当前位置: 首页 > 软件库 > 云计算 > >

gradle-docker-compose-plugin

授权协议 MIT License
开发语言 Java
所属分类 云计算
软件类型 开源软件
地区 不详
投 递 者 吴建中
操作系统 跨平台
开源组织
适用人群 未知
 软件概览

gradle-docker-compose-plugin Build

Simplifies usage of Docker Compose for local development and integration testing in Gradle environment.

composeUp task starts the application and waits till all containers become healthy and all exposed TCP ports are open (so till the application is ready). It reads assigned host and ports of particular containers and stores them into dockerCompose.servicesInfos property.

composeDown task stops the application and removes the containers, only if 'stopContainers' is set to 'true' (default value).

composeDownForced task stops the application and removes the containers.

composePull task pulls and optionally builds the images required by the application. This is useful, for example, with a CI platform that caches docker images to decrease build times.

composeBuild task builds the services of the application.

composePush task pushes images for services to their respective registry/repository.

composeLogs task stores logs from all containers to files in containerLogToDir directory.

Quick start

buildscript {
    repositories {
        mavenCentral()
    }
    dependencies {
        classpath "com.avast.gradle:gradle-docker-compose-plugin:$versionHere"
    }
}

apply plugin: 'docker-compose'

// Or use the new Gradle Portal plugins (then you don't have to add the dependency as above):
// plugins {
//  id 'com.avast.gradle.docker-compose' version "$versionHere"
// }

dockerCompose.isRequiredBy(test)
  • docker-compose up is executed in the project directory, so it uses the docker-compose.yml file.
  • If the provided task (test in the example above) executes a new process then environment variables and Java system properties are provided.
    • The name of environment variable is ${serviceName}_HOST and ${serviceName}_TCP_${exposedPort} (e.g. WEB_HOST and WEB_TCP_80).
    • The name of Java system property is ${serviceName}.host and ${serviceName}.tcp.${exposedPort} (e.g. web.host and web.tcp.80).
    • If the service is scaled then the serviceName has _1, _2... suffix (e.g. WEB_1_HOST and WEB_1_TCP_80, web_1.host and web_1.tcp.80)

Why to use Docker Compose?

  1. I want to be able to run my application on my computer, and it must work for my colleagues as well. Just execute docker-compose up and I'm done - e.g. the database is running.
  2. I want to be able to test my application on my computer - I don't wanna wait till my application is deployed into dev/testing environment and acceptance/end2end tests get executed. I want to execute these tests on my computer - it means execute docker-compose up before these tests.

Why this plugin?

You could easily ensure that docker-compose up is called before your tests but there are few gotchas that this plugin solves:

  1. If you execute docker-compose up -d (detached) then this command returns immediately and your application is probably not able to serve requests at this time. This plugin waits till all containers become healthy and all exported TCP ports of all services are open.
    • If waiting for healthy state or open TCP ports timeouts (default is 15 minutes) then it prints log of related service.
  2. It's recommended not to assign fixed values of exposed ports in docker-compose.yml (i.e. 8888:80) because it can cause ports collision on integration servers. If you don't assign a fixed value for exposed port (use just 80) then the port is exposed as a random free port. This plugin reads assigned ports (and even IP addresses of containers) and stores them into dockerCompose.servicesInfo map.
  3. There are minor differences when using Linux containers on Linux, Windows and Mac, and when using Windows Containers. This plugin handles these differences for you so you have the same experience in all environments.

Usage

The plugin must be applied on project that contains docker-compose.yml file. It supposes that Docker Engine and Docker Compose are installed and available in PATH.

Starting from plugin version 0.10.0, Gradle 4.9 or newer is required (because it uses Task Configuration Avoidance API).

