the Kubernetes scheduler, and then performing preselection priority recommendation for the next scheduling according to the optimal execution result obtained by the statistical analyzer in the scheduling preselection.
In a second aspect, the present invention discloses a scheduling apparatus for a container cluster, including: the starting module is used for receiving a request to be scheduled and starting the working node grouping device; the pre-grouping module is used for pre-grouping a plurality of working nodes in the container cluster, and performing grouping numbering and numbering sequencing; the preselection module is used for preselecting according to groups according to the grouping numbers of the working nodes, judging whether the working nodes in the node groups after preselection meet the container scheduling binding requirement, if so, the preselection is successful, the binding module is executed, and if not, the preselection is failed, and the filtering module is executed; the binding module is used for binding the container scheduling to the working node which is successfully preselected, judging whether the container scheduling to the working node is normal or not, if so, recording a scheduling result to the statistical analyzer, and if not, executing the health check module; the filtering module is used for sending an instruction to the preselection executor, filtering the working nodes which fail in preselection and recording the result to the statistical analyzer; the health check module is used for checking the health running condition of the working node, executing a preselection actuator and recording an execution result to the statistical analyzer; and the round judgment module is used for judging whether the current round container scheduling is finished, if so, preheating work is carried out before the next round of scheduling, and if not, other work nodes are continuously preselected.
In a third aspect, the invention discloses an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.
In a fourth aspect, the invention discloses a computer-readable storage medium, storing a computer program, which when executed by a processor, performs the steps of any of the methods described above.
Compared with the prior art, the invention has the beneficial effects that: the scheduling method of the container cluster is completely adapted to the existing cloud native technology architecture, the improvement and optimization of the scheduling strategy method are realized on the basis of not changing the basic capability framework of the existing Kubernets scheduler under the condition of not influencing the normal operation of the existing online service, and the effect improvement method of large-scale container scheduling is provided. In the scenario where the cloud platform is most widely used, such as: in high-frequency use scenes such as newly creating Pod containers through an API (application programming interface), creating the Pod containers for complementing copies by a control manager, triggering elastic telescopic scenes of the containers, rescheduling of the created Pod containers when a working node is down and the like, performing efficient intelligent pre-selection on the Pod containers to be created through a distributed intelligent pre-selection controller program to obtain computing nodes to be scheduled, traversing all target nodes, and screening out candidate nodes meeting requirements, so that unnecessary pre-selection algorithm and pre-selection strategy analysis in the computing nodes failing to be preliminarily determined and pre-selected and filtered can be avoided, the computing scheduling time of a cloud platform of the cloud native containers is greatly prolonged, and the accuracy and the pre-selection success rate of the pre-selection nodes are improved. Therefore, the performance of large-scale container cluster scheduling is improved, the cloud platform stability of the cloud native container is improved, the container scheduling success rate and scheduling real-time performance are guaranteed, the user business system quick response and system operation smoothness experience are greatly improved, and then business continuity, service level indexes, service level targets, service quality and other enterprise business system core elements are guaranteed. Enterprises can also realize the improvement of the utilization rate of various resources such as servers, storage, networks and the like through the large-scale container cluster scheduling efficiency-improving method, and the aim of saving the resource cost of various infrastructures is fulfilled.
Drawings
The disclosure of the present invention is illustrated with reference to the accompanying drawings. It is to be understood that the drawings are designed solely for the purposes of illustration and not as a definition of the limits of the invention. In the drawings, like reference numerals are used to refer to like parts. Wherein:
fig. 1 is a flowchart illustrating a scheduling method of a container cluster according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a scheduling apparatus of a container cluster according to an embodiment of the present invention.
Detailed Description
It is easily understood that according to the technical solution of the present invention, a person skilled in the art can propose various alternative structures and implementation ways without changing the spirit of the present invention. Therefore, the following detailed description and the accompanying drawings are merely illustrative of the technical aspects of the present invention, and should not be construed as all of the present invention or as limitations or limitations on the technical aspects of the present invention.
An embodiment according to the present invention is shown in connection with fig. 1. The invention discloses a scheduling method of a container cluster, which comprises the following steps:
and S1, receiving the request to be scheduled, and starting the work node grouping device.
