TW202340942A - System for deploying high availability service, method and computer readable medium thereof - Google Patents

Abstract

The present invention is a system and a method for deploying high availability service. Based on the content of a service to be deployed and a cluster data of all the clusters in the mesh, the clusters involved in a service deployment are obtained and the deployment sequence is arranged by a control platform, so as commands in the cluster is generated according to parameters for the service deployment and the deployment sequence. The commands in the cluster is obtained and parsed by a cluster agent, so as the deployment instructions that need to be executed at the local end is executed, and the deployment instructions that need to be executed at the remote end is executed by the corresponding cluster agent that is notified through the control platform. After the service deployment is completed by all cluster agents, deployment results are reported to the control platform for confirmation. The present invention also provides a computer-readable medium for executing the method of the present invention.

Description

High-availability service deployment system, method and computer-readable medium

The present invention relates to the technology of network service deployment, and in particular, to a high-availability service deployment system and method and its computer-readable medium.

With the advancement of network technology and service demand, many service deployments and installations can be automatically executed through the system, reducing time and space constraints. For service providers or operators, it not only saves manpower but also shortens the time. It is indeed a major advancement in technology.

With the diversified development of network architecture, the concept of container has obviously become a necessary configuration. Containers include elements that allow applications to execute smoothly in different environments, that is, a container that allows application code, running processes, The practice of integrating system tools, system libraries and related configurations into one entity, and the action of deploying and organizing application containers is called container orchestration. There are currently many container orchestration tools on the market, such as those by Google. The Kubernetes (K8S for short) system developed. For shared software deployment and installation solutions, it is common for a container to involve multiple different clusters, that is, a service will be deployed to multiple clusters. However, in the existing technology, each cluster can only be deployed separately. The lack of correlation between these clusters makes the installation process inconvenient and complicated. This is cumbersome, especially for high availability (High Availability) services, which usually require multiple clusters to execute. If services are only installed in separate clusters, there may be problems such as inefficiency and difficulty in overall inspection.

In view of the above problems, how to find a way to deploy network services that can make high-availability services involving multiple clusters more smoothly deployed and installed. In particular, in addition to reducing the problems derived from individual deployments, At the same time, it also strengthens the communication between various clusters, which will become a goal that people in this technical field are currently eager to pursue.

In order to solve the above-mentioned problems of the prior art, the present invention discloses a high-availability service deployment system, which includes: a control platform device, which is used to analyze the content of the service to be deployed and obtain cluster data in the grid to obtain the deployment of the service. The parameters required for the service and the deployment sequence of the required clusters are used to generate and send intra-cluster instructions based on the parameters required to deploy the service and the deployment sequence; and the cluster agent device is connected to the control platform device for communication , used to receive and parse the instructions in the cluster, deploy the service through the execution of the exec command, and return the deployment results to the control platform device so that the control platform device can confirm whether the service deployment is completed. .

In the above-mentioned system, the control platform device includes: a deployment management module, which is used to analyze the content of the service to obtain the parameters required to deploy the service, and based on the parameters required to deploy the service and the deployment sequence. Generate the intra-cluster command; the cluster grid module is used to obtain the cluster data in the grid to concatenate the deployment sequence after confirming the order and relationship of the clusters required to deploy the service; and the message module, It is used to transmit the intra-cluster command and receive the deployment result returned by the cluster agent device and forward it to the deployment management module.

In one embodiment, the message module is used to manage message sending and transmission between clusters, and the messages transmitted between the clusters include synchronous messages and/or asynchronous messages.

In the above system, the cluster agent device includes: a deployment module, which is used to parse the instructions in the cluster to deploy the service; an execution module, which is used to execute the local execution of the instructions in the cluster. The execution of the program; and the delivery and reception module are used to receive deployment messages from the control platform device, send driver messages between the cluster agent devices, and return the deployment results to the control platform device.

In one embodiment, the deployment module confirms that the instructions in the cluster belong to the deployment instructions executed by the local end or the deployment instructions executed by the remote end, so that when the instructions in the cluster belong to the deployment instructions executed by the remote end, make The message module calls the cluster of the remote terminal so that the deployment command executed by the remote terminal is executed in its corresponding cluster.

In one embodiment, the message reception between the control platform device and the cluster agent device is asynchronous, and the message drive between the cluster agent devices is synchronized or driven by the Web Application Programming Interface (Web) of the cluster agent device. API) for paging.

In one embodiment, the deployment results include deployment results of all clusters on the local side and the remote side.

