Shulou(Shulou.com)05/31 Report--

In this issue, the editor will bring you about Rainbond v3.7.0 how to achieve the stability of enterprise PaaS. The article is rich in content and analyzes and narrates it from a professional point of view. I hope you can get something after reading this article.

As an IT basic system platform, Rainbond evaluates and ensures the availability of distributed systems from three aspects: low-coupling architecture design, highly available deployment mode, self-recovery and fault-tolerant design, in order to achieve the effect of unattended.

In the aspect of low coupling architecture design, Rainbond abstracts the distributed system into three types of nodes: management, computing and storage. Different node attributes are composed of different service components to uncouple the services. At the same time, the minimum availability requirements are not the same for different nodes.

Management node: user-oriented, providing a series of management functions such as application construction, control, scheduling, delivery and data storage. If there is no management node in abnormal cases, existing applications can still run normally.

Computing node: responsible for the actual operation of the application and provide environmental protection for the application operation, can tolerate degradation in abnormal cases, and schedule the application to other computing nodes to run.

Storage node: used to store application persistent data and provide data access services. If the storage node is abnormal, the stateless application can still run normally.

To better ensure a highly available deployment, all modules and components of Rainbond itself support high availability--

Management nodes: support idempotent deployment of multiple nodes to ensure high availability

Compute nodes: idempotently deploy multiple compute nodes to form redundant resource pools, thus tolerating single-node resource limitations or failures

Storage node: a storage resource pool is formed by redundant deployment to provide stable services to computing nodes. If any storage node fails, the business will not be interrupted and the data will not be lost.

In the aspect of self-recovery and fault-tolerant service design, in order to eliminate the manual work of users when maintaining large cluster components and services, Rainbond also divides node attributes into management nodes, computing nodes and storage nodes, and makes a strict distinction to make different service lists run under different node attributes. All services are guarded by Systemd. While ensuring the health of the process, each node Node component checks and evaluates all the services, hardware, and network communications of the current node in real time, in order to check out the health status of each service and node, and finally converge into the overall health state of the data center. Rainbond users do not need to pay attention to every component in most cases. You only need to observe the health of the data center or node.

When an exception occurs in the service, different types of exceptions trigger different actions, such as "restart", "alarm" and so on, to restore the availability of the service.

In addition, the independent node inspection and management mechanism provided by Rainbond provides a powerful supplement to the automatic maintenance of computing nodes by Kubernetes. For example, for a large cluster, when a node has errors such as network allocation and image acquisition that lead to unhealthy nodes, the platform automatically disconnects the node, forbids new applications to be dispatched to this node, and minimizes application scheduling errors.

The above is the three aspects of thinking and design of Rainbond to evaluate and ensure the usability of the system. The specific embodiment in the functional point is as follows:

All platform services use Systemd process-level daemon plus Rainbond-Node business-level health check daemon to ensure the healthy operation of the service.

All platform services support health check and Prometheus monitoring indicator exposure

The management node supports online and offline to isolate the platform from being unavailable due to node failure

Automatic isolation and recovery are supported when computing node health check abnormalities

Support to configure custom alarm rules for physical monitoring of nodes and alarms for service monitoring

The statistics of tenants' usage of resources (memory, disk) are done by a single node (automatic switching in case of Rainbond-Worker Master node failure)

Support for convenient query of data center health status and all node health status through command line tools.

In terms of stability, Rainbond will increase the practice of chaos engineering in the next step.

Other update application management functions

This update further improves the application management and delivery system.

First of all, Rainbond provides platform support for svn users and .net core users, with a built-in compilation and running environment for version 2.1 of. Net Core. Users only need to provide standard code to go to the cloud with one click (users of other versions of .Net Core can use a defined Dockerfile).

Secondly, Rainond further provides support for directly interfacing Jenkins CI\ CD processes with API. It is recommended that users deploy Jenkins directly within the platform and interface with the platform through API.

In addition, for the standardized delivery of micro-service architecture products, Rainbond supports the unified sharing of applications and plug-ins into an integrated business system, and quick delivery through the application market.

The specific update list is as follows--

Supports. NetCore (2. 1) language to build applications with one click and run on Linux system

Support continuous construction of SVN code repositories

Added automatic build entry to support automatic build triggered by custom API, Gitee-Webhook and Gogs-Webhook for better integration with third-party CI systems.

Support the application + plug-in complete delivery of the application market, and install the application + plug-in complete business system from the market, providing a complete software delivery model of business + governance function extension binding.

Dockerfile build supports ARG parameters

Support code Tag construction application based on Git warehouse

Support to re-identify language types after application creation

Security.

Data center egress API uses TLS two-way security authentication with console, command line tools and other clients.

The administrator can control the user registration function, and the user joining the team needs to be reviewed by the administrator.

System installation

By simplifying installation steps and separating resource acquisition from node installation, Rainbond will uniformly support online installation and offline installation after this update, solving the problem that some users are unable to install Rainbond due to network environment.

Fully offline installation of Centos7.4.1709, Ubuntu16.04 and Debian9.5 is supported.

Support horizontal expansion of management nodes

About Rainbond

Rainbond is an application-centric open source PaaS, independently developed by Haoyu based on Docker, Kubernetes and other container technologies. It can be used as an application delivery platform, DevOps platform, automated operation and maintenance platform and industry cloud platform in public or private cloud environments, or as an enterprise-level hybrid cloud management tool, Kubernetes container management tool or Service Mesh micro services architecture governance tool.

The above is the editor for you to share Rainbond v3.7.0 how to achieve enterprise-level PaaS stability, if you happen to have similar doubts, you might as well refer to the above analysis to understand. If you want to know more about it, you are welcome to follow the industry information channel.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.