Build open source automation integration platform

Previous post: Macaca – Node.js-based automated testing

This is the official REPO of Macacaalibaba/macacaWelcome to issue, star.

Reliable introduction

Reliable is a distributed architecture continuous integration system developed by members of the Macaca team. Applicable to integration construction and integration testing scenarios. It is a typical master-slave structure, which is divided into two parts: reliable-master and reliable-slave.

It has the following characteristics:

  • Cluster load is properly allocated
  • Provide plug-in mechanism, easy to expand
  • Deployment is very simple
  • Easy access to Gitlab, Github and other community systems

Meanwhile, she blends seamlessly with Macaca.

It is a good choice to use Macaca + Reliable to build an open source automation integration platform for your team and company.

Master one-click deployment

For a more convenient installation experience, reliable-Master is deployed using Docker. Clone the source code to the server or local machine and make deploy env=prod.

For details about the configuration, see github/ control-master /deploy

Configuration and Functions

email

Support simple mail transfer protocol, through the configuration of SMTP notification creators, subscribers of the mail push function.

multilingual

Supports multiple languages. The default is English. You can set the site to Chinese through the configuration file.

Add user

Run the following command to add an initial user or administrator for the site:

$ make adduser
Copy the code

Next you can access the site using port 8080

Log in to the created user

You’re done

Slave deployment

We use ZeromQ for message queuing, so we need to install ZeromQ first.

$ brew install pkg-config
$ brew install zeromq

# Then

$ npm install reliable-slave -g
Copy the code

To complete the deployment and mount, specify the Master IP address or domain name:

reliable server -m  --verbose
Copy the code

As shown in the document

A variety of patterns

  • Single point pattern
    • Both control-slave and control-Master are deployed on the same machine
    • This mode saves equipment and is the least expensive to use
  • Cluster pattern
    • Reliable slave is recommended to be deployed on Windows or OSX (MacMini is recommended), and reliable Master is recommended to be deployed on a professional server
  • The crowdsourcing model
    • Users use lapid-slave to mount the deployed lapID-master to provide resources

Welcome to discuss and learn from each other.

Weibo: weibo.com/xudafeng Github: github.com/xudafeng