Run multiple machines in a single Vagrant file

Last Updated on by

Post summary: How to run multiple machines on Vagrant described in a single Vagrantfile.

Code bellow can be found in GitHub sample-dropwizard-rest-stub repository in Vagrantfile file. This post is part of Vagrant series. All of other Vagrant related posts as well as more theoretical information what is Vagrant and why to use it can be found in What is Vagrant and why to use it post.

Vagrantfile

As described in Vagrant introduction post all configurations are done in a single text file called Vagrantfile. Bellow is a Vagrant file which can be used to initialise two machines. One is same as described in Run Dropwizard Java application on Vagrant post, the other is the one described in Run Docker container on Vagrant post.

Vagrant.configure('2') do |config|

  config.vm.hostname = 'dropwizard'
  config.vm.box = 'opscode-centos-7.2'
  config.vm.box_url = 'http://opscode-vm-bento.s3.amazonaws.com/vagrant/virtualbox/opscode_centos-7.2_chef-provisionerless.box'

  config.vm.synced_folder './', '/vagrant'

  config.vm.define 'jar' do |jar|
    jar.vm.network :forwarded_port, guest: 9000, host: 9100
    jar.vm.network :forwarded_port, guest: 9001, host: 9101

    jar.vm.provider :virtualbox do |vb|
      vb.name = 'dropwizard-rest-stub-jar'
    end

    jar.vm.provision :shell do |shell|
      shell.inline = <<-SHELL
        sudo service dropwizard stop
        sudo yum -y install java
        sudo mkdir -p /var/dropwizard-rest-stub
        sudo mkdir -p /var/dropwizard-rest-stub/logs
        sudo cp /vagrant/target/sample-dropwizard-rest-stub-1.0-SNAPSHOT.jar /var/dropwizard-rest-stub/dropwizard-rest-stub.jar
        sudo cp /vagrant/config-vagrant.yml /var/dropwizard-rest-stub/config.yml
        sudo cp /vagrant/linux_service_file /etc/init.d/dropwizard
        # Replace CR+LF with LF because of Windows
        sudo sed -i -e 's/\r//g' /etc/init.d/dropwizard
        sudo service dropwizard start
      SHELL
    end
  end

  config.vm.define 'docker' do |docker|
    docker.vm.network :forwarded_port, guest: 9000, host: 9000
    docker.vm.network :forwarded_port, guest: 9001, host: 9001

    docker.vm.provider :virtualbox do |vb|
      vb.name = 'dropwizard-rest-stub-docker'
      vb.customize ['modifyvm', :id, '--memory', '768', '--cpus', '2']
    end
  
    docker.vm.provision :shell do |shell|
      shell.inline = <<-SHELL
        sudo yum -y install epel-release
        sudo yum -y install python-pip
        sudo pip install --upgrade pip
        sudo pip install six==1.4
        sudo pip install docker-py
      SHELL
    end
  
    docker.vm.provision :docker do |docker|
      docker.build_image '/vagrant/.', args: '-t dropwizard-rest-stub'
      docker.run 'dropwizard-rest-stub', args: '-it -p 9000:9000 -p 9001:9001 -e ENV_VARIABLE_VERSION=1.1.1'
    end
  end
  
end

Vagrantfile explanation

File starts with a Vagrant.configure(‘2’) do |config| which states that version 2 of Vagrant API will be used and defines constant with name config to be used bellow. Guest operating system hostname is set with config.vm.hostname. If you use vagrant-hostsupdater plugin it will add it to your hosts file and you can access it from browser in case you are developing web applications. With config.vm.box you define which would be the guest operating system. Vagrant maintains config.vm.box = “hashicorp/precise64” which is Ubuntu 12.04 (32 and 64-bit), they also recommend to use Bento’s boxes, but I found issues with Vagrant’s as well as Bento’s boxes so I’ve decided to use one I know is working. I specify where it is located with config.vm.box_url. It is It is CentOS 7.2. With config.vm.synced_folder command you specify that Vagrantfile location folder is shared as /vagrant/ in guest operating system. This makes it easy to transfer files between guest and host operating systems. Now comes the part where two different machines are defined. First one is defined with config.vm.define ‘jar’ do |jar|, which declares variable jar to be used later in configurations. All other configurations are well described in Run Dropwizard Java application on Vagrant post. Specific part here is port mapping. In order to avoid port collision port 9000 from guest is mapped to port 9100 to host with jar.vm.network :forwarded_port, guest: 9000, host: 9100 line. This is because second machine uses port 9000 from the host. Second machine is defined in config.vm.define ‘docker’ do |docker|, which declares variable docker to be used in further configurations. All other configurations are described in Run Docker container on Vagrant post.