The present invention further discloses a method for deploying a high-availability service, which is executed by computer equipment. The method includes the following steps: causing the deployment management module of the control platform device to analyze the content of the service to be deployed to obtain the requirements for deploying the service. Parameters; make the cluster grid module of the control platform device provide cluster data in the grid, confirm the order and relationship of the clusters required to deploy the service, and connect the deployment sequence; make the deployment management module deploy the The parameters required by the service and the deployment sequence generate intra-cluster instructions for the message module of the control platform device to transmit the intra-cluster instructions; the deployment of the cluster agent device The module parses the instructions in the cluster to deploy the service, and the execution program module of the cluster agent device performs the execution of the execution program; and causes the pass-and-receive module of the cluster agent device to return the deployment results to the control The platform device allows the control platform device to confirm whether the service deployment is completed.

In one embodiment, the message module is used to manage message sending and transmission between clusters, and the messages transmitted between the clusters include synchronous messages and/or asynchronous messages.

In one embodiment, the message reception between the control platform device and the cluster agent device is asynchronous, and the message drive between the cluster agent devices is synchronized or driven by the Web Application Programming Interface (Web) of the cluster agent device. API) for paging.

In the aforementioned method, the step of parsing the intra-cluster command by the deployment module of the cluster agent device to deploy the service includes confirming that the intra-cluster command is a deployment command executed locally or remotely. , when the command in the cluster is a deployment command executed by the remote end, the message module is made to call the cluster of the remote end, so that the deployment command executed by the remote end is executed in its corresponding cluster.

In the above method, the step of causing the transmission and reception module of the cluster agent device to return the deployment result to the control platform device so that the control platform device can confirm whether the service deployment is completed means that the control platform device uses the message The module receives the deployment result and forwards it to the deployment management module to determine whether the deployment is completed.

In the foregoing method, the deployment result includes the deployment results of all clusters on the local side and the remote side.

The invention further discloses a computer-readable medium, which is used in a computing device or a computer and stores instructions to execute the aforementioned high-availability service deployment method.

It can be seen from the above that the high-availability service deployment system and method of the present invention, specifically, the control platform device considers the content of the service to be deployed and the cluster data of all clusters in the grid, and finds out the information involved in this service deployment. The cluster arranges the deployment order, and then generates intra-cluster instructions based on the obtained parameters and deployment order. The cluster agent device obtains the intra-cluster instructions and parses them. If it is a deployment instruction to be executed locally, the cluster agent device will execute it. If it is a deployment command to be executed by the remote end, the corresponding cluster agent device will be notified through the control platform device to execute it. After all cluster agent devices have completed their own deployment, the deployment status will be reported to the control platform device for confirmation. Deployment results. It can be seen from the above that the present invention uses a distributed deployment method to simultaneously activate user services in multiple clusters in the service grid, thereby achieving high-availability service backup and offloading effects.

1: Highly available service deployment system

11:Control platform device

111: Deployment management module

112: Cluster grid module

113:Message module

12, 12’: Cluster proxy device

121, 121’: Deployment module

122, 122’: Execution module

123, 123’: transmitting and receiving module

501-507:Process

601-606:Process

701-706:Process

S401-S405: Steps

Figure 1 is an architectural diagram of a high-availability service deployment system of the present invention.

FIG. 2 is an architecture diagram of a high-availability service deployment system according to an embodiment of the present invention.

FIG. 3 is a schematic diagram of the implementation of the high-availability service deployment system of the present invention.

Figure 4 is a step diagram of the deployment method of a high-availability service of the present invention.

Figure 5 is a flow chart of a method for deploying high-availability services according to the present invention in a specific embodiment.

Figure 6 is a flow chart of the control platform device constructing and deploying clusters of the present invention.

FIG. 7 is a flow chart of internal deployment of the cluster agent device of the present invention.

The following describes the technical content of the present invention through specific embodiments. Those familiar with the art can easily understand the advantages and effects of the present invention from the content disclosed in this specification. However, the present invention can also be implemented or applied through other different specific implementation forms.

Figure 1 is an architectural diagram of a high-availability service deployment system of the present invention. The present invention is based on a service mesh network architecture, uses cloud container platform services as a cluster architecture, and uses a distributed deployment method to simultaneously activate user services in multiple clusters within the service mesh to achieve High Availability (High Availability) service backup and offloading effect. As shown in the figure, the deployment system 1 of high availability services includes a control platform device 11 and a cluster agent device 12. Among them, the control platform device 11 and the cluster agent device 12 communicate with each other. Connection, the control platform device 11 mainly controls the service to be deployed. Because the service is highly available, it will be deployed in several clusters, and the cluster agent device 12 represents one of several clusters, and mainly receives data from Control the messages of the platform device 11 to perform service installation.

