WO2020164265A1

WO2020164265A1 - Method and related device for adjusting deployment of docker component on basis of affinity

Info

Publication number: WO2020164265A1
Application number: PCT/CN2019/116704
Authority: WO
Inventors: 赵骏
Original assignee: 平安科技（深圳）有限公司
Priority date: 2019-02-11
Filing date: 2019-11-08
Publication date: 2020-08-20
Also published as: CN109933406A; CN109933406B

Abstract

The present application provides a method, apparatus, computer device, and readable storage medium for adjusting the deployment of a Docker component on the basis of affinity, relating to the technical field of distributed deployment, said method comprising: according to a preset frequency, obtaining run information and run occupancy information between Docker components within a time period corresponding to the preset frequency; analyzing the run information between each component and the run occupancy information of each component to obtain an affinity relationship; generating a frequent item set of components according to each affinity relationship; according to each affinity relationship and the frequent item set of components, adjusting the deployment of each component in each host group. In the present application, the run information and run occupancy information of each Docker component in a system is monitored according to a preset frequency, and the affinity relationship between each component is automatically parsed; according to the priority and the affinity between the components, the deployment of various components on each host group is adjusted such that components having higher affinity may be deployed on the same geographical host group, thus facilitating data communication transmission and execution efficiency.

Description

Method and related equipment for adjusting Docker component deployment based on affinity

This application claims the priority of a Chinese patent application filed with the Chinese Patent Office on February 11, 2019, the application number is 201910110281.9, and the invention title is "Methods and Related Equipment for Adjusting Docker Component Deployment Based on Affinity". The reference is incorporated in this application.

Technical field

This application relates to the field of distributed deployment technology, and in particular to a method and related equipment for adjusting the deployment of Docker components based on affinity.

Background technique

In large financial systems, Docker containers are often used to deploy micro-service distributed systems. A dozen or even dozens of components are often involved in a distributed system. Each component is set in a Docker container, and the distributed system is run based on the Docker container. The operation of each Docker component requires efficient interaction and reasonable resource utilization. Therefore, it is necessary to parse out the affinity relationship between the components in order to manage which Docker components are deployed together and which Docker components need to be isolated from each other during deployment. Among them, the affinity relationship refers to the affinity relationship between components. Affinity is used to show whether the internal connection between components is close, for example, according to whether the mutual calls and resource sharing between components and components are frequent, so Determine the affinity between components. Existing users use k8s (Kubernetes), an internal container management system, to deploy Docker components based on affinity. However, when deploying Docker components through k8s, the affinity of each component requires the user to manually set a very detailed label and configuration, and then k8s can determine the affinity between components based on the label and configuration of each component. Moreover, once the configuration and label are determined, the affinity between the components cannot be adjusted, unless the system is restarted to complete the deployment of the components again.

technical problem

The main purpose of this application is to provide a method and related equipment for adjusting the deployment of Docker components based on affinity, which aims to solve the disadvantages that the affinity relationship between existing Docker components needs to be manually set and the deployment of Docker components is not instantaneous.

Technical solutions

In order to achieve the above objective, this application provides a method for adjusting the deployment of Docker components based on affinity. The Docker components are deployed in a host cluster, which is a cluster formed by all the hosts located in the same region. The method includes:

According to a preset frequency, the operating information between the Docker components and the operating occupation information corresponding to each of the Docker components in the time period corresponding to the preset frequency are acquired, where the operating information is each Docker The calling relationship information and resource sharing information between components, the calling relationship information is the calling relationship and the number of calls corresponding to each other between the components, and the resource sharing information is the corresponding information of the resources commonly called between the components and the components at runtime ；

Parse the operating information between the Docker components and the operating occupation information of each Docker component to obtain the affinity relationship between the Docker components;

Invoke the Apriori algorithm to analyze the affinity relationship between the Docker components, and generate a frequent item set of components, where the frequent item set of the component is a set containing the affinity relationship between the Docker components;

Adjust the deployment of each Docker component on each host group according to the affinity relationship between the Docker components and the frequent item set of the components.

The present application also provides a device for adjusting the deployment of Docker components based on affinity. The Docker components are deployed in a host group, and the host group is a cluster formed by all the hosts located in the same area, and the device includes :

The obtaining module is configured to obtain the operation information among the Docker components within the time period corresponding to the preset frequency and the operation occupation information corresponding to each of the Docker components according to the preset frequency, wherein the operation information Is the calling relationship information and resource sharing information between the Docker components, the calling relationship information is the calling relationship and the number of calls corresponding to each other between the components, and the resource sharing information is the common calling between the components and the components at runtime The corresponding information of the resource;

The parsing module is used to analyze the operation information between the Docker components and the operation occupation information of the Docker components to obtain the affinity relationship between the Docker components;

The first generation module is used to call the Apriori algorithm to analyze the affinity relationship between the Docker components and generate a frequent item set of components, wherein the frequent item set of the component is a set containing the affinity relationship between the Docker components ；

The adjustment module is used to adjust the deployment of each Docker component on each host group according to the affinity relationship between each Docker component and the frequent item set of the component.

The present application also provides a computer device, including a memory and a processor, the memory stores a computer program, and the processor implements the steps of any one of the above methods when the computer program is executed by the processor.

The present application also provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the steps of any one of the above methods are implemented.

