CN111538569B

CN111538569B - Cloud platform-based system one-key deployment method

Info

Publication number: CN111538569B
Application number: CN202010371770.2A
Authority: CN
Inventors: 谢成冬; 李耀; 田骏; 夏咸辉
Original assignee: Wuhan Zhongbang Bank Co Ltd
Current assignee: Wuhan Zhongbang Bank Co Ltd
Priority date: 2020-05-04
Filing date: 2020-05-04
Publication date: 2023-04-28
Anticipated expiration: 2040-05-04
Also published as: CN111538569A

Abstract

The invention discloses a cloud platform-based system one-key deployment method, which relates to the technical field of information and aims to provide a cloud platform-based system one-key deployment method capable of rapidly copying IaaS resources of a service system contained in a target regression testing environment and keeping an IP address unchanged. The method comprises the following steps: establishing a source VPC network in a cloud platform; migrating a virtual machine where the existing system is located to a source VPC network established by 1; establishing a clone VPC network in a cloud platform; the VPC router data, the virtual subnet data and the load balancing data in the source VPC network are collected through a resource arrangement unit, and the data are manufactured into a resource arrangement template; executing a resource arrangement template in the clone VPC network, and transmitting the VPC router data, the virtual subnet data and the load balancing data in the resource arrangement template to the clone VPC network; the virtual machines of the required systems in the source VPC network are copied into virtual subnets in the clone VPC network.

Description

Cloud platform-based system one-key deployment method

Technical Field

The invention relates to the field of network development, in particular to a system one-key deployment method based on a cloud platform.

Background

Currently, more than 100 sets of business systems are operated in the development and test cloud of our company, and the systems are usually supported and operated by different numbers of virtual machines, and the total number of the virtual machines in the development and test cloud is more than 1000.

Every time a new version of a business system is released, the relevant project group organizes to develop a regression test to verify whether the previously discovered and repaired errors reappear on the new software version. In addition, not only is the step retest performed when the defect is present, but all functions that may be affected are tested when the repaired defect is verified. Therefore, in daily development testing, whenever a certain project group proposes to perform regression testing on a certain service system that is responsible for the project group, IT is often required that the IT infrastructure portion newly create a complete set of regression testing environments for the project group in the cloud platform, where the tested subject service system and the service systems that are responsible for other project groups associated with the functions of the tested subject service system are included. The test case is then re-executed to confirm that the modification did not introduce new errors or caused other business functions to produce errors.

Currently, assuming that the service system a needs to perform regression testing, the service system associated with the service system a may include A, B, C three service system functions, and a target test environment is created in the existing development test cloud platform, which generally needs to undergo the following steps:

1. a cloud platform manager re-creates corresponding IaaS resources, such as a virtual machine, a basic network and the like, for the service A, B, C according to the service utilization condition of the cloud resources of the service A, B, C in the existing development test cloud platform;

2. three development and test teams of service A, B, C log in to a newly built virtual machine, re-build and configure a service environment meeting test requirements, which may include complex and repetitive operations such as software installation, code compilation, service data copying, configuration file modification, performance optimization, and the like.

In the step 2 above, a plurality of people who need to invest A, B, C three business system development test teams take a few weeks to finish, and consistency of a newly built test environment and business versions and data existing in an original development test cloud platform cannot be guaranteed. If the service a system to be tested is associated with more than one other service system, for example, 20 service systems or more, this means that 20 service system-related project group manpower may be put into.

Regression testing is used as a component of the software lifecycle, and takes up a great amount of work in the whole software testing process, and multiple regression tests are performed at each stage of software development. In progressive and rapid iterative development, the continued release of new versions makes regression testing more frequent. If the tested several service systems are changed frequently, and the system a, the system B and the system C are likely to be the same today, then building a similar regression testing environment consumes a lot of repeated manpower investment which cannot be estimated, and the cost increases exponentially.

In view of the foregoing, as a foundation for operation of modern IT infrastructure, a cloud platform needs to provide a set of mechanisms that can quickly copy or clone IaaS resources (including but not limited to virtual machines, network configuration, firewall policies, etc.) of a service system included in a target regression testing environment based on an existing running development test virtual machine, while keeping an IP address unchanged; and the gap between data before and after the replication or cloning of the storage volume of the virtual machine is reduced as much as possible, so that the service system after the replication or cloning can be tested without logging in to modify configuration and repair service. In addition, the cloud platform should also be able to ensure that the newly created suite of test environments does not conflict with existing or other newly created test environments and is easy to manage and maintain.