The control platform device 11 is used to analyze the content of the service to be deployed to obtain the parameters required to deploy the service. In addition, it also obtains cluster data in the service mesh to obtain the parameters required to deploy the service. cluster, to obtain the deployment sequence of the cluster based on the relationship and ordering between the clusters. Finally, generate intra-cluster instructions based on the parameters required to deploy the service and the deployment sequence, and then send the intra-cluster instructions for clustering The agent device 12 receives the intra-cluster command.

The cluster agent device 12 analyzes the command after receiving the command in the cluster, deploys the service through the execution of the exec command, and returns the deployment result to the control platform device 11 after the deployment is completed for the control platform device. 11 Confirm whether the service is deployed. In particular, the intra-cluster instructions received by the cluster agent device 12 will include instructions to be executed by the local end and the remote end. The local end refers to the cluster that received the intra-cluster instruction, and the remote end refers to the cluster that received the intra-cluster instruction. Other clusters other than the cluster of instructions in the cluster. In other words, each cluster will only execute what it can process or needs to process. For other related instructions, the control platform device 11 will then notify the corresponding cluster that needs to process the instruction. The technical details of this part will be described in detail later.

In one embodiment, the message reception between the control platform device 11 and the cluster agent device 12 is asynchronous, and the message drive between the cluster agent devices is synchronous or driven by the Web Application Programming Interface (Web API) of the cluster agent device. ) to page. In other words, after the control platform device 11 sends the deployment message including the instructions in the cluster, the cluster agent device 12 receives it by itself. At this time, the control platform device 11 does not need to wait for the response of the cluster agent device 12 to perform other tasks. Therefore, both The messaging between them is asynchronous. In addition, the cluster agent device 12 may include multiple agent devices that act as agents for different clusters. During the deployment and installation process, agent device A may require agent device B to perform certain procedures before continuing the installation. In other words, if agent device A does not receive Before agent device B responds, it cannot continue to perform subsequent operations, that is, agent device A must wait for a response from agent device B, so the message transmission between agent devices is synchronous.

In short, when exchanging data, if it is important data (such as a key), it can be carried out through the designated message transmission module, but if it is general data, if it is under a mesh network (mesh) architecture, If two cluster networks communicate with each other, they can communicate directly via API (for example, server API). If the two cluster networks do not communicate with each other, they will communicate through a designated message transmission module. Therefore, the control platform device 11 knows the cluster from the beginning. After the mesh relationship between them is issued, the communication method between the two clusters can be determined.

In one embodiment, the deployment results returned by the cluster proxy device 12 include the deployment results of all clusters at the local end and the remote end. In short, all cluster agent devices 12 need to return deployment results to know the deployment result of this service. For example, the control platform device 11 sends a deployment message, which is received and executed by agent device A and agent B, where , if agent device A is local, agent device B can be called a remote terminal. In addition, if agent device A also requires agent device C to perform a certain configuration before installation, agent device A has to wait for agent device C. the reply, so for agent device A Specifically, the deployment results to be reported include the deployment results of agent device C and the deployment results of agent device A itself. Therefore, the deployment results are collectively referred to as covering all clusters on the local and remote ends.

In terms of the overall architecture, the control platform device 11 is a system platform cluster, and the cluster agent device 12 is a cluster in the service grid. In a specific embodiment, the control platform device 11 is a Kubernetes cluster that the system mainly executes, and the cluster agent device 12 Kubernetes cluster within the service mesh that users need to deploy. The technical details of the control platform device 11 and the cluster agent device 12 are further described below respectively.

FIG. 2 is an architectural diagram of a high-availability service deployment system according to an embodiment of the present invention. As shown in the figure, when the user wants to deploy a Kubernetes cluster in the service grid, the control platform device 11 performs service deployment and management. In addition, the corresponding cluster agent device 12 also needs to be driven to perform deployment, as shown in the figure. As shown, the control platform device 11 includes a deployment management module 111, a cluster grid module 112 and a message module 113, and the cluster agent device 12 includes a deployment module 121, an execution module 122 and a delivery and reception module 123. .

The deployment management module 111 is used to analyze the content of the service to be deployed to obtain the parameters required to deploy the service, and to generate instructions in the cluster according to the parameters and deployment sequence required to deploy the service. Specifically, the deployment management module 111 is responsible for analyzing the deployment content, required instructions and actions required for service startup, including the execution corresponding to deployment (kubernetes instructions) or execution sequence (exec command). The deployment management module 111 obtains Parameters required to deploy the service to be deployed, as well as cluster relationships within the grid provided by the cluster grid module 112, analyze which clusters are to be executed for this deployment, and confirm the network relationships and message communication between them The mode includes asynchronous message drivers, synchronous message drivers and service Restful API calls of the cluster agent device 12, etc., and finally generates intra-cluster instructions. The aforementioned Restful API refers to conforming to or being compatible with the presentation layer state transfer (REST). ) of a web server or web browser.