In the embodiment of the invention, a container Pod queue scheduler is used for realizing scheduling queuing control and triggering and starting the work node grouping device. When the start condition is satisfied, the queue scheduler starts the work node grouper.
And S2, pre-grouping the plurality of working nodes in the container cluster, and performing grouping numbering and numbering sequencing.
Specifically, pre-grouping a plurality of working nodes in a container cluster includes: and sequencing the plurality of working nodes in the container cluster according to the quantity of the working loads on each working node, and grouping the working nodes into a group according to 30-50 working nodes, wherein the working nodes are sequenced from small to large according to the quantity of the working loads on each working node.
And S3, preselecting according to groups according to the grouping numbers of the working nodes, judging whether the working nodes in the preselected node groups meet the scheduling binding requirements of all containers, if so, preselecting successfully, executing a step S4, and if not, preselecting unsuccessfully, and executing a step S5.
Specifically, the pre-selection controller is started according to the packet ID according to the packet number information provided by the working node grouping device. And the preselection controller distributes preselection strategies and preselection algorithms to each working node group in a concurrent process calling mode according to the grouping number information sent by the working node grouping device. And the system is in a stop state during the execution period of no preselected control, when the working node receives the request to be scheduled and bound, the preselected controller is started, and then the preselected actuator completes the action of preselected scheduling execution.
The above-mentioned pre-selection strategy and pre-selection algorithm, namely the container scheduling binding requirement, includes: checking whether the name of the working node is matched with the NodeName expected to be dispatched by the Pod, judging whether a port in a port list used by the preselected Pod is occupied in the preselected working node, judging whether a requested storage volume is available in the current working node, checking whether the resource availability on the working node meets the operation requirement of the Pod object, and the like.
And the preselection controller judges the working nodes with failure preselection and filtration according to the various budget strategies, if one of the working nodes fails to have a preselection strategy, the preselection controller does not perform other preselection algorithm and preselection strategy analysis on the working node, and performs preselection work on the next working node again.
And S4, after the preselection is successful, the working nodes enter a scheduling queue, the container scheduling is bound with the working nodes which are successfully preselected, whether the container scheduling to the working nodes is normal or not is judged, if yes, the scheduling result is recorded to a statistical analyzer, and if not, the step S6 is executed.
And S5, sending an instruction to a preselection actuator, filtering the working nodes which fail to preselection, and recording the result to a statistical analyzer.
And S6, checking the health operation condition of the working node, executing a preselection actuator, and recording the execution result to a statistical analyzer.
Judging whether the current condition of the container cluster meets the condition of starting a preselection actuator or not according to the health condition of the working nodes in the container cluster, if so, executing the preselection actuator, finishing the execution of each preselection actuator, reporting the execution result to a statistical analyzer, and summarizing the execution result of preselection scheduling of each working node by the statistical analyzer.
Checking the health operation condition of the working node, comprising the following steps: checking whether various resource loads of the working nodes, such as CPU utilization rate, memory utilization rate, storage IO, network IO, working load and the like, exceed a threshold value, and whether various network communication services can normally provide services to the outside.
Prior to said executing the preselected actuator, further comprising: and checking whether the preselection controller and the preselection actuator are on line or not and whether the daemon process is on line when the preselection controller and the preselection actuator run, if so, executing the preselection actuator, otherwise, starting the daemon process first and then executing the preselection actuator.
Checking whether the pre-selection controller and the pre-selection actuator, and the runtime daemon are online, comprising:
and judging whether the preselected controller is running according to the process survival state and the process running time of the preselected controller. The preselection controller runs on each working node group in a Daemon process mode (Daemon Set). The daemon acts like a Watch Dog program, and if the process survival state of the preselection controller is detected to be abnormal, the preselection controller process is restarted.
And judging whether the last preselection execution result is reported to the statistical analyzer or not according to the timestamp of the data received by the statistical analyzer.
And judging whether the time condition for starting the preselection controller is met or not according to the completion time of the container scheduling period of the current round.
And S7, judging whether the container scheduling of the current round is finished, if so, carrying out preheating operation before the next round of scheduling, and if not, continuing to preselect other operation nodes.