Disclosure of Invention

The invention aims to solve the problems and provides a cloud platform-based system one-key deployment method capable of rapidly copying IaaS resources of a service system contained in a target regression testing environment and keeping an IP address unchanged.

In order to solve the technical problems, the invention adopts the following technical scheme:

a cloud platform-based system one-key deployment method is characterized by comprising the following steps of: the method comprises the following steps:

s1, establishing a source VPC network in a cloud platform;

s2, migrating a virtual machine where the existing system is located into a source VPC network established in the step S1;

s3, establishing a clone VPC network in the cloud platform;

s4, collecting VPC router data, virtual subnet data and load balancing data in a source VPC network, and manufacturing the data into a resource arrangement template;

s5, executing a resource arrangement template in the clone VPC network, and transmitting VPC router data, virtual subnet data and load balancing data in the resource arrangement template to the clone VPC network;

s6, copying a virtual machine of a system required in the source VPC network into a virtual subnet in the clone VPC network;

s7, detecting the system copied into the clone VPC network to confirm that the system works normally.

In the above technical solution, the resource allocation unit in step S4 includes:

and a data searching module: the virtual sub-network load balancing method is used for searching the VPC router data, the virtual sub-network data and the load balancing data;

and a data acquisition module: the virtual sub-network data processing method is used for copying and collecting the searched VPC router data, virtual sub-network data and load balancing data;

and a storage module: the method comprises the steps of storing collected VPC router data, virtual subnet data, load balancing data and a manufactured resource arrangement template;

a data arrangement module: the virtual sub-network data and the load balancing data are arranged according to the sequence;

and a template manufacturing module: the virtual sub-network data and the load balancing data are used for making the arranged VPC router data, the virtual sub-network data and the load balancing data into a resource arrangement template.

In the above technical solution, the step S4 includes the following steps:

s41, enabling a data searching module to enter a system of a source VPC network, finding a default storage position of VPC router data, virtual subnet data and load balancing data in the source VPC network, entering the default storage position to search the VPC router data, virtual subnet data and load balancing data in the source VPC network, and transmitting the found VPC router data, virtual subnet data and load balancing data to a data acquisition module;

s42, the data acquisition module copies the found VPC router data, virtual subnet data and load balancing data, and pastes and stores the data into the storage module;

s43, the storage module transmits the stored VPC router data, virtual subnet data and load balancing data to the data arrangement module, the data arrangement module arranges the data according to the format and sequence in the preset template to obtain a template file consistent with the format and sequence of the preset template, and the template file is transmitted to the template making module;

s44, the template making module converts the received template file into an application program configuration file, takes the application program configuration file as a resource arrangement template and stores the resource arrangement template in the storage module.

In the above technical solution, the step S6 includes the following steps:

s61, a virtual machine of a required system is found in a source VPC network through a search program and is selected;

s62, performing copy operation on the selected virtual machine;

s63, selecting a virtual subnet in the cloned VPC network as a target network path;

s64, performing pasting operation, and copying the virtual machine of the required system to a virtual subnet in the clone VPC network.

In the above technical solution, the step S7 includes the following steps:

s71, opening application programs and databases of the cloned VPC network system, and detecting whether the cloned VPC network system can be normally used;

s72, logging in a browser and a client of a system in a clone VPC network, and detecting whether the browser and the client can normally access the network;

and S73, after the detection results of the steps S71 and S72 are normal, completing the deployment operation of the system in the clone VPC network.

Compared with the prior art, the invention has the advantages and positive effects that:

1. according to the invention, the construction work of a new regression testing environment system does not need to input manpower for a relevant service system project group any more, and the manpower input of each service system department in the development testing cloud platform in the early development testing can be continued, so that the high manpower cost generated by a large amount of repeated labor which needs to be input when constructing the regression testing environment system is saved;

2. the invention enables the project group to put more energy into the development test of the new service, and improves the new version of each service system and the online period of the new service system, thereby improving the overall level and efficiency of the software development test;

3. the invention enables the cloud platform administrator to more easily identify the development test environment similar to that of managing a plurality of sets of parallel development test environments, and is convenient for the daily maintenance operation of IT operation.