buildscript {
    repositories {
        mavenCentral()
    }
    dependencies {
        classpath "com.avast.gradle:gradle-docker-compose-plugin:$versionHere"
    }
}

apply plugin: 'docker-compose'

dockerCompose.isRequiredBy(test) // hooks 'dependsOn composeUp' and 'finalizedBy composeDown', and exposes environment variables and system properties (if possible)

dockerCompose {
    useComposeFiles = ['docker-compose.yml', 'docker-compose.prod.yml'] // like 'docker-compose -f <file>'; default is empty
    startedServices = ['web'] // list of services to execute when calling 'docker-compose up' or 'docker-compose pull' (when not specified, all services are executed)
    scale = [${serviceName1}: 5, ${serviceName2}: 2] // Pass docker compose --scale option like 'docker-compose up --scale serviceName1=5 --scale serviceName2=2'
    forceRecreate = false // pass '--force-recreate' and '--renew-anon-volumes' when calling 'docker-compose up' when set to 'true`
    noRecreate = false // pass '--no-recreate' when calling 'docker-compose up' when set to 'true`
    buildBeforeUp = true // performs 'docker-compose build' before calling the 'up' command; default is true
    buildBeforePull = true // performs 'docker-compose build' before calling the 'pull' command; default is true
    ignorePullFailure = false // when set to true, pass '--ignore-pull-failure' to 'docker-compose pull'
    ignorePushFailure = false // when set to true, pass '--ignore-push-failure' to 'docker-compose push'
    pushServices = [] // which services should be pushed, if not defined then upon `composePush` task all defined services in compose file will be pushed (default behaviour)
    buildAdditionalArgs = ['--force-rm']
    pullAdditionalArgs = ['--ignore-pull-failures']
    upAdditionalArgs = ['--no-deps']
    downAdditionalArgs = ['--some-switch']
    composeAdditionalArgs = ['--context', 'remote', '--verbose', "--log-level", "DEBUG"] // for adding more [options] in docker-compose [-f <arg>...] [options] [COMMAND] [ARGS...]

    waitForTcpPorts = true // turns on/off the waiting for exposed TCP ports opening; default is true
    waitForTcpPortsTimeout = java.time.Duration.ofMinutes(15) // how long to wait until all exposed TCP become open; default is 15 minutes
    waitAfterTcpProbeFailure = java.time.Duration.ofSeconds(1) // how long to sleep before next attempt to check if a TCP is open; default is 1 second
    tcpPortsToIgnoreWhenWaiting = [1234] // list of TCP ports what will be ignored when waiting for exposed TCP ports opening; default: empty list
    waitForHealthyStateTimeout = java.time.Duration.ofMinutes(15) // how long to wait until a container becomes healthy; default is 15 minutes
    waitAfterHealthyStateProbeFailure = java.time.Duration.ofSeconds(5) // how long to sleep before next attempt to check healthy status; default is 5 seconds
    checkContainersRunning = true // turns on/off checking if container is running or restarting (during waiting for open TCP port and healthy state); default is true

    captureContainersOutput = false // if true, prints output of all containers to Gradle output - very useful for debugging; default is false
    captureContainersOutputToFile = project.file('/path/to/logFile') // sends output of all containers to a log file
    captureContainersOutputToFiles = project.file('/path/to/directory') // sends output of all services to a dedicated log file in the directory specified, e.g. 'web.log' for service named 'log'
    composeLogToFile = project.file('build/my-logs.txt') // redirect output of composeUp and composeDown tasks to this file; default is null (ouput is not redirected)
    containerLogToDir = project.file('build/logs') // directory where composeLogs task stores output of the containers; default: build/containers-logs
    includeDependencies = false // calculates services dependencies of startedServices and includes those when gathering logs or removing containers; default is false

    stopContainers = true // doesn't call `docker-compose down` if set to false - see below the paragraph about reconnecting; default is true
    removeContainers = true // default is true
    retainContainersOnStartupFailure = false // if set to true, skips running ComposeDownForced task when ComposeUp fails - useful for troubleshooting; default is false
    removeImages = com.avast.gradle.dockercompose.RemoveImages.None // Other accepted values are All and Local
    removeVolumes = true // default is true
    removeOrphans = false // removes containers for services not defined in the Compose file; default is false
    
    projectName = 'my-project' // allow to set custom docker-compose project name (defaults to a stable name derived from absolute path of the project and nested settings name), set to null to Docker Compose default (directory name)
    projectNamePrefix = 'my_prefix_' // allow to set custom prefix of docker-compose project name, the final project name has nested configuration name appended
    executable = '/path/to/docker-compose' // allow to set the path of the docker-compose executable (useful if not present in PATH)
    dockerExecutable = '/path/to/docker' // allow to set the path of the docker executable (useful if not present in PATH)
    dockerComposeWorkingDirectory = project.file('/path/where/docker-compose/is/invoked/from')
    dockerComposeStopTimeout = java.time.Duration.ofSeconds(20) // time before docker-compose sends SIGTERM to the running containers after the composeDown task has been started
    environment.put 'BACKEND_ADDRESS', '192.168.1.100' // environment variables to be used when calling 'docker-compose', e.g. for substitution in compose file
}