Beneficial effect

This application provides a method, device and computer equipment for adjusting the deployment of Docker components based on affinity. By monitoring the operation information and operation occupancy information of each Docker component in the system according to a preset frequency, it automatically analyzes and obtains the information between the components. Affinity. Then according to the preset priority and the affinity relationship between the components, the deployment of each component on the host groups in different regions is dynamically adjusted, so that the components with higher affinity can be deployed on the host groups in the same region , Conducive to data communication transmission and execution efficiency.

Description of the drawings

FIG. 1 is a schematic diagram of the method steps for affinity adjustment of Docker component deployment in an embodiment of the present application;

Fig. 2 is a block diagram of the overall structure of an apparatus for affinity adjustment of Docker component deployment in an embodiment of the present application;

FIG. 3 is a schematic block diagram of the structure of a computer device according to an embodiment of the present application.

The realization, functional characteristics, and advantages of the purpose of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings.

The best mode of the invention

In order to make the purpose, technical solutions, and advantages of this application clearer, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the application, and not used to limit the application.

1, an embodiment of the present application provides a method for adjusting the deployment of Docker components based on affinity. The Docker components are deployed in a host group formed by all the hosts located in the same area. , The method includes:

S1: According to the preset frequency, obtain the operation information among the Docker components within the time period corresponding to the preset frequency and the operation occupation information corresponding to each of the Docker components, where the operation information is each The calling relationship information and resource sharing information between the Docker components, the calling relationship information is the calling relationship and the number of calls between the components and the components, and the resource sharing information is the resource sharing between the components and the components at runtime. Corresponding information

S2: Parse the operation information between the Docker components and the operation occupation information of each Docker component to obtain the affinity relationship between the Docker components;

S3: Invoke the Apriori algorithm to analyze the affinity relationship between the Docker components to generate a frequent item set of components, where the frequent item set of the component is a set containing the affinity relationship between the Docker components;

S4: Adjust the deployment of each Docker component on each host group according to the affinity relationship between each Docker component and the frequent item set of the component.

In this embodiment, the monitoring service in each host group monitors in real time the operation information between the docker components in the docker container and the operation occupation information corresponding to each docker component. Among them, the operation information includes the calling relationship information and resource sharing information between the docker components; the calling relationship information includes the calling relationship and the number of calls between the docker components. Because the docker container component, that is, the services in the docker component are calling each other When you need to send the call request to the API Gateway component first, the API Gateway component is responsible for forwarding the call request to the service of the destination component, so the adjustment system can monitor the forwarding of the call request in the API Gateway component to obtain the corresponding call request Initiating component and destination component of each component, so as to obtain the corresponding calling relationship between each component, and by counting the number of forwarding of the calling request with the calling relationship, the number of calls corresponding to the two components with the calling relationship is obtained; resource sharing Information is the information of the resources that are commonly called between components and components at runtime. The adjustment system can obtain service access resources in the docker container component through network monitoring, and obtain two components with resource sharing based on the requesting end of the accessed resources. That is, the corresponding information of the resources commonly called between components and components at runtime; the operating occupation information includes the CPU occupancy rate and memory occupancy of the component service at runtime, and the adjustment system can be monitored in real time through the monitoring service in the docker container. The adjustment system adjusts the deployment of each docker component on each host group according to the preset frequency. Therefore, according to the preset frequency, the adjustment system intercepts all the information of the real-time monitoring of each host group from the operating information of each docker component in the time period corresponding to the preset frequency and the operating occupation information corresponding to each docker component for the next analysis . Among them, the operation occupancy information corresponding to each docker component is the average value in the time period corresponding to the preset frequency. The adjustment system first calculates the calling frequency of each docker component according to the number of calls between each docker component and the time period corresponding to the preset frequency. Then, input the calling frequency between each docker component, the resource sharing information between each docker component, and the operation occupation information of each docker component into the pre-built affinity library. Among them, the affinity relationship library stores the calling frequency between components, the resource sharing information between components, the operation occupation information of the components and the affinity relationship mapping relationship table. The calling system obtains the affinity between each docker component according to the mapping relationship table. relationship. Among them, the affinity relationship includes affinity information and two components corresponding to the affinity information, and the affinity information includes affinity, closer affinity, violation of affinity, and anti-affinity. During the matching process, some components do not have any affinity with other components, which is specifically expressed as "no affinity". According to the affinity relationship between the docker components obtained after the matching, the adjustment system filters each component to obtain each first component that has a definite affinity relationship with other components, that is, removes the "non-affinity" from all components Part of the components, get each first component. The adjustment system generates an affinity set corresponding to the affinity relationship of each first component according to each first component obtained through screening and the affinity relationship between each component. Among them, the affinity set is a set composed of two first components with a certain affinity relationship and corresponding affinity information, for example, [component 1, component 2, affinity]. The adjustment system analyzes all affinity sets through the Apriori algorithm, and obtains statistics on the number of appearances of each first component and the number of appearances corresponding to each affinity information of the first component, and according to the appearance times of each first component and each first component And the number of occurrences corresponding to each affinity information of the first component to generate a component frequent item set. For example, the affinity set of each first component is: {[component 1, component 2, affinity], [component 2, component 3, anti-affinity], [component 2, component 4, more affinity]}, The component frequent itemset generated after analysis is: {[component 1, occurrences: 1, affinity: 1, relatively affinity: 0, violation of affinity: 0, anti-affinity: 0], [Component 2, number of occurrences: 3 times, affinity: 1 time, more affinity: 1 time, violation of affinity: 0 times, anti-affinity: 1 time], [component 3, number of occurrences: 1 time, affinity : 0 times, more affinity: 0 times, violation of affinity: 0 times, anti-affinity: 1 time], [component 4, number of occurrences: 1 time, affinity: 0 times, more affinity: 1 time, violation Affinity: 0 times, anti-affinity: 0 times]}. The adjustment system first selects the first component with the first priority from the component frequent item set according to the preset priority, and deploys the first component with the first priority in the first host group. Among them, the preset priority is: number of occurrences> affinity> relatively affinity> violation of affinity> anti affinity, and each item is arranged in descending order of the corresponding value; the first host group is each host group Any one of Then, the adjustment system screens out the first component with the second priority from the component frequent item set according to the priority, and judges whether the first component with the second priority and the first component with the first priority are connected according to the affinity relationship. It is affinity or more affinity, that is, whether the preset conditions are met to complete the judgment action of the preset conditions. If it does, the first component with the second priority is deployed in the first host group; if it does not, the first component with the second priority is deployed in the second host group, and the component deployment action has been completed. Wherein, the second host group is any host group excluding the first host group in each host group, and the same host group is a cluster formed by all host groups set up in the same area. The adjustment system sequentially selects the first component from the component frequent item set in descending order of priority, and executes the judgment action of the preset conditions with the first component deployed in the host group one by one, until all the first components in the component frequent item set are deployed carry out.