Running Vagrant

Command to start Vagrant machine is: vagrant up. Then in order to invoke provisioning section with actual deployment you have to call: vagrant provision. All can be done in one step: vagrant up –provision. To shutdown the machine use vagrant halt. To delete machine: vagrant destroy.

Conclusion

It is very easy to create Vagrantfile that builds and runs several machines with different applications. It possible to make those machine communicate with each other, hence simulation real environment. Once created file can be reused by all team members. It is executed over and over again making provisioning extremely easy.

Read more...

Run Docker container on Vagrant

Last Updated on by

Post summary: How to run Docker container on Vagrant.

Code bellow can be found in GitHub sample-dropwizard-rest-stub repository in Vagrantfile-docker file. Since Vagrant requires to have only one Vagrantfile if you want to run this example you have to rename Vagrantfile-docker to Vagrantfile then run Vagrant commands described in the end of this post. This post is part of Vagrant series. All of other Vagrant related posts as well as more theoretical information what is Vagrant and why to use it can be found in What is Vagrant and why to use it post.

Vagrantfile

As described in Vagrant introduction post all configurations are done in a single text file called Vagrantfile. Bellow is a Vagrant file which can be used to deploy and start Docker container on Vagrant. Example here uses Dockerised application that is described in Run Dropwizard application in Docker with templated configuration using environment variables post.

Vagrant.configure('2') do |config|

  config.vm.hostname = 'dropwizard'
  config.vm.box = 'opscode-centos-7.2'
  config.vm.box_url = 'http://opscode-vm-bento.s3.amazonaws.com/vagrant/virtualbox/opscode_centos-7.2_chef-provisionerless.box'

  config.vm.synced_folder './', '/vagrant'

  config.vm.network :forwarded_port, guest: 9000, host: 9000
  config.vm.network :forwarded_port, guest: 9001, host: 9001

  config.vm.provider :virtualbox do |vb|
    vb.name = 'dropwizard-rest-stub-docker'
    vb.customize ['modifyvm', :id, '--memory', '768', '--cpus', '2']
  end

  config.vm.provision :shell do |shell|
    shell.inline = <<-SHELL
      sudo yum -y install epel-release
      sudo yum -y install python-pip
      sudo pip install --upgrade pip
      sudo pip install six==1.4
      sudo pip install docker-py
    SHELL
  end

  config.vm.provision :docker do |docker|
    docker.build_image '/vagrant/.', args: '-t dropwizard-rest-stub'
    docker.run 'dropwizard-rest-stub', args: '-it -p 9000:9000 -p 9001:9001 -e ENV_VARIABLE_VERSION=1.1.1'
  end

end

Vagrantfile explanation

File starts with a Vagrant.configure(‘2’) do |config| which states that version 2 of Vagrant API will be used and defines constant with name config to be used bellow. Guest operating system hostname is set with config.vm.hostname. If you use vagrant-hostsupdater plugin it will add it to your hosts file and you can access it from browser in case you are developing web applications. With config.vm.box you define which would be the guest operating system. Vagrant maintains config.vm.box = “hashicorp/precise64” which is Ubuntu 12.04 (32 and 64-bit), they also recommend to use Bento’s boxes. I have found issues with Vagrant’s as well as Bento’s boxes so I’ve decided to use one I know is working. I specify where it is located with config.vm.box_url. It is CentOS 7.2. With config.vm.synced_folder command you specify that Vagrantfile location folder is shared as /vagrant/ in guest operating system. This makes it easy to transfer files between guest and host operating systems. This mount is done by default, but it is good to explicitly state it for better readability. With config.vm.network :forwarded_port port from guest OS is forwarded to your hosting OS. Without exposing any port you will not have access to guest OS, only port open by default is 22 for SSH. With config.vm.provider :virtualbox do |vb| you access VirtualBox provider for more configurations, vb.name = ‘dropwizard-rest-stub-docker’ sets the name that you see in Oracle VirtualBox Manager. With vb.customize [‘modifyvm’, :id, ‘–memory’, ‘768’, ‘–cpus’, ‘2’] you modify default hardware settings for the machine, RAM is set to 768MB and 2 CPUs are configured. Finally the provisioning part takes place which is done by shell commands inside config.vm.provision :shell do |shell| block. This block installs Python as well as docker-py. It is CentOS specific as it uses YUM which is CentOS package manager. Next provisioning part is to run docker provisioner that builds docker image and then runs it by mapping ports and setting environment variable. For more details how to build and run Docker containers read Run Dropwizard application in Docker with templated configuration using environment variables post.

Running Vagrant

Command to start Vagrant machine is: vagrant up. Then in order to invoke provisioning section with actual deployment you have to call: vagrant provision. All can be done in one step: vagrant up –provision. To shutdown the machine use vagrant halt. To delete machine: vagrant destroy.

Conclusion

It is very easy to create Vagrantfile that builds and runs Docker container. Once created file can be reused by all team members. It is executed over and over again making provisioning extremely easy.

Read more...

Run Dropwizard application in Docker with templated configuration using environment variables

Last Updated on by

Post summary: How to run Dropwizard application inside a Docker container with configuration template file which later get replaced by environment variables.

In Build a RESTful stub server with Dropwizard post I have described how to build a Dropwizard application and run it. In current post I will show how this application can be inserted into Docker container and run with different configurations, based on different environment variables. Code sample here can be found as buildable and runable project in GitHub sample-dropwizard-rest-stub repository.

Dropwizard templated config

Dropwizard supports replacing variables from config file with environment variables if such are defined. First step is to make the config file with special notation.

version: ${ENV_VARIABLE_VERSION:- 0.0.2}

Dropwizard substitution is based on Apache’s StrSubstitutor library. ${ENV_VARIABLE_VERSION:- 0.0.2} means that Dropwizard will search for environment variable with name ENV_VARIABLE_VERSION and replace its value with given variable. If no variable is configured then it will replace with default value of 0.0.2. If default is not needed then just ${ENV_VARIABLE_VERSION} can be used.

Next step is to make Dropwizard substitute config file with environment variables on its startup before file is being read. This is done with following code:

@Override
public void initialize(Bootstrap<RestStubConfig> bootstrap) {
	bootstrap.setConfigurationSourceProvider(new SubstitutingSourceProvider(
			bootstrap.getConfigurationSourceProvider(), 
			new EnvironmentVariableSubstitutor(false)));
}

EnvironmentVariableSubstitutor is used with false in order to suppress throwing of UndefinedEnvironmentVariableException in case environment variable is not defined. There is approach to pack config.yml file into JAR file and use it from there, then bootstrap.getConfigurationSourceProvider() should be changed with path -> Thread.currentThread().getContextClassLoader().getResourceAsStream(path).

Docker

Docker is platform for building software inside containers which then can be deployed on different environments. A container is like a virtual machine with significant difference that it does not build full operation system. In this way host’s resources are optimised, container consumes as much memory as needed by application. Virtual machine itself consumes memory to run the operation system as well. Containers have resource (CPU, RAM, HDD) isolation so that application sees the container as separate operating system.

Dockerfile

A Dockerfile is a text document that contains all the commands a user could call on the command line to assemble an image. Using docker build users can create an automated build that executes several command-line instructions in succession. In given example Dockerfile looks like this:

FROM openjdk:8u121-jre-alpine
MAINTAINER Automation Rhapsody https://automationrhapsody.com/

WORKDIR /var/dropwizard-rest-stub

ADD target/sample-dropwizard-rest-stub-1.0-SNAPSHOT.jar /var/dropwizard-rest-stub/dropwizard-rest-stub.jar
ADD config-docker.yml /var/dropwizard-rest-stub/config.yml

EXPOSE 9000 9001

ENTRYPOINT ["java", "-jar", "dropwizard-rest-stub.jar", "server", "config.yml"]

Dockerfile starts with FROM openjdk:8u121-jre-alpine which instructs docker to use this already prepared image as a base image for container. At https://hub.docker.com/ there are numerous of images maintained by different organisations or individuals that provide different frameworks and tools. They can save you a lot of work allowing you to skip configuration and installation of general things. In given case openjdk:8u121-jre-alpine has OpenJDK JRE 1.8.121 already installed with minimalist Linux libraries. MAINTAINER documents who is the author of this file. WORKDIR sets working directory of the container. Then ADD copies JAR file and config.yml to container. ENTRYPOINT configures container to be run as executable. This instruction translates to java -jar dropwizard-rest-stub.jar server config.yml command. More info can be found in Dockerfile reference link.

Build Docker container

Docker container can be build with following command:

docker build -t dropwizard-rest-stub .

dropwizard-rest-stub is the name of container, it is later used to run container with this name. Note that dot in the end is mandatory.

Run Docker container

docker run -it -p 9000:9000 -p 9001:9001 -e ENV_VARIABLE_VERSION=1.1.1 dropwizard-rest-stub

Name of the container is dropwizard-rest-stub and is put in the end. With -p 9000:9000 port 9000 form guest machine is exposed to host machine, otherwise container will not be accessible. With -e ENV_VARIABLE_VERSION=1.1.1 environment variable is being set. If not passed it will be substituted with 0.0.2 in config.yml file as described above. In case of many environment variables it is more comfortable to put them in a file with content KEY=VALUE. Then use this file with –env-file= instead of specifying each individual variable.

Conclusion

Dropwizard has very easy mechanism for making configuration dependant on environment variables. This makes Dropwizard applications extremely suitable to be build into Docker containers and run on different environments just by changing environment variables being passed to container.

Read more...