Drawings

In order to more clearly illustrate the embodiments of the invention or the technical solutions of the prior art, the drawings which are used in the description of the embodiments or the prior art will be briefly described, it being obvious that the drawings in the description below are only some embodiments of the invention, and that other drawings can be obtained according to these drawings without inventive faculty for a person skilled in the art.

FIG. 1 is a schematic diagram of a VPC network;

FIG. 2 is a frame structure diagram of a resource orchestration unit;

fig. 3 is a schematic diagram of the effect after copying.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, modifications, equivalents, improvements, etc., which are apparent to those skilled in the art without the benefit of this disclosure, are intended to be included within the scope of this invention.

The invention provides a cloud platform-based system one-key deployment method capable of rapidly copying IaaS resources of a service system contained in a target regression testing environment and keeping an IP address unchanged, which can build a plurality of sets of environments based on the original environment in a time as short as possible, maximally save manpower, time and deployment configuration problems, and realize one-key deployment of a demand environment, resources and system application; the implementation process is as follows:

1. virtual machine migration to VPC

The current development and test cloud IaaS adopts a cloud platform, all virtual machines run in a basic network of the cloud platform, computing resources of the cloud platform are cloud platform virtual machines, network resources of the cloud platform are virtual VPCs, load balancing is software loads, IP address division is achieved through virtual subnets of the cloud platform, external routes are achieved through virtual routers, and a firewall is achieved through a cloud platform virtual firewall.

In order to ensure that IP addresses in a basic network are not overlapped, the cloud platform must change the IP addresses when cloning virtual machines, when virtual machines of a plurality of service systems need to be cloned, a great number of configuration change operations must be carried out on each service system project group, unrecoverable errors are extremely easy to generate, and even some systems do not support to modify the IP addresses. Therefore, in order to implement "one-key deployment", it must be ensured that virtual machines with the same IP address can coexist and have no mutual interference on the network, so that a virtual private network VPC is introduced on the cloud platform; the virtual private networks VPCs can be established in the cloud platform, and are independent, and the information such as IP addresses and the like of the virtual private networks VPCs are not affected, so that the problem that virtual machines with the same IP addresses cannot be simultaneously in the cloud platform can be solved through the independent virtual private networks VPCs.

The virtual private network VPC is a proprietary large network that can be preconfigured for users within the cloud platform environment. In the VPC network, the IP address range can be customized, the subnet can be created, and various cloud resources such as a host, a database and the like can be created in the subnet. Each VPC is composed of a router, at least one private network segment and at least one switch, and when a private network and a switch are created, virtual subnets, i.e. network IP addresses, need to be specified, the VPCs are 100% isolated, and hosts do not interfere with each other.

As shown in fig. 1, the virtual private network is similar to a local area network formed by connecting a plurality of servers together using a virtual Switch (L2 Switch) in the physical world. The private networks of different users are 100% isolated, and the private networks can be added as subnets to the VPC. The virtual router is used for interconnection among a plurality of private networks and provides a plurality of additional functions: DHCP, port forwarding, VPN, tunnel services and access control, cover the usual network configuration and management effort.

The node communication of the virtual private network is changed from the traditional tree structure into the mesh structure, and all nodes are in point-to-point direct communication, so that the communication performance among the nodes is improved, and the number of private networks which can be connected by a single router is also improved. The VPC network can ensure high performance and high availability of the network cluster under the condition of large-scale deployment.

In a VPC network, a management router is only responsible for management functions such as VPN, tunnel, DNS, port forwarding, and the like, and forwarding and routing of these management traffic, and forwarding traffic between subnets is no longer handled. The host in the VPC network can bind the own elastic IP; and a dedicated firewall is arranged, no membership exists among the IP, the firewall and the management router, and meanwhile, the load balancer can also be directly connected with a host in the VPC network.

2. Quasi-production cloud construction and test cloud transformation

The sub-production cloud platform design adopts a VPC plus boundary router (matched with traditional hardware F5) to simulate the IT infrastructure of the current bank production area.