Further, the operation occupation information includes CPU occupancy rate and memory occupancy, and the operation information among the Docker components and the operation occupation information of the Docker components are analyzed to obtain the affinity between the Docker components. The steps of the relationship include:

S201: Calculate the calling frequency between the Docker components according to the number of calls between the Docker components and the time period corresponding to the preset frequency;

S202: Input the calling frequency among the Docker components, the resource sharing information among the Docker components, and the operation occupation information of each Docker component into a pre-built affinity library, and match each Docker component to obtain The affinity relationship between the components, wherein the affinity relationship library stores the calling frequency between the components, the resource sharing information between the components, the operation occupation information of the components and the affinity relationship mapping relationship table.

In this embodiment, the adjustment system inputs the number of calls between the Docker components and the time period corresponding to the preset frequency into the preset calculation formula, and calculates the call frequency between the Docker components. The calculation formula is F=N/S, F is the calling frequency, N is the number of calls, and S is the time period corresponding to the preset frequency. For example, if the preset frequency is 30 minutes/time, then S is 30 minutes. After obtaining the calling frequency between each Docker component, the adjustment system inputs the calling frequency between each Docker component, the resource sharing information between each Docker component, and the operation occupancy information of each Docker component into the pre-built affinity library. Match the calling frequency between the components, the resource sharing information between the components, the operation occupancy information of the components and the affinity relationship mapping relationship table stored in the relationship library, and obtain the affinity relationship between the components. Among them, the mapping table is:

Further, the affinity relationship between the Docker components includes the affinity information corresponding to two Docker components, and the step of invoking the Apriori algorithm to analyze the affinity relationship between the Docker components and generating a frequent item set of components includes:

S301: According to the affinity relationship between the Docker components, filter each Docker component to obtain each first component that has a definite affinity relationship with other Docker components;

S302: According to each of the first components and the affinity relationship between each of the Docker components, generate an affinity set corresponding to the affinity relationship between each of the first components;

S303: According to each of the affinity sets, statistically obtain the number of appearances of each of the first components and the number of appearances of each affinity information corresponding to each of the first components;

S304: Generate the component frequent item set according to each of the first components, the number of appearances of each of the first components, and the number of appearances corresponding to each affinity information of each of the first components.

In this embodiment, the affinity relationship between the Docker components obtained by the adjustment system according to the matching includes the affinity information corresponding to the two Docker components. For example, the affinity relationship is: component 1, component 2, affinity, where " "Affinity" is the information of affinity. The adjustment system filters each Docker component to obtain each first component that has a definite affinity with other Docker components, that is, removes the "non-affinity" part of the Docker components from all Docker components to obtain each first component. The adjustment system generates an affinity set corresponding to the affinity relationship of each first component according to each first component obtained through screening and the affinity relationship between each Docker component. Among them, the affinity set is a set composed of two first components with a definite affinity relationship and corresponding affinity information. For example, the affinity set is: [component 1, component 2, affinity]. The adjustment system analyzes all affinity sets through the Apriori algorithm, and obtains statistics on the number of appearances of each first component and the number of appearances corresponding to each affinity information of the first component, and according to the appearance times of each first component and each first component And the number of occurrences corresponding to each affinity information of the first component to generate a component frequent item set. For example, the affinity set of each first component is: {[component 1, component 2, affinity], [component 2, component 3, anti-affinity], [component 2, component 4, more affinity]}, The component frequent itemset generated after analysis is: {[component 1, occurrences: 1, affinity: 1, relatively affinity: 0, violation of affinity: 0, anti-affinity: 0], [Component 2, number of occurrences: 3 times, affinity: 1 time, more affinity: 1 time, violation of affinity: 0 times, anti-affinity: 1 time], [component 3, number of occurrences: 1 time, affinity : 0 times, more affinity: 0 times, violation of affinity: 0 times, anti-affinity: 1 time], [component 4, number of occurrences: 1 time, affinity: 0 times, more affinity: 1 time, violation Affinity: 0 times, anti-affinity: 0 times]}.