The cluster grid module 112 is used to obtain the cluster data in the grid, so as to concatenate the deployment sequence after confirming the order and relationship of the clusters required to deploy the service. As mentioned above, when the deployment management module 111 generates intra-cluster instructions, it needs to understand the clusters involved in the deployment service and the relationships between these clusters. Therefore, the cluster grid module 112 is used to obtain the clusters in the grid. Data, and after confirming the clusters required to deploy the service and the order and relationship of these clusters, the cluster deployment sequence for deploying the service is concatenated.

The message module 113 is used to transmit intra-cluster instructions generated by the deployment management module 111, and receive the deployment results returned by the cluster agent device 12 and forward them to the deployment management module 111. In short, the message module 113 manages the sending and delivery of messages between clusters, including synchronous messages and/or asynchronous messages. Among them, synchronous messages involve transmission by both parties and have a timeout time, that is, the caller After sending a message, the caller will wait for a response from the other party before proceeding to the next step. When asynchronous messages are transmitted through a channel, the caller can handle other tasks in a short period of time after sending the message without waiting for a response from the other party. Using Which can be adjusted according to design requirements; in addition, all clusters need to be registered to the message module 113 to receive or transmit messages.

The deployment module 121 is used to parse the instructions in the cluster from the control platform device 11 to deploy services. In short, after receiving the driver message sent by the control platform device 11, the deployment module 121 will execute the container command or the kubernetes command. As mentioned above, these commands may include the local side (i.e., after receiving the message cluster) and the instructions to be executed by the remote end (non-local end).

The execution program module 122 is used to execute the local execution program in the instructions in the cluster. Specifically, the executor module 122 is used to execute the exec command using host entity instructions, that is, the local service installation in the instructions in the cluster will be handed over to the local device (such as a computer host). , server) to execute, at this time, the installation is performed through the internal entity instructions of the device.

The transmission and reception module 123 is used to receive deployment messages from the control platform device 11 , transmit driving messages between cluster agent devices, and return deployment results to the control platform device 11 . In short, the transmission and reception module 123 mainly receives the deployment information of the control platform device 11 and is used to transmit the driving messages between the cluster agent devices. When the cluster agent device 12 completes the entire deployment, the deployment result is returned. The control platform device 11, as mentioned above, the message reception between the control platform device 11 and the cluster agent device 12 is asynchronous, and the message driving between the cluster agent devices is synchronous or called by the cluster agent device service Restful API. This speeds up execution between deployment clusters.

In one embodiment, the deployment module 121 confirms that the instructions in the cluster belong to the deployment instructions executed by the local end or the deployment instructions executed by the remote end, so that when the instructions in the cluster belong to the deployment instructions executed by the remote end, the control The message module 113 of the platform device 11 makes a call to the cluster at the remote end, so that the deployment command executed by the remote end is executed in its corresponding cluster. Therefore, deployment messages are sent to different cluster agent devices 12 through the message module 123 of the control platform device 11 .

FIG. 3 is a schematic diagram of the implementation of the high-availability service deployment system of the present invention. As shown in the figure, after the deployment management module 111 generates an intra-cluster command for service deployment, the message is sent to the message module 113. After receiving the deployment message from the delivery and reception module 123, the cluster agent device 12 notifies the deployment module. Group 121 analyzes the instructions in the cluster, and transfers the instructions belonging to the local end to the execution module 122 for execution. If the deployment module 121 finds that there are instructions from the remote end in the instructions in the cluster, it notifies the message module 113 and transfers them to the execution module 122. After the corresponding cluster agent device 12' is received by the delivery and reception module 123' of the cluster agent device 12', it is transferred to the execution module 122' for execution, and the deployment result is returned to the cluster agent device 12. Similarly, the cluster When the deployment module 121' of the agent device 12' discovers that there is a remote command in the cluster, it will transfer it to the corresponding agent device for processing. After the deployment and installation are completed, the delivery and reception module 123 will The deployment result is returned to the message module 113, and the message module 113 then forwards it to the deployment management module 111 to provide the deployment management module 111 with the deployment status of the service.