test.doFirst {
    // exposes "${serviceName}_HOST" and "${serviceName}_TCP_${exposedPort}" environment variables
    // for example exposes "WEB_HOST" and "WEB_TCP_80" environment variables for service named `web` with exposed port `80`
    // if service is scaled using scale option, environment variables will be exposed for each service instance like "WEB_1_HOST", "WEB_1_TCP_80", "WEB_2_HOST", "WEB_2_TCP_80" and so on
    dockerCompose.exposeAsEnvironment(test)
    // exposes "${serviceName}.host" and "${serviceName}.tcp.${exposedPort}" system properties
    // for example exposes "web.host" and "web.tcp.80" system properties for service named `web` with exposed port `80`
    // if service is scaled using scale option, environment variables will be exposed for each service instance like "web_1.host", "web_1.tcp.80", "web_2.host", "web_2.tcp.80" and so on
    dockerCompose.exposeAsSystemProperties(test)
    // get information about container of service `web` (declared in docker-compose.yml)
    def webInfo = dockerCompose.servicesInfos.web.firstContainer
    // in case scale option is used, dockerCompose.servicesInfos.containerInfos will contain information about all running containers of service. Particular container can be retrieved either by iterating the values of containerInfos map (key is service instance name, for example 'web_1')
    def webInfo = dockerCompose.servicesInfos.web.'web_1'
    // pass host and exposed TCP port 80 as custom-named Java System properties
    systemProperty 'myweb.host', webInfo.host
    systemProperty 'myweb.port', webInfo.ports[80]
    // it's possible to read information about exposed UDP ports using webInfo.updPorts[1234]
}

Nested configurations

It is possible to create a new set of ComposeUp/ComposeBuild/ComposePull/ComposeDown/ComposeDownForced/ComposePush tasks using following syntax:

Groovy
dockerCompose {
    // settings as usual
    myNested {
        useComposeFiles = ['docker-compose-for-integration-tests.yml']
        isRequiredBy(project.tasks.myTask)
    }
}
  • It creates myNestedComposeUp, myNestedComposeBuild, myNestedComposePull, myNestedComposeDown, myNestedComposeDownForced and myNestedComposePush tasks.
  • It's possible to use all the settings as in the main dockerCompose block.
  • Configuration of the nested settings defaults to the main dockerCompose settings (declared before the nested settings), except following properties: projectName, startedServices, useComposeFiles, scale, captureContainersOutputToFile, captureContainersOutputToFiles, composeLogToFile, containerLogToDir, pushServices

When exposing service info from myNestedComposeUp task into your task you should use following syntax:

test.doFirst {
    dockerCompose.myNested.exposeAsEnvironment(test)
}
Kotlin
test.doFirst {
    dockerCompose.nested("myNested").exposeAsEnvironment(project.tasks.named("test").get())
}

It's also possible to use this simplified syntax:

dockerCompose {
    isRequiredByMyTask 'docker-compose-for-integration-tests.yml'
}

Reconnecting

If you specify stopContainers to be false then the plugin automatically tries to reconnect to the containers from the previous runinstead of calling docker-compose up again. Thanks to this, the startup can be very fast.

It's very handy in scenarios when you iterate quickly and e.g. don't want to wait for Postgres to start again and again.

Because you don't want to check-in this change to your VCS, you can take advantage of this init.gradle initialization script (in short, copy this file to your USER_HOME/.gradle/ directory).