Further, the affinity information includes affinity, relative affinity, violation of affinity, and anti-affinity. According to the affinity relationship between the Docker components and the frequent item set of the components, each Docker is adjusted. The steps to deploy components on each host cluster include:

S401: According to a preset priority, filter out the first component with the first priority from the frequent item set of components, and deploy the first component with the first priority in a first host group, and A host group is any host group in each of the host groups;

S402: According to the preset priority, filter out the first component with the second priority from the component frequent item set, and determine the first component with the second priority and the first component with the first priority Whether the affinity relationship between the components meets the preset condition to complete the judgment action of the preset condition, wherein the preset condition is that the affinity relationship between the components is affinity or more affinity;

S403: If it meets, deploy the first component with the second priority in the first host group; if it does not meet, deploy the first component with the second priority in the second host group to complete the deployment Component action, wherein the second host group is any host group in each of the host groups except the first host group;

S404: Filter the first component from the component frequent item set in descending order of priority, and execute the judgment action of the preset condition and the deployment component action one by one with each first component currently deployed in the host group, until All the first components in the component frequent item set are deployed.

In this embodiment, the adjustment system first selects the first component with the first priority from the component frequent item set according to the preset priority, and deploys the first component with the first priority in the first host group. Among them, the preset priority is: number of occurrences> affinity> relatively affinity> violation of affinity> anti affinity, and each item is arranged in descending order of the corresponding value; the first host group is each host group Any one of Then, the adjustment system screens out the first component with the second priority from the component frequent item set according to the priority, and judges whether the first component with the second priority and the first component with the first priority are connected according to the affinity relationship. It is affinity or more affinity, that is, whether the preset conditions are met to complete the judgment action of the preset conditions. If it matches, deploy the first component with the second priority in the first host group; if it does not match, deploy the first component with the second priority in the second host group to complete the component deployment action. The second host group is any host group excluding the first host group in each host group. The adjustment system sequentially filters the first component from the component frequent item set in descending order of priority, and executes the judgment action of the preset conditions and the deployment component action with each first component currently deployed in the host group, until the component frequent item set All the first components are deployed. For example, the component frequent item set is: {[Component 1, Occurrence: 1, Affinity: 1, Relative Affinity: 0, Violation Affinity: 0, Anti-Affinity: 0], [Component 2 , Occurrences: 3 times, Affinity: 1 time, Relative affinity: 1 time, Affinity violation: 0 times, Anti-affinity: 1 time], [Component 3, Occurrences: 1 time, Affinity: 0 times , More affinity: 0 times, violation of affinity: 0 times, anti-affinity: 1 time], [component 4, occurrences: 1 time, affinity: 0 times, more affinity: 1 time, violation of affinity: 0 times, anti-affinity: 0 times]}. Component 2 appears 3 times and has 1 affinity. The priority is first. Component 2 is first deployed to host group 1. The number of occurrences of component 1 is the second, the number of affinity is more, and the priority is second, so component 1 is deployed next. Because component 1 and component 2 are compatible, component 1 is also deployed to host group 1. Although the number of occurrences of component 4 and component 3 is 1, and the number of affinity is 0, component 4 is more affinity, and component 3 is anti-affinity, so component 4 is deployed first. Component 4 is not related to component 1 that has been deployed on host cluster 1, and is more compatible with component 2, so it can be deployed on host cluster 1. And if there is an association that violates affinity and anti-affinity, they cannot be deployed together, and they must be deployed independently on a new host group. Finally, component 3 is deployed. Component 3 and component 2 deployed in host cluster 1 are anti-affinity and cannot be deployed to host cluster 1, but must be deployed to new host cluster 2.

Further, the step of deploying the first component with the second priority in the first host group includes:

S4011: Determine whether the affinity relationship between the first component with the second priority and the first component with the first priority is an affinity;

S4012: If it is affinity, deploy the first component with the second priority and the first component with the first priority on the same host; if it is more affinity, deploy the first component with the second priority The component is deployed on any host in the first host group.

In this embodiment, the host group includes multiple hosts. When the adjustment system determines that the first component with the second priority needs to be deployed in the same host group as the first component with the first priority, it also needs to determine whether the relationship between the two first components is affinity, that is, the highest affinity. And degree. If so, deploy the first component with the second priority and the first component with the first priority on the same host, so that the data communication transmission and execution efficiency between the two are higher. If the relationship between the two first components is not relatively close, the first component with the second priority is randomly deployed on any host in the first host group.

Further, according to the preset priority, the first component with the second priority is filtered out from the frequent item set of the components, and the first component with the second priority and the first component with the first priority are determined The step of determining whether the affinity relationship between the first components meets the preset condition to complete the preset condition includes:

S40201: Determine whether the number of host groups currently deployed with the first component is not less than 2;

S40202: If it is not less than 2, determine one by one whether the affinity relationship between the first component currently deployed and each Docker component in the host group on which the first component is currently deployed meets the preset Condition to complete the judgment action of the preset condition;

If it matches, deploy the first component with the second priority in the first host group; if it does not match, deploy the first component with the second priority in the second host group to complete the deployment The steps of component actions include:

S40301: Randomly deploy the first component that is currently deployed to any one of the host groups where the first component is currently deployed.