Figure 4 is a step diagram of the deployment method of high-availability services of the present invention. It illustrates that when deploying high-availability services, the sharing concept can be used to deploy high-availability services in multiple clusters through clusters and service parsing, so as to achieve The effect of remote backup is that in the architecture of the present invention, the control platform device communicates with the control platform device, the control platform device performs deployment planning and management, and the cluster agent device performs individual deployment and installation. The following will describe the high-level Step-by-step instructions for deploying available services.

Step S401, causing the deployment management module of the control platform device to analyze the content of the service to be deployed to obtain parameters required for deploying the service. This step explains that the deployment management module analyzes the deployment content, required instructions and actions required for service startup, so as to obtain the parameters required to deploy the service to be deployed.

Step S402: Instruct the cluster grid module of the control platform device to provide cluster data in the grid, confirm the order and relationship of the clusters required for deploying the service, and connect the deployment sequence. This step explains that the cluster grid module will obtain the cluster relationships within the grid, then analyze the clusters involved in the secondary deployment and confirm the network relationships and message communication modes between these clusters. In short In other words, find out the clusters to which the services are deployed, and understand the relationships and communication methods between these clusters as a basis for deployment.

Step S403, causing the deployment management module to generate an intra-cluster command based on the parameters required for deploying the service and the deployment sequence, so that the information module of the control platform device can transmit the intra-cluster command. This step explains how the deployment management module generates intra-cluster instructions based on the parameters required for the deployment service obtained in the first two steps and the deployment sequence between the involved clusters. These intra-cluster instructions are the driver agents. The device is used for deployment and installation. After that, the commands in the cluster are sent by the message module of the control platform device.

Step S404: The deployment module of the cluster agent device parses the instructions in the cluster to deploy the service, and the execution program module of the cluster agent device executes the execution program. This step shows that the cluster agent device can receive the deployment message containing the instructions in the cluster sent by the message module of the control platform device through its internal transmission and reception module. Afterwards, the deployment module will analyze it to obtain the information to be carried out. Deploy the installation content and execute the execution program by the execution module.

In one embodiment, when the deployment module parses the instructions in the cluster and deploys the service, it further includes confirming that the instructions in the cluster belong to the deployment instructions executed locally or remotely, so that the instructions in the cluster belong to When a deployment command is executed by the remote end, the message module is caused to call the cluster of the remote end, so that the deployment command executed by the remote end is executed in its corresponding cluster. In short, the instructions in the cluster may include local and remote instructions. If it is a local instruction, it can be directly executed by the cluster agent device that receives the instruction in the cluster. If it is a remote instruction, the control platform device will be notified. The message module makes a call to the remote terminal's cluster, and then transfers the remote terminal's instructions so that the deployment instructions executed by the remote terminal can be executed in the corresponding cluster.

Step S405, causing the transmission and reception module of the cluster agent device to transmit the deployment result back to the control platform device, so that the control platform device can confirm whether the service deployment is completed. This step means that after the deployment module completes the deployment and installation of its own cluster, it will generate a deployment result and can transmit the deployment result back to the control platform device through the delivery and reception module, so that the control platform device can confirm the deployment status of the service. In one embodiment, the control platform device receives the deployment result through its message module and forwards it to the deployment management module to determine whether the deployment is completed.

In other embodiments, the message module mainly manages the sending and transmission of messages between clusters, and the messages transmitted between the clusters include synchronous messages and/or asynchronous messages. Specifically, the message reception between the control platform device and the cluster agent device is asynchronous, that is, the control platform device After sending the deployment message, you do not need to wait for a reply from the cluster agent device before you can execute other procedures. In addition, the message driver between the cluster agent devices is synchronized or performed by the Web Application Programming Interface (Web API) of the cluster agent device. Paging, in short, in order to ensure communication between each agent device, each agent device uses a synchronous method to transmit messages, and communication can also be carried out through the Web Application Programming Interface (Web API) of each agent device.

FIG. 5 is a flow chart of a method for deploying high-availability services according to the present invention in a specific embodiment.

In process 501, parse service. First of all, Kubernetes (K8S) is a system that can manage microservices (microservices). It can automatically deploy and manage containers (Containers) on multiple machines. Helm is a tool for managing configuration files, and a service may be composed of multiple machines. A helm chart is generated together. This process analyzes the contents of the chart to confirm parameters that may need to be replaced due to different environments.

In process 502, clusters are deployed in the grid to string out cluster sequences. This process explains that when a service is deployed in a service grid, the cluster data in the grid will be obtained first, the primary cluster and the secondary cluster will be found, the cluster ordering and relationship will be confirmed, and the deployment sequence of the clusters will be connected accordingly. , where the deployment architecture can be divided into a main cluster (Master-Master architecture) or primary and secondary clusters of different levels (Master-Slave architecture).