Before each preselection scheduling is started, the working node grouping device obtains the summary information of the scheduling preselection execution result of the last time through the statistical analyzer, and the working node grouping device prepares for preheating work for the next intelligent preselection scheduling in a large-scale container cluster scene. This preheating operation includes: and clearing the queue cache of the previous scheduling of the Kubernetes scheduler, and then performing preselection priority recommendation for the next scheduling according to the optimal execution result obtained by the statistical analyzer in the scheduling preselection.
Referring to fig. 2, the present invention discloses a scheduling apparatus for a container cluster, comprising:
and the starting module is used for receiving the request to be scheduled and starting the work node grouping device.
And the pre-grouping module is used for pre-grouping a plurality of working nodes in the container cluster, and performing grouping numbering and numbering sequencing.
And the preselection module is used for preselecting according to groups according to the grouping numbers of the working nodes, judging whether the working nodes in the node groups after preselection meet the container scheduling binding requirement, if so, the preselection is successful, the binding module is executed, and if not, the preselection is failed, and the filtering module is executed.
And the binding module is used for binding the container scheduling to the working node which is successfully preselected, judging whether the container scheduling to the working node is normal or not, recording a scheduling result to the statistical analyzer if the container scheduling to the working node is normal, and executing the health check module if the container scheduling to the working node is not normal.
And the filtering module is used for sending an instruction to the preselection executor, filtering the working nodes which fail in preselection and recording the result to the statistical analyzer.
And the health check module is used for checking the health running condition of the working node, executing a preselected actuator and recording an execution result to the statistical analyzer.
And the round judgment module is used for judging whether the current round container scheduling is finished, if so, preheating work is carried out before the next round of scheduling, and if not, other work nodes are continuously preselected.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
The invention also discloses an electronic device, which comprises a memory and a processor, wherein the memory stores a computer program, and the processor realizes the steps of any one of the methods when executing the computer program.
The invention also discloses a computer-readable storage medium, in which a computer program is stored, which computer program, when being executed by a processor, realizes the steps of any of the above-mentioned methods.
In summary, the beneficial effects of the invention include: the scheduling method of the container cluster is completely adapted to the existing cloud native technology architecture, the improvement and optimization of the scheduling strategy method are realized on the basis of not changing the basic capability framework of the existing Kubernets scheduler under the condition of not influencing the normal operation of the existing online service, and the effect improvement method of large-scale container scheduling is provided. In the scenario where the cloud platform is most widely used, such as: in high-frequency use scenes such as newly creating Pod containers through an API (application programming interface), creating the Pod containers for complementing copies by a control manager, triggering elastic telescopic scenes of the containers, rescheduling of the created Pod containers when a working node is down and the like, performing efficient intelligent pre-selection on the Pod containers to be created through a distributed intelligent pre-selection controller program to obtain computing nodes to be scheduled, traversing all target nodes, and screening out candidate nodes meeting requirements, so that unnecessary pre-selection algorithm and pre-selection strategy analysis in the computing nodes failing to be preliminarily determined and pre-selected and filtered can be avoided, the computing scheduling time of a cloud platform of the cloud native containers is greatly prolonged, and the accuracy and the pre-selection success rate of the pre-selection nodes are improved. Therefore, the performance of large-scale container cluster scheduling is improved, the cloud platform stability of the cloud native container is improved, the container scheduling success rate and scheduling real-time performance are guaranteed, the user business system quick response and system operation smoothness experience are greatly improved, and then business continuity, service level indexes, service level targets, service quality and other enterprise business system core elements are guaranteed. Enterprises can also realize the improvement of the utilization rate of various resources such as servers, storage, networks and the like through the large-scale container cluster scheduling efficiency-improving method, and the aim of saving the resource cost of various infrastructures is fulfilled.
It should be understood that the integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention essentially or partially contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The technical scope of the present invention is not limited to the above description, and those skilled in the art can make various changes and modifications to the above-described embodiments without departing from the technical spirit of the present invention, and such changes and modifications should fall within the protective scope of the present invention.