After applying for the border router, the physical switch running the border router creates a corresponding routing instance. The boundary router can be used as a gateway between an SDN network in the VPC and a traditional physical network to perform conversion between the vxlan and the vlan, and the hardware devices such as a server, a switch, F5, a firewall and the like in the traditional physical network are accessed into a virtual subnet in the VPC, so that high-efficiency packet forwarding communication is ensured. The border router may also be used to open connections between the VPC internal networks.

The production area IT network is composed of a plurality of physically isolated zones on the network, and except that the servers of the database zones are operated in a physical machine mode, the servers of other zones provide a virtual machine bearing system and services by operating VMware virtualization software. Each zone has hardware devices such as an access switch, a convergence switch, a physical firewall, an F5 and the like which are independently deployed.

The test cloud and the quasi-production cloud are deployed in the same cluster, and related advantages and values are as follows:

a) The scheme design gives consideration to the SDN network of the cloud platform and the traditional physical network, ensures network isolation of a host in a zone, and can realize access control as required by configuring a boundary router and a firewall in the cloud.

b) The core switch is produced by simulating the hardware switch, so that the large-scale networking is supported, and the performance is not lost.

c) Compared with the current IT network architecture of the production area, the system architecture can realize the target networking architecture by using a small amount of network equipment, and saves the investment for building the IT system of the quasi-production area.

d) And multi-VPC is used for simulating a multi-production zone environment, so that IaaS cloud resource replication is facilitated.

3. Custom development: host batch cloning

When the virtual machines in the development test cloud are all in the VPC, preconditions are created for IP addresses not to conflict after the virtual machines are copied or cloned. Then, implementing "one-touch deployment" can be split into two aspects:

1) The VPC environment where the fast replication virtual machine is located comprises a VPC router, a virtual subnet, load balancing and the like;

2) The virtual machines are cloned to the newly built VPC in batches quickly while keeping the IP address unchanged.

Regarding point 1, the cloud platform may be implemented by an existing "resource orchestration unit", where the structure of the resource orchestration unit is shown in fig. 2; the cloud platform manager can extract IaaS network resources such as VPC routers, virtual subnets, load balancing and the like used in the original test environment to form a resource arrangement template by utilizing the existing function of a resource arrangement unit of the cloud platform.

Compiling and discharging a set of entity resources (Topology) from the combination relation described by the template, wherein the entity resources comprise IaaS network resources such as VPC routers, virtual subnets, load balancing and the like which are configured in the same way as the original test environment;

regarding point 2, the "host batch clone" scheme by cloud platform.

And selecting virtual machines related to a plurality of service systems in an original development and test environment, executing batch cloning host operation, wherein a target network selects a VPC virtual subnet newly copied through 'resource arrangement', the host after default cloning keeps an IP address unchanged, and a corresponding host firewall policy also keeps unchanged, so that the implementation effect of 'host batch cloning' related background logic is shown in figure 3.

The newly created VPC contains the target test environment through the resource arrangement unit and the host batch clone. Because cloning of the virtual machine storage volume is based on snapshot implementation, the version and data of each service system in the new test environment are almost consistent with those of each service system in the original development test cloud platform (the difference of data in volume after specific cloning depends on the storage volume replication time and how much of source volume IO is during cloning). Compared with the prior art, the method can develop target test work without putting into a plurality of service development and test teams to build an environment, can clone virtual machines to a new environment in batches quickly, meanwhile, a service system does not need to modify configuration, keeps the server IP, the network environment and the server data unchanged, ensures that the deployed service system can be put into use without modifying any configuration, and makes a certain contribution to the development and test work of software.

Claims

1. A cloud platform-based system one-key deployment method is characterized by comprising the following steps of: the method comprises the following steps:

s1, establishing a source VPC network in a cloud platform;

s3, establishing a clone VPC network in the cloud platform;

2. The cloud platform based system one-key deployment method of claim 1, wherein: the step S4 is performed by a resource orchestration unit comprising:

3. The cloud platform based system one-key deployment method of claim 1, wherein: the step S4 includes the steps of:

4. The cloud platform based system one-key deployment method of claim 3, wherein: the step S6 includes the steps of:

s62, performing copy operation on the selected virtual machine;

5. The cloud platform based system one-key deployment method of claim 4, wherein: the step S7 includes the steps of: