Vagrant

Type: vagrant Artifact BuilderId: mitchellh.post-processor.vagrant

The Packer Vagrant post-processor takes a build and converts the artifact into a valid Vagrant box, if it can. This lets you use Packer to automatically create arbitrarily complex Vagrant boxes, and is in fact how the official boxes distributed by Vagrant are created.

If you've never used a post-processor before, please read the documentation on using post-processors. This knowledge will be expected for the remainder of this document.

Because Vagrant boxes are provider-specific, the Vagrant post-processor is hardcoded to understand how to convert the artifacts of certain builders into proper boxes for their respective providers.

Currently, the Vagrant post-processor can create boxes for the following providers.

• AWS
• Azure
• DigitalOcean
• Docker
• File
• Hyper-V
• LXC
• Parallels
• QEMU
• VirtualBox
• VMware

Please note that if you are using the Vagrant builder, then the Vagrant post-processor is unnecessary because the output of the Vagrant builder is already a Vagrant box; using this post-processor with the Vagrant builder will cause your build to fail.

Configuration

The simplest way to use the post-processor is to just enable it. No configuration is required by default. This will mostly do what you expect and will build functioning boxes for many of the built-in builders of Packer.

However, if you want to configure things a bit more, the post-processor does expose some configuration options. The available options are listed below, with more details about certain options in following sections.

• architecture (string) - The architecture of the Vagrant box. This will be set to the detected architecture of the builder host by default. Supported values: amd64, i386, arm, arm64, ppc64le, ppc64, mips64le, mips64, mipsle, mips, and s390x.

• compression_level (number) - An integer representing the compression level to use when creating the Vagrant box. Valid values range from 0 to 9, with 0 being no compression and 9 being the best compression. By default, compression is enabled at level 6.

• include (array of strings) - Paths to files to include in the Vagrant box. These files will each be copied into the top level directory of the Vagrant box (regardless of their paths). They can then be used from the Vagrantfile.

• keep_input_artifact (boolean) - When true, preserve the artifact we use to create the vagrant box. Defaults to false, except when you set a cloud provider (e.g. aws, azure, google, digitalocean). In these cases deleting the input artifact would render the vagrant box useless, so we always keep these artifacts -- even if you specifically set "keep_input_artifact":false

• output (string) - The full path to the box file that will be created by this post-processor. This is a template engine. Therefore, you may use user variables and template functions in this field. The following extra variables are also available in this engine:

  • Architecture: The architecture of the Vagrant box
  • Provider: The Vagrant provider the box is for
  • ArtifactId: The ID of the input artifact.
  • BuildName: The name of the build.

  By default, the value of this config is packer_{{.BuildName}}_{{.Provider}}_{{.Architecture}}.box.

• provider_override (string) - this option will override the internal logic that decides which Vagrant provider to set for a particular Packer builder's or post-processor's artifact. It is required when the artifact comes from the Artifice post-processor, but is otherwise optional. Valid options are: digitalocean, virtualbox, azure, vmware, libvirt, docker, lxc, scaleway, hyperv, parallels, aws, or google.

• vagrantfile_template (string) - Path to a template to use for the Vagrantfile that is packaged with the box. This option supports the usage of the template engine for JSON and the contextual variables for HCL2.

• vagrantfile_template_generated (boolean) - By default, Packer will exit with an error if the file specified using the vagrantfile_template variable is not found. However, under certain circumstances, it may be desirable to dynamically generate the Vagrantfile during the course of the build. Setting this variable to true skips the start up check and allows the user to script the creation of the Vagrantfile at some previous point in the build. Defaults to false.

Using together with the Artifice post-processor

Sometimes you may want to run several builds in a pipeline rather than running this post-processor inside a long-running Packer build. Here is an example of how to do this:

JSON

{  "builders": [    {      "type": "null",      "communicator": "none"    }  ],  "post-processors": [    [      {        "type": "artifice",        "files": [          "output-virtualbox-iso/vbox-example-disk001.vmdk",          "output-virtualbox-iso/vbox-example.ovf"        ]      },      {        "type": "vagrant",        "keep_input_artifact": true,        "provider_override": "virtualbox"      }    ]  ]}

HCL2

source "null" "example" {    communicator = "none"} build {  sources = [    "source.null.example"  ]  post-processors {    post-processor "artifice" {      files = [        "output-virtualbox-iso/vbox-example-disk001.vmdk",        "output-virtualbox-iso/vbox-example.ovf"      ]    }    post-processor "vagrant" {      keep_input_artifact = true      provider_override   = "virtualbox"    }  }}

Provider-Specific Overrides

If you have a Packer template with multiple builder types within it, you may want to configure the box creation for each type a little differently. For example, the contents of the Vagrantfile for a Vagrant box for AWS might be different from the contents of the Vagrantfile you want for VMware. The post-processor lets you do this.

Specify overrides within the override configuration by provider name:

JSON

{  "type": "vagrant",  "compression_level": 1,  "override": {    "vmware": {      "compression_level": 0    }  }}

HCL2

## This feature is not implemented in HCL.

In the example above, the compression level will be set to 1 except for VMware, where it will be set to 0.

The available provider names are:

• aws
• azure
• digitalocean
• file
• google
• hyperv
• parallels
• libvirt
• lxc
• scaleway
• virtualbox
• vmware
• docker

Input Artifacts

By default, Packer will delete the original input artifact, assuming you only want the final Vagrant box as the result. If you wish to keep the input artifact (the raw virtual machine, for example), then you must configure Packer to keep it.

Please see the documentation on input artifacts for more information.

Docker

Using a Docker input artifact will include a reference to the image in the Vagrantfile. If the image tag is not specified in the post-processor, the sha256 hash will be used.

The following Docker input artifacts are supported:

• docker builder with commit: true, always uses the sha256 hash
• docker-import
• docker-tag
• docker-push

QEMU/libvirt

The libvirt provider supports QEMU artifacts built using any these accelerators: none, kvm, tcg, or hvf.

VMWare

If you are using the Vagrant post-processor with the vmware-esxi builder, you must export the builder artifact locally; the Vagrant post-processor will not work on remote artifacts.

Artifice

If you are using this post-processor after defining an artifact using the Artifice post-processor, then you must set the "provider_override" template option so that the Vagrant post-processor knows what provider to use to create the Vagrant box.