In process 503, a deployment instruction is generated. This process refers to parsing the parameters in the service, such as supporting helm command (helm command), kubernets application programming interface (kubernets api), linux shell command (linux shell command), etc., and placing the command sequence according to the service settings and adding the required correspondence. Which cluster performs the join to generate intra-cluster instructions.

In process 504, a deployment command is sent to each agent device for the cluster sequence. This process describes the use of message content packaging including deployment instructions (i.e. instructions within the cluster) for cluster subscription messages. Channel sending can be sent to the message management center (that is, the message module of the present invention), using asynchronous messages. When the agent device in the cluster receives the message, the platform completes sending the message.

In process 505, the agent device is deployed and executed. This process explains that after the agent device in the cluster receives the message, it analyzes it into instructions that are executable in the cluster and not executable in the cluster. The deployment instructions executed directly in the cluster include helm command (helm command), kubernets application programming interface (kubernets api) ), linux shell command, etc. Instructions that are not executable in the cluster are processed in the next process.

In process 506, multiple agent devices execute deployment instructions with each other. This process explains that instructions that are not executable in this cluster may be instructions that need to read or obtain data related to another cluster. Synchronous message transmission can be used to drive the execution of another cluster to obtain the result and obtain the data. The result of this instruction is compared with the original cluster. The results of the instructions within are combined to produce the deployment result of this cluster agent device.

In process 507, all cluster information is collected. This process describes the return of deployment results from the agent device to the control platform device. After the control platform device collects all the deployment results, it determines whether to complete the deployment or change of a service.

Figure 6 is a flow chart of the control platform device constructing and deploying clusters of the present invention. This flow chart is the process that users perform on the main control platform to deploy or update a service. Each process is explained in detail below.

In process 601, parse service. In this process, when users need to deploy or update on the edge, they first perform service analysis to confirm how many services are updated, how to deploy them, and which ones need to be adjusted using the kubernets application programming interface (kubernets api). Those require the use of host commands.

In process 602, the clusters deployed in the grid are parsed to string out the cluster sequence. This process is to analyze which service grid the edge terminal for service installation is located in. It uses the cluster information that has been set in the grid to analyze the primary and secondary cluster relationships, cluster content, and the clusters and sequences that need to be deployed.

In process 603, a deployment instruction is generated. This process is to generate the instruction set required to deploy the service by correlating the information obtained by the above two processes. There are agent devices inside corresponding to various connectors (adaptors), such as using Restful API, kubernetes instructions, and host linux shell commands (host linux shell command) and library common language architecture (lib cli) commands.

In process 604, a deployment command is sent to each agent device for the cluster sequence. This process uses the instruction set generated in process 603 to issue an asynchronous message notification for the cluster sequence in the grid, and issues a subscription to the cluster message sent, waits for a reply from this subscription, and changes the service status to deploying, which is stored in the data in the library.

In process 605, asynchronously waits for the agent device to respond to the deployment result. When the service is sent out in process 604, this process will enter a waiting state, waiting for all proxy devices in the cluster to reply messages.

In process 606, the service deployment is parsed and the deployment is completed. After receiving the reply from the agent device, this process needs to parse the reply content of each cluster agent, analyze all clusters in the grid for this service, and confirm whether a service deployment has been completed.

The technical content of the present invention will be explained below through the description of a specific embodiment.

When you want to create a highly available mongo database in two clusters, you can set one cluster as primary and the other cluster as secondary, and the created service content can correspond to the helm chart file directory, that is, add parsable adapter instructions in values.yaml (yaml is a text format that describes settings), which can execute attributes before and after deploying the service. At the virtual machine (VM) level instructions, kubectl instructions, etc., the above content is packaged into a helm chart file, uploaded to the control platform device, and the plan is created after analysis.

The control platform device sends a message to the two clusters. The content is the aforementioned command. After cluster A receives the message, it parses it in sequence, assigns the command executor, and completes the deployment command. After completing the deployment command, it replies the message to the control platform device. After cluster B receives the message, Analyze in sequence, assign the command executor, and after completing the deployment command, reply the message to the control platform device. The control platform device receives the cluster A message, retrieves the deployment data stored in the database, and updates the deployment target status to complete deployment. The control platform device receives the cluster B message, retrieves the deployment data stored in the database, and updates the deployment target status to complete. deploy. After the platform device receives the message from the two clusters, it can notify the user that the deployment is completed.