Claims (9)
1. A scheduling method for a container cluster is characterized by comprising the following steps:
s1, receiving a request to be scheduled, and starting a working node grouping device;
s2, pre-grouping a plurality of working nodes in the container cluster, and carrying out grouping numbering and numbering sequencing;
s3, preselecting according to groups according to the grouping numbers of the working nodes, judging whether the working nodes in the preselected node groups meet the scheduling binding requirements of all containers, if so, preselecting successfully, executing a step S4, and if not, preselecting fails, and executing a step S5;
s4, binding the container scheduling with the working nodes which are successfully preselected, judging whether the container scheduling to the working nodes is normal, if so, recording a scheduling result to a statistical analyzer, and if not, executing a step S6;
s5, sending an instruction to a preselection actuator, filtering the working nodes which fail to preselection, and recording the result to a statistical analyzer;
s6, checking the health running condition of the working node, executing a preselection actuator, and recording the execution result to a statistical analyzer;
and S7, judging whether the container scheduling of the current round is finished, if so, carrying out preheating operation before the next round of scheduling, and if not, continuing to preselect other operation nodes.
2. The method for scheduling a container cluster according to claim 1, wherein the pre-grouping a plurality of working nodes in the container cluster comprises:
and sequencing the plurality of working nodes in the container cluster according to the workload quantity of each working node, and grouping the working nodes into a group according to 30-50 working nodes.
3. The method for scheduling of container cluster according to claim 1, wherein in step S3, if the working node in the node group checks that there is one of the container scheduling binding requirements that is not satisfied, the preselection fails and the remaining container binding requirements are not checked.
4. The method of claim 1, wherein the container scheduling binding requirement comprises: whether the name of the working node is matched with the NodeName expected to be dispatched by the Pod is checked, whether a port in a port list used by the preselected Pod is occupied in the preselected working node or not is judged, whether a requested storage volume is available in the current working node or not is judged, and whether the resource availability on the working node meets the operation requirement of the Pod object or not is checked.
5. The method of scheduling a cluster of containers of claim 1, further comprising, prior to said executing a pre-selection executor: and checking whether the preselection controller and the preselection actuator are on line or not and whether the daemon process is on line when the preselection controller and the preselection actuator run, if so, executing the preselection actuator, otherwise, starting the daemon process first and then executing the preselection actuator.
6. The method for scheduling a container cluster according to claim 1, wherein in step S7, the preheating operation includes: and clearing the queue cache of the previous scheduling of the Kubernetes scheduler, and then performing preselection priority recommendation for the next scheduling according to the optimal execution result obtained by the statistical analyzer in the scheduling preselection.
7. A scheduling apparatus for a container cluster, comprising:
the starting module is used for receiving a request to be scheduled and starting the working node grouping device;
the pre-grouping module is used for pre-grouping a plurality of working nodes in the container cluster, and performing grouping numbering and numbering sequencing;
the preselection module is used for preselecting according to groups according to the grouping numbers of the working nodes, judging whether the working nodes in the node groups after preselection meet the container scheduling binding requirement, if so, the preselection is successful, the binding module is executed, and if not, the preselection is failed, and the filtering module is executed;
the binding module is used for binding the container scheduling to the working node which is successfully preselected, judging whether the container scheduling to the working node is normal or not, if so, recording a scheduling result to the statistical analyzer, and if not, executing the health check module;
the filtering module is used for sending an instruction to the preselection executor, filtering the working nodes which fail in preselection and recording the result to the statistical analyzer;
the health check module is used for checking the health running condition of the working node, executing a preselection actuator and recording an execution result to the statistical analyzer;
and the round judgment module is used for judging whether the current round container scheduling is finished, if so, preheating work is carried out before the next round of scheduling, and if not, other work nodes are continuously preselected.
8. An electronic device, characterized in that the electronic device comprises a memory storing a computer program and a processor implementing the steps of the method according to any of claims 1-6 when the processor executes the computer program.
9. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111393111.XA CN113821328A (en) | 2021-11-23 | 2021-11-23 | Scheduling method and device for container cluster, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111393111.XA CN113821328A (en) | 2021-11-23 | 2021-11-23 | Scheduling method and device for container cluster, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113821328A true CN113821328A (en) | 2021-12-21 |
Family
ID=78919711
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111393111.XA Pending CN113821328A (en) | 2021-11-23 | 2021-11-23 | Scheduling method and device for container cluster, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113821328A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9015724B2 (en) * | 2009-09-23 | 2015-04-21 | International Business Machines Corporation | Job dispatching with scheduler record updates containing characteristics combinations of job characteristics |
| CN111694641A (en) * | 2020-06-16 | 2020-09-22 | 中电科华云信息技术有限公司 | Storage management method and system for container application |
| CN111770162A (en) * | 2020-06-24 | 2020-10-13 | 重庆紫光华山智安科技有限公司 | Network bandwidth limiting method, device, main node and storage medium |
| CN112035248A (en) * | 2020-08-20 | 2020-12-04 | 浙江大华技术股份有限公司 | Container group scheduling method, electronic device and storage medium |
| CN112463349A (en) * | 2021-01-28 | 2021-03-09 | 北京睿企信息科技有限公司 | Load balancing method and system for efficiently scheduling GPU (graphics processing Unit) capability |
-
2021
- 2021-11-23 CN CN202111393111.XA patent/CN113821328A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9015724B2 (en) * | 2009-09-23 | 2015-04-21 | International Business Machines Corporation | Job dispatching with scheduler record updates containing characteristics combinations of job characteristics |
| CN111694641A (en) * | 2020-06-16 | 2020-09-22 | 中电科华云信息技术有限公司 | Storage management method and system for container application |
| CN111770162A (en) * | 2020-06-24 | 2020-10-13 | 重庆紫光华山智安科技有限公司 | Network bandwidth limiting method, device, main node and storage medium |
| CN112035248A (en) * | 2020-08-20 | 2020-12-04 | 浙江大华技术股份有限公司 | Container group scheduling method, electronic device and storage medium |
| CN112463349A (en) * | 2021-01-28 | 2021-03-09 | 北京睿企信息科技有限公司 | Load balancing method and system for efficiently scheduling GPU (graphics processing Unit) capability |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109960585B (en) | Resource scheduling method based on kubernets | |
| CN109918198B (en) | Simulation cloud platform load scheduling system and method based on user characteristic prediction | |
| CN112162865B (en) | Scheduling method and device of server and server | |
| CN111399989B (en) | Container cloud-oriented task preemption and scheduling method and system | |
| CN108632365B (en) | Service resource adjusting method, related device and equipment | |
| CN110134495B (en) | Container cross-host online migration method, storage medium and terminal equipment | |
| CN111818159B (en) | Management method, device, equipment and storage medium of data processing node | |
| CN114153580A (en) | Cross-multi-cluster work scheduling method and device | |
| CN111381928B (en) | Virtual machine migration method, cloud computing management platform and storage medium | |
| CN112416585A (en) | GPU resource management and intelligent scheduling method for deep learning | |
| TW202133055A (en) | Method for establishing system resource prediction and resource management model through multi-layer correlations | |
| CN112799817A (en) | Micro-service resource scheduling system and method | |
| CN110727508A (en) | Task scheduling system and scheduling method | |
| CN111190691A (en) | Automatic migration method, system, device and storage medium suitable for virtual machine | |
| CN106101212A (en) | Big data access method under cloud platform | |
| CN113391913A (en) | Distributed scheduling method and device based on prediction | |
| CN111767145A (en) | Container scheduling system, method, device and equipment | |
| CN112799837A (en) | Container dynamic balance scheduling method | |
| CN111796933A (en) | Resource scheduling method, device, storage medium and electronic equipment | |
| CN109614210B (en) | Storm big data energy-saving scheduling method based on energy consumption perception | |
| CN114911613A (en) | Cross-cluster resource high-availability scheduling method and system in inter-cloud computing environment | |
| Hu et al. | Cloud model-based security-aware and fault-tolerant job scheduling for computing grid | |
| CN113778627A (en) | Scheduling method for creating cloud resources | |
| CN113821328A (en) | Scheduling method and device for container cluster, electronic equipment and storage medium | |
| CN109783304B (en) | Energy-saving scheduling method and corresponding device for data center |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20211221 |