In this embodiment, when the number of host clusters currently deployed with the first component is not less than 2, if the current deployment of the first component and each component in the host cluster currently deployed with the first component If the affinity relationship of both meets the preset condition, the first component currently deployed is randomly deployed in any one of the host groups where the first component is currently deployed. For example, the host group currently deployed with the first component is: host group A, the first deployed component is a; host group B, the first deployed component is b; host group C, the first deployed component Is c. The first component currently deployed is d, and the affinity with the first components a, b, and c are all affinity or more affinity, then adjust the system to randomly deploy the first component d in host group A and host group Any one of B and host group C.

Further, before the step of obtaining the operation information between the Docker components and the operation occupancy information corresponding to each Docker component in the time period corresponding to the preset frequency according to the preset frequency, the step includes:

S5: Real-time monitoring of the call data of the call request in the API Gateway, where the call request is a request initiated when the Docker components call each other, and the call data includes the request-side component, the destination-side component, and the number of call requests ；

S6: According to the request-end component and the destination-end component, generate the corresponding invocation relationship between the component and the component, and update the number of invocations corresponding to the request-end component and the destination-end component according to the number of invocation requests ；

S7: Generate the calling relationship information according to the calling relationship corresponding to each other between the components and the updated calling times.

In this embodiment, when the services of the components in the docker container call each other, the call request needs to be sent to the API Gateway component first, and then the API Gateway component is responsible for forwarding the call request to the service of the destination component. The adjustment system monitors and collects interactive call data between Docker container services from the API Gateway component through the monitoring service agent installed in each host. The call data includes the request-side component, the destination-side component, and the number of invocation requests. The adjustment system generates a corresponding calling relationship between the component and the component based on the request-side component and the destination-side component, and the two components in each group with the calling relationship record the corresponding number of calls. Adjust the system according to the number of call requests, add the number of call requests this time to the original call times of the two components that have a call relationship in the group, so as to complete the corresponding call to the two components of the group. Update of the number of calls. The adjustment system obtains the calling relationship information of the two components with the calling relationship in the group according to the corresponding calling relationship between the components and the component and the updated calling times.

This embodiment provides a method for adjusting the deployment of Docker components based on affinity. By monitoring the operating information and operating occupancy information of each Docker component in the system according to a preset frequency, the affinity relationship between the components is automatically analyzed. Then according to the preset priority and the affinity relationship between the components, the deployment of each component on the host groups in different regions is dynamically adjusted, so that the components with higher affinity can be deployed on the host groups in the same region , Conducive to data communication transmission and execution efficiency.

2, an embodiment of the present application also provides an apparatus for adjusting the deployment of Docker components based on affinity. The Docker components are deployed in a host group, and the host group is all the hosts set in the same area. Formed cluster, the device includes:

The obtaining module 1 is configured to obtain, according to a preset frequency, the operation information among the Docker components within a time period corresponding to the preset frequency and the operation occupation information corresponding to each of the Docker components, wherein the operation The information is the calling relationship information and resource sharing information between each of the Docker components, the calling relationship information is the calling relationship and the number of calls corresponding to each other between the components, and the resource sharing information is that the components and the components are shared at runtime The corresponding information of the called resource;

The parsing module 2 is used to analyze the operation information between the Docker components and the operation occupation information of the Docker components to obtain the affinity relationship between the Docker components;

The first generation module 3 is used to call the Apriori algorithm to analyze the affinity relationship between each of the Docker components, and generate a frequent item set of components, wherein the frequent item set of the component includes the affinity relationship between the Docker components set;

The adjustment module 4 is configured to adjust the deployment of each Docker component on each host group according to the affinity relationship between each Docker component and the frequent item set of the component.

In this embodiment, the embodiments of the modules and units of the device are consistent with the corresponding method steps described above, and will not be described in detail here.

This embodiment provides an apparatus for adjusting the deployment of Docker components based on affinity. By monitoring the operating information and operating occupancy information of each Docker component in the system according to a preset frequency, the affinity relationship between the components is automatically analyzed. Then according to the preset priority and the affinity relationship between the components, the deployment of each component on the host groups in different regions is dynamically adjusted, so that the components with higher affinity can be deployed on the host groups in the same region , Conducive to data communication transmission and execution efficiency.

3, an embodiment of the present application also provides a computer device. The computer device may be a server, and its internal structure may be as shown in FIG. 3. The computer equipment includes a processor, a memory, a network interface and a database connected through a system bus. Among them, the computer designed processor is used to provide calculation and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer readable instructions, and a database. The internal memory provides an environment for the operation of the operating system and computer-readable instructions in the non-volatile storage medium. The database of the computer equipment is used to store data such as the affinity relationship database. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer-readable instructions are executed by the processor, the processes of the foregoing method embodiments are executed. Those skilled in the art can understand that the structure shown in FIG. 3 is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied.

An embodiment of the present application further provides a computer-readable storage medium. The readable storage medium may be a volatile readable storage medium or a non-volatile readable storage medium on which computer-readable instructions are stored. When the computer-readable instructions are executed by the processor, the processes of the above-mentioned method embodiments are executed. The above are only the preferred embodiments of this application, and do not limit the scope of this application. Any equivalent structure or equivalent process transformation made using the content of this application description and drawings, or directly or indirectly applied to other related The technical field is equally included in the scope of patent protection of this application.