FIG. 7 is a flow chart of internal deployment of the cluster agent device of the present invention. This flowchart is a process to be executed on the cluster agent device. Each process is described in detail below.

In process 701, the agent device receives the deployment message. This process means that the cluster agent device can receive the message, that is, after the cluster joins the platform to establish a grid, it will establish a connection with the message module on the control platform device and wait for the message to be sent.

In process 702, the message is parsed to confirm local execution. This process explains that when a deployment message is received, it will enter the message parsing to confirm whether it is an instruction to be executed in this cluster. If not, the message module will be used to call the process 703. If it is, the process will be entered. 704.

In process 703, other cluster devices execute the deployment command synchronously and respond. This process means that when the instruction requires other clusters to be executed, the synchronization message function of the message module will be used as the call mode to specify the cluster call. At this time, the called cluster agent will perform its agent processing, that is, similar to the process 702. After the message, the local deployment command execution and other procedures of process 704 are performed to reply to the calling end cluster after completion.

In process 704, the local terminal executes the deployment instruction. This process refers to the deployment instructions that need to be executed on the local side of the agent. After all instructions on the local side of the agent are completed, process 705 is entered.

In process 705, the instruction is confirmed to be completed and the deployment status is changed. This process must wait for the remote end and the local end to complete the reply before judging the deployment service status of the entire cluster. If one of the replies fails, it is judged that the deployment of the entire cluster has failed. If all replies are successful, it is judged that the deployment of the entire cluster has failed. The deployment is successful and proceeds to process 706.

In process 706, the agent device sends a deployment message. In this process, the agent device is responsible for replying the last deployed message to the message module and completing all tasks of the agent device.

In addition, the present invention also discloses a computer-readable medium, which is applied to a computing device or computer having a processor (eg, CPU, GPU, etc.) and/or a memory, and stores instructions, and can utilize the computing device or computer. The computer executes the computer-readable medium through the processor and/or memory to perform the above methods and steps when executing the computer-readable medium.

The modules, units, devices, etc. of the present invention include a microprocessor and a memory, and algorithms, data, programs, etc. are stored in the memory or chip, and the microprocessor can load data or algorithms or programs from the memory. Processing such as data analysis or calculation will not be described in detail here. In other words, the high-availability service deployment system of the present invention can be executed on an electronic device, such as a general computer, a tablet, or a server, and performs analysis and calculations after receiving the data. Therefore, the high-availability service deployment system performs the procedures , can be designed and constructed through software on electronic devices with processors, memories and other components to run on various electronic devices; in addition, each module or unit of the high-availability service deployment system can also be independently configured Components, such as those designed as calculators, memories, storages or firmware with processing units, can all become components for implementing the present invention. That is, the present invention can be presented in the form of software programs, hardware or firmware architectures.

In summary, the high-availability service deployment system and method of the present invention are based on the service grid network architecture, use the cloud container platform service as the cluster architecture, and use the distributed deployment method to simultaneously launch user services on the service network. Multiple clusters in the grid can achieve high-availability service backup and offloading effects. When the same service is deployed in multiple clusters, different settings are required to use the master-master architecture or master-slave architecture. (Master-Slave) architecture, therefore, the present invention utilizes the differentiation of deployment instructions. The agent device can drive other cluster agent devices to adjust settings by receiving and sending instructions to other clusters or interoperating with RESTful APIs in the service grid to complete the Service high availability deployment mechanism. Therefore, the present invention has the following effects.

First, the present invention connects all clusters using a service grid and uses the extension of network connectivity to exchange deployment messages, thereby making clusters relevant.

Secondly, the present invention only needs to target one cluster in the service grid for deployment and update, and can be extended to other clusters to achieve a distributed deployment effect.

Furthermore, the present invention uses the agent device deployment program to support Restful API, kubernetes commands, host linux shell commands, and library common language architecture (lib cli) commands according to different service requirements. Add command requirements inside, and you can also inject commands into the agent device.

In addition, the cluster agent device of the present invention can complete self-cluster deployment, and can also drive other cluster agent devices to obtain relevant data through synchronization messages to assist the original agent device to complete the configuration.

The above embodiments are only illustrative and not intended to limit the present invention. Anyone skilled in the art can modify and modify the above embodiments without departing from the spirit and scope of the invention. change. Therefore, the scope of rights protection of the present invention is defined by the scope of the patent application attached to the present invention. As long as it does not affect the effect and implementation purpose of the present invention, it should be covered by this disclosed technical content.