Tips

  • You can call dockerCompose.isRequiredBy(anyTask) for any task, for example for your custom integrationTest task.
  • If some Dockerfile needs an artifact generated by Gradle then you can declare this dependency in a standard way, like composeUp.dependsOn project(':my-app').distTar
  • All properties in dockerCompose have meaningful default values so you don't have to touch it. If you are interested then you can look at ComposeSettings.groovy for reference.
  • dockerCompose.servicesInfos contains information about running containers so you must access this property after composeUp task is finished. So doFirst of your test task is perfect place where to access it.
  • Plugin honours a docker-compose.override.yml file, but only when no files are specified with useComposeFiles (conform command-line behavior).
  • Check ContainerInfo.groovy to see what you can know about running containers.
  • You can determine the Docker host in your Gradle build (i.e. docker-machine start) and set the DOCKER_HOST environment variable for compose to use: dockerCompose { environment.put 'DOCKER_HOST', '192.168.64.9' }
  • If the services executed by docker-compose are running on a specific host (different than Docker, like in CirceCI 2.0), then SERVICES_HOST environment variable can be used. This value will be used as the hostname where the services are expected to be listening.
  • If you need to troubleshoot a failing ComposeUp task, set retainContainersOnStartupFailure to prevent containers from begin forcibly deleted. Does not override removeContainers, so if you run ComposeDown, it will not be affected.
  • 在java web项目中,本地开发经常会需要在本地使用docker启动数据库等之类的服务。gradle提供了一个插件,允许通过gradle task启动docker的容器。在这里我们介绍的一个gralde插件com.avast.gradle.docker-compose 简单使用 引入插件 plugins { id "com.avast.gradle.docker-compose" ver

  • docker-maven-plugin是spotify出品的一款针对spring boot项目的docker插件,可将spring boot项目打包到docker镜像中。 如果在编译docker镜像时需要设置build arg,只需要在maven的配置文件pom.xml中,configuration下增加buildArgs。标签的key和值对应build arg的key和值,如下所示,docker

  • 简介 docker compose是用于定义和运行多容器docker应用程序的工具,compose 通过一个配置文件来管理多个docker容器。可以使用docker-compose.yml脚本来启动、停止、重启应用,进行docker容器的编排和管理。但是docker compose并没有实现容器的负载均衡,还需要借助其他工具实现。 安装 docker官网地址:https://docs.docker

  • Docker群:在堆栈部署期间忽略docker中的“构建”配置文件 | 码农家园 Docker swarm: 'build' configuration in docker compose file ignored during stack deployment

 相关资料
  • 主要内容:实例,Compose 安装,使用,composetest/app.py 文件代码,docker-compose.yml 配置文件,yml 配置指令参考Compose 简介 Compose 是用于定义和运行多容器 Docker 应用程序的工具。通过 Compose,您可以使用 YML 文件来配置应用程序需要的所有服务。然后,使用一个命令,就可以从 YML 文件配置中创建并启动所有服务。 如果你还不了解 YML 文件配置,可以先阅读 YAML 入门教程。 Compose 使用的三个步骤:

  • 使用 Docker Compose,您可以使用一个命令启动本地测试网络。 需求 安装 tendermint 安装 docker 安装 docker-compose 构建 构建 tendermint 二进制文件和可选 tendermint/localnode docker 映像。 注意,二进制文件将被挂载到容器中,因此可以在不重新构建镜像的情况下更新它。 cd $GOPATH/src/github.

  • Docker Compose 是 Docker 官方编排(Orchestration)项目之一,负责快速的部署分布式应用。 本章将介绍 Compose 项目情况以及安装和使用。

  • 在安装 Compose之前,你需要先安装好 Docker 。然后你需要使用 curl 指令来安装 Compose 安装 Docker 首先,你需要安装大于或者等于1compose/.6版本的 Docker 。 MAC OSX 安装指南 Ubuntu 安装指南 其它系统安装指南 安装 Compose 运行下边的命令来安装 Compose: curl -L https://githubcompose/

  • 解决多容器的 APP 部署问题。 要从 Dockerfile build image 或从 Dockerhub 拉取或从 Tar export iamge。 要创建多个 container。 要管理多个 container。(启动停止删除) 介绍Docker Compose 通过一个 yml 文件定义多容器的 docker 应用,通过一条命令就可以根据 yml 文件的定义去创建或管理这些容器。 版

  • Docker Compose 是 Docker 编排服务的一部分,Machine 可以让用户在其它平台快速安装Docker,Swarm 可以让 Docker 容器在集群中高效运转,而 Compose 可以让用户在集群中部署分布式应用。简单的说,Docker Compose 属于一个“应用层”的服务,用户可以定义哪个容器组运行哪个应用,它支持动态改变应用,并在需要时扩展。 使用Compose的第一步