Claims

A method for adjusting the deployment of Docker components based on affinity, where the Docker components are deployed in a host group, the host group being a cluster formed by all the hosts set up in the same area, characterized in that the method includes:

According to a preset frequency, the operating information between the Docker components and the operating occupation information corresponding to each of the Docker components in the time period corresponding to the preset frequency are acquired, where the operating information is each Docker The calling relationship information and resource sharing information between components, the calling relationship information is the calling relationship and the number of calls corresponding to each other between the components, and the resource sharing information is the corresponding information of the resources commonly called between the components and the components at runtime ；

Parse the operating information between the Docker components and the operating occupation information of each Docker component to obtain the affinity relationship between the Docker components;

Invoke the Apriori algorithm to analyze the affinity relationship between the Docker components, and generate a frequent item set of components, where the frequent item set of the component is a set containing the affinity relationship between the Docker components;

Adjust the deployment of each Docker component on each host group according to the affinity relationship between the Docker components and the frequent item set of the components.
The method for adjusting the deployment of Docker components based on affinity according to claim 1, wherein the operating occupancy information includes a CPU occupancy rate and a memory occupancy, and the analysis of the operating information and each of the Docker components The steps of obtaining the affinity relationship between the Docker components by the operation occupation information of the Docker components include:

According to the number of calls between the Docker components and the time period corresponding to the preset frequency, the call frequency between the Docker components is calculated;

Input the calling frequency between the Docker components, the resource sharing information between the Docker components, and the operation occupation information of the Docker components into the pre-built affinity library, and match to obtain the information between the Docker components. Affinity relationship, wherein the affinity relationship library stores a calling frequency between components, resource sharing information between components, operation occupation information of components, and an affinity relationship mapping relationship table.
The method for adjusting the deployment of Docker components based on affinity according to claim 1, wherein the affinity relationship between the Docker components includes the affinity information corresponding to the two Docker components, and the Apriori algorithm is called to analyze the Describe the affinity between Docker components and the steps to generate frequent itemsets of components, including:

According to the affinity relationship between each of the Docker components, filter each of the Docker components to obtain each first component that has a certain affinity with other Docker components;

Generating an affinity set corresponding to the affinity relationship between each of the first components according to each of the first components and the affinity relationship between each of the Docker components;

According to each of the affinity sets, the number of appearances of each of the first components and the number of appearances of each affinity information corresponding to each of the first components are obtained by statistics;

The component frequent item set is generated according to each of the first components, the number of appearances of each of the first components, and the number of appearances corresponding to each affinity information of each of the first components.
The method for adjusting the deployment of Docker components based on affinity according to claim 3, wherein the affinity information includes affinity, relatively affinity, violation of affinity, and anti-affinity, and the Docker The affinity relationship between the components and the frequent item set of the components, and the steps to adjust the deployment of each Docker component on each host group include:

According to the preset priority, the first component with the first priority is selected from the frequent item set of the components, and the first component with the first priority is deployed in the first host group. The first host The group is any host group in each of the host groups;

According to the preset priority, filter out the first component with the second priority from the frequent item set of components, and determine whether the first component with the second priority and the first component with the first priority Whether the affinity relationship between the components meets the preset condition to complete the judgment action of the preset condition, wherein the preset condition is that the affinity relationship between the components is affinity or more affinity;

If it does, the first component with the second priority is deployed in the first host group; if it does not, the first component with the second priority is deployed in the second host group to complete the component deployment action , Wherein the second host group is any host group in each of the host groups except the first host group;

The first component is selected from the frequent item set of components in descending order of priority, and the judgment action of the preset condition and the deployment component action are executed one by one with each first component currently deployed in the host group, until the All the first components in the component frequent item set are deployed.
The method for adjusting the deployment of Docker components based on affinity according to claim 4, wherein the step of deploying the first component with the second priority in the first host group comprises:

Determining whether the affinity relationship between the first component with the second priority and the first component with the first priority is an affinity;

If it is affinity, deploy the first component with the second priority and the first component with the first priority on the same host; if it is more affinity, deploy the first component with the second priority Any host in the first host group.
The method for adjusting the deployment of Docker components based on affinity according to claim 4, wherein the first component with the second priority is selected from the frequent item set of components according to the preset priority, And judging whether the affinity relationship between the first component with the second priority and the first component with the first priority meets a preset condition, and the step of completing the judgment action of the preset condition includes:

Determine whether the number of host groups currently deployed with the first component is not less than 2;

If it is not less than 2, then determine one by one whether the affinity relationship between the first component currently deployed and each Docker component in each host group on which the first component is currently deployed meets the preset condition, To complete the judgment action of the preset conditions;

If it matches, deploy the first component with the second priority in the first host group; if it does not match, deploy the first component with the second priority in the second host group to complete the deployment The steps of component actions include

The first component that is currently deployed is randomly deployed on any one of the host groups where the first component is currently deployed.
The method for adjusting the deployment of Docker components based on affinity according to claim 1, wherein the operation information between the Docker components in a time period corresponding to the preset frequency is obtained according to a preset frequency And before the steps of operating occupancy information corresponding to each of the Docker components, including:

Real-time monitoring of the call data of the call request in the API Gateway, where the call request is a request initiated when the Docker components call each other, and the call data includes the request-side component, the destination-side component, and the number of call requests;