11:Control platform device

111: Deployment management module

112: Cluster grid module

113:Message module

12:Cluster proxy device

121:Deploy module

122: Execution module

123:Transmitting and receiving module

Claims (14)

A high-availability service deployment system includes: The control platform device is used to analyze the content of the service to be deployed and obtain the cluster data in the grid to obtain the parameters required to deploy the service and the deployment sequence of the required clusters, so as to deploy the service according to the parameters required and The deployment sequence generates intra-cluster instructions and sends the intra-cluster instructions; and The cluster agent device communicates with the control platform device and is used to receive and parse the instructions in the cluster, deploy the service through execution of the execution program, and return the deployment results to the control platform device for the control. The platform device confirms whether the service deployment is complete. The high-availability service deployment system as described in claim 1, wherein the control platform device includes: The deployment management module is used to analyze the content of the service to obtain the parameters required to deploy the service, and to generate instructions in the cluster based on the parameters required to deploy the service and the deployment sequence; The cluster grid module is used to obtain the cluster data in the grid to concatenate the deployment sequence after confirming the order and relationship of the clusters required to deploy the service; and The message module is used to transmit instructions within the cluster, receive the deployment result returned by the cluster agent device, and forward it to the deployment management module. The deployment system for high-availability services as described in request 2, wherein the message module is used to manage the sending and delivery of messages between clusters, and the messages passed between the clusters include synchronization messages and/or Asynchronous messages. The high-availability service deployment system as described in claim 1, wherein the cluster agent device includes: The deployment module is used to parse the instructions in the cluster to deploy the service; The execution module is used to execute the local execution sequence of the instructions in the cluster; and The transmitting and receiving module is used to receive deployment messages from the control platform device, transmit driving messages between the cluster agent devices, and return the deployment results to the control platform device. The deployment system for high-availability services as described in claim 4, wherein the deployment module confirms that the instructions in the cluster are deployment instructions executed locally or remotely, so that the instructions in the cluster are When the remote terminal executes a deployment command, the message module is caused to call the cluster of the remote terminal, so that the deployment command executed by the remote terminal is executed in its corresponding cluster. The deployment system for high-availability services as described in claim 1, wherein the message reception between the control platform device and the cluster agent device is asynchronous, and the message drive between the cluster agent devices is synchronous or driven by the cluster. Proxy device's Web Application Programming Interface (Web API) for paging. The high-availability service deployment system as described in claim 1, wherein the deployment results include the deployment results of all clusters on the local side and the remote side. A method for deploying high-availability services is executed by computer equipment. The method includes the following steps: Let the deployment management module of the control platform device analyze the content of the service to be deployed to obtain the parameters required to deploy the service; Let the cluster grid module of the control platform device provide the cluster data in the grid, confirm the order and relationship of the clusters required to deploy the service, and connect the deployment sequence; Cause the deployment management module to generate an intra-cluster command based on the parameters required to deploy the service and the deployment sequence, so that the message module of the control platform device can transmit the intra-cluster command; The intra-cluster instructions are parsed by the deployment module of the cluster agent device to deploy the service, and the execution program is executed by the execution module of the cluster agent device; and The transmitting and receiving module of the cluster agent device is caused to transmit the deployment result back to the control platform device, so that the control platform device can confirm whether the service deployment is completed. The method of deploying a high-availability service as described in request 8, wherein the message module is used to manage message sending and transmission between clusters, and the messages transmitted between the clusters include synchronization messages and/or Asynchronous messages. The method for deploying high-availability services as described in claim 8, wherein the message reception between the control platform device and the cluster agent device is asynchronous, and the message drive between the cluster agent devices is synchronous or driven by the cluster. Proxy device's Web Application Programming Interface (Web API) for paging. The method of deploying a high-availability service as described in claim 8, wherein the step of parsing the intra-cluster command by the deployment module of the cluster agent device to deploy the service includes confirming that the intra-cluster command belongs to local execution. The deployment command or the deployment command executed by the remote end, so that when the command in the cluster is a deployment command executed by the remote end, the message module is called to the cluster of the remote end, so that the deployment command executed by the remote end Executed within its corresponding cluster. The method of deploying a high-availability service as described in claim 8, wherein the transmitting and receiving module of the cluster agent device is caused to transmit the deployment result back to the control platform device, so that the control platform device can confirm whether the service deployment is completed. Step means that the control platform device receives the deployment result through the message module and forwards it to the deployment management module to determine whether the deployment is completed. The method of deploying a high-availability service as described in claim 8, wherein the deployment results include the deployment results of all clusters on the local side and the remote side. A computer-readable medium, used in a computing device or computer, stores instructions to execute the method for deploying a high-availability service as described in any one of claims 8 to 13.

Images