Generating the corresponding invocation relationship between the components and the components according to the request-end component and the destination-end component, and update the number of invocations corresponding to the request-end component and the destination-end component according to the number of invocation requests;

The invocation relationship information is generated according to the invocation relationship corresponding to each other between the components and the updated invocation times.
A device for adjusting the deployment of Docker components based on affinity, the Docker components being deployed in a host group, the host group being a cluster formed by all the hosts set up in the same area, characterized in that the device comprises:

The obtaining module is configured to obtain the operation information among the Docker components within the time period corresponding to the preset frequency and the operation occupation information corresponding to each of the Docker components according to the preset frequency, wherein the operation information Is the calling relationship information and resource sharing information among the Docker components, the calling relationship information is the calling relationship and the number of calls corresponding to each other between the components, and the resource sharing information is the common calling between the components and the components at runtime The corresponding information of the resource;

The parsing module is used to analyze the operation information between the Docker components and the operation occupation information of the Docker components to obtain the affinity relationship between the Docker components;

The first generation module is used to call the Apriori algorithm to analyze the affinity relationship between the Docker components, and generate a frequent item set of components, wherein the frequent item set of the component is a set containing the affinity relationship between the Docker components ；

The adjustment module is used to adjust the deployment of each Docker component on each host group according to the affinity relationship between each Docker component and the frequent item set of the component.
The device for adjusting the deployment of Docker components based on affinity according to claim 8, wherein the operating occupancy information includes CPU occupancy and memory occupancy, and the analysis module includes:

The calculation unit is configured to calculate the calling frequency between the Docker components according to the number of calls between the Docker components and the time period corresponding to the preset frequency;

The matching unit is used to input the calling frequency between the Docker components, the resource sharing information between the Docker components, and the operation occupation information of the Docker components into the pre-built affinity library, and the matching is used to obtain the The affinity relationship between Docker components is described, wherein the affinity relationship library stores the calling frequency between the components, the resource sharing information between the components, the operation occupation information of the components and the affinity relationship mapping table.
The device for adjusting the deployment of Docker components based on affinity according to claim 8, wherein the affinity relationship between the Docker components includes affinity information corresponding to two Docker components, and the first generation module includes :

The screening unit is configured to screen each of the Docker components to obtain each first component that has a definite affinity with other Docker components according to the affinity relationship between each of the Docker components;

The first generating unit is configured to generate an affinity set corresponding to the affinity relationship between each of the first components according to each of the first components and the affinity relationship between each of the Docker components;

A statistical unit, configured to calculate the number of appearances of each of the first components and the number of appearances corresponding to each affinity information corresponding to each of the first components according to each of the affinity sets;

The second generating unit is configured to generate the component frequent item set according to each of the first components, the number of appearances of each of the first components, and the number of appearances corresponding to each affinity information of each of the first components.
The device for adjusting the deployment of Docker components based on affinity according to claim 10, wherein the affinity information includes affinity, relative affinity, violation of affinity, and anti-affinity, and the adjustment module includes:

The first deployment unit is configured to filter out the first component with the first priority from the component frequent item set according to the preset priority, and deploy the first component with the first priority on the first host Group, the first host group is any host group in each of the host groups;

The judging unit is configured to filter out the first component with the second priority from the set of frequent component items according to the preset priority, and determine the first component with the second priority and the first component with the first priority Whether the affinity relationship between the first components meets the preset condition, so as to complete the judgment action of the preset condition, where the preset condition is that the affinity relationship between the components is affinity or more affinity;

The second deployment unit is configured to deploy the first component with the second priority in the first host group; or, to deploy the first component with the second priority in the second host group to complete the deployment Component action, wherein the second host group is any host group in each of the host groups except the first host group;

The circulation unit is used to sequentially filter the first component from the component frequent item set in descending order of priority, and execute the judgment action of the preset condition and the deployment component one by one with each first component currently deployed in the host group Until all the first components in the component frequent item set are deployed.
The device for adjusting deployment of Docker components based on affinity according to claim 11, wherein the second deployment unit comprises:

A first judging subunit for judging whether the affinity relationship between the first component with the second priority and the first component with the first priority is an affinity;

A first deployment subunit, configured to deploy the first component with the second priority and the first component with the first priority on the same host;

The second deployment subunit is configured to deploy the first component with the second priority on any host in the first host group.
The device for adjusting the deployment of Docker components based on affinity according to claim 11, wherein the determining unit comprises:

The second judgment subunit is used to judge whether the number of host groups currently deployed with the first component is not less than 2;

The third judging subunit is used to judge one by one whether the affinity relationship between the first component currently deployed and each Docker component in each of the host groups on which the first component is currently deployed meets the preset Condition to complete the judgment action of the preset condition;

The second deployment unit includes:

The third deployment subunit is configured to randomly deploy the first component that is currently deployed to any one of the host groups where the first component is currently deployed.
The device for adjusting Docker component deployment based on affinity according to claim 8, wherein the device further comprises:

The monitoring module is used to monitor the call data of the call request in the API Gateway in real time, where the call request is a request initiated when the Docker components call each other, and the call data includes the request-side component, the destination-side component, and the call The number of requests;

The second generation module is configured to generate the corresponding calling relationship between the components and the components according to the request-side component and the destination-side component, and update the request-side component and the destination-side component according to the number of the call requests The number of calls corresponding to the component;

The third generation module is configured to generate the calling relationship information according to the calling relationship corresponding to each other between the components and the updated calling times.
A computer device includes a memory and a processor, and the memory stores computer-readable instructions, wherein the processor implements a method for adjusting the deployment of Docker components based on affinity when executed, and the Docker components are deployed in In a host group, the host group is a cluster formed by all the hosts located in the same area, and the method includes:

According to a preset frequency, the operating information between the Docker components and the operating occupation information corresponding to each of the Docker components in the time period corresponding to the preset frequency are acquired, where the operating information is each Docker The calling relationship information and resource sharing information between components, the calling relationship information is the calling relationship and the number of calls corresponding to each other between the components, and the resource sharing information is the corresponding information of the resources commonly called between the components and the components at runtime ；

Parse the operating information between the Docker components and the operating occupation information of each Docker component to obtain the affinity relationship between the Docker components;

Invoke the Apriori algorithm to analyze the affinity relationship between the Docker components, and generate a frequent item set of components, where the frequent item set of the component is a set containing the affinity relationship between the Docker components;

Adjust the deployment of each Docker component on each host group according to the affinity relationship between the Docker components and the frequent item set of the components.
The computer device according to claim 15, wherein the operating occupation information includes a CPU occupancy rate and a memory occupancy, and the analyzing the operating information among the Docker components and the operating occupation information of the Docker components The processor executes the step of parsing the operation information between the Docker components and the operation occupation information of each Docker component to obtain the affinity relationship between the Docker components, including:

According to the number of calls between the Docker components and the time period corresponding to the preset frequency, the call frequency between the Docker components is calculated;

Input the calling frequency between the Docker components, the resource sharing information between the Docker components, and the operation occupation information of the Docker components into the pre-built affinity library, and match to obtain the information between the Docker components. Affinity relationship, wherein the affinity relationship library stores a calling frequency between components, resource sharing information between components, operation occupation information of components, and an affinity relationship mapping relationship table.
The computer device according to claim 15, wherein the affinity relationship between the Docker components includes affinity information corresponding to two Docker components; and the processor executes the calling Apriori algorithm to analyze each of the Docker components The affinity relationship between the components, the steps to generate frequent itemsets of components, including:

According to the affinity relationship between each of the Docker components, filter each of the Docker components to obtain each first component that has a certain affinity with other Docker components;

Generating an affinity set corresponding to the affinity relationship between each of the first components according to each of the first components and the affinity relationship between each of the Docker components;

According to each of the affinity sets, the number of appearances of each of the first components and the number of appearances of each affinity information corresponding to each of the first components are obtained by statistics;

The component frequent item set is generated according to each of the first components, the number of appearances of each of the first components, and the number of appearances corresponding to each affinity information of each of the first components.
A computer-readable storage medium having computer-readable instructions stored thereon, wherein the computer-readable instructions are executed by a processor to implement a method for adjusting the deployment of Docker components based on affinity, and the Docker components are deployed in In a host group, the host group is a cluster formed by all the hosts located in the same area, and is characterized in that the method includes:

According to a preset frequency, the operating information between the Docker components and the operating occupation information corresponding to each of the Docker components in the time period corresponding to the preset frequency are acquired, where the operating information is each Docker The calling relationship information and resource sharing information between components, the calling relationship information is the calling relationship and the number of calls corresponding to each other between the components, and the resource sharing information is the corresponding information of the resources commonly called between the components and the components at runtime ；

Parse the operating information between the Docker components and the operating occupation information of each Docker component to obtain the affinity relationship between the Docker components;

Invoke the Apriori algorithm to analyze the affinity relationship between the Docker components, and generate a frequent item set of components, where the frequent item set of the component is a set containing the affinity relationship between the Docker components;

Adjust the deployment of each Docker component on each host group according to the affinity relationship between the Docker components and the frequent item set of the components.
The computer-readable storage medium according to claim 18, wherein the operating occupancy information includes a CPU occupancy rate and a memory occupancy; the processor executes the parsing of the operating information among the Docker components and each The steps of obtaining the affinity relationship between the Docker components by the operation occupation information of the Docker components include:

According to the number of calls between the Docker components and the time period corresponding to the preset frequency, the call frequency between the Docker components is calculated;

Input the calling frequency between the Docker components, the resource sharing information between the Docker components, and the operation occupation information of the Docker components into the pre-built affinity library, and match to obtain the information between the Docker components. Affinity relationship, wherein the affinity relationship library stores a calling frequency between components, resource sharing information between components, operation occupation information of components, and an affinity relationship mapping relationship table.
The computer-readable storage medium according to claim 18, wherein the affinity relationship between the Docker components includes the affinity information corresponding to the two Docker components, and the processor executes the calling Apriori algorithm to analyze each Describe the affinity between Docker components and the steps to generate frequent itemsets of components, including:

According to the affinity relationship between each of the Docker components, filter each of the Docker components to obtain each first component that has a certain affinity with other Docker components;

Generating an affinity set corresponding to the affinity relationship between each of the first components according to each of the first components and the affinity relationship between each of the Docker components;

According to each of the affinity sets, the number of appearances of each of the first components and the number of appearances of each affinity information corresponding to each of the first components are obtained by statistics;

The component frequent item set is generated according to each of the first components, the number of appearances of each of the first components, and the number of appearances corresponding to each affinity information of each of the first components.