HashiTalks 2025 Learn about unique use cases, homelab setups, and best practices at scale at our 24-hour virtual knowledge sharing event. Register
Packer
Packer Home

Local Shell Post Processor

Official

Type: shell-local

The local shell post processor executes scripts locally during the post processing stage. Shell local provides a convenient way to automate executing some task with packer outputs and variables.

Basic example

The example below is a fully functional self-contained build.

source "file" "example" {
    content = "example content"
}

build {
  source "source.file.example" {
    target = "./test_artifact.txt"
  }

  post-processor "shell-local" {
    inline = ["echo foo"]
  }
}

Configuration Reference

The reference of available configuration options is listed below. The only required element is either "inline" or "script". Every other option is optional.

Exactly one of the following is required:

  • command (string) - This is a single command to execute. It will be written to a temporary file and run using the execute_command call below.

  • inline (array of strings) - This is an array of commands to execute. The commands are concatenated by newlines and turned into a single file, so they are all executed within the same context. This allows you to change directories in one command and use something in the directory in the next and so on. Inline scripts are the easiest way to pull off simple tasks within the machine.

  • script (string) - The path to a script to execute. This path can be absolute or relative. If it is relative, it is relative to the working directory when Packer is executed.

  • scripts (array of strings) - An array of scripts to execute. The scripts will be executed in the order specified. Each script is executed in isolation, so state such as variables from one script won't carry on to the next.

Optional parameters:

  • env (map of strings) - A map of key/value pairs to inject prior to the execute_command. Packer injects some environmental variables by default into the environment, as well, which are covered in the section below. Duplicate env settings override environment_vars settings.

  • environment_vars (array of strings) - An array of key/value pairs to inject prior to the execute_command. The format should be key=value. Packer injects some environmental variables by default into the environment, as well, which are covered in the section below.

  • env_var_format (string) - When we parse the environment_vars that you provide, this gives us a string template to use in order to make sure that we are setting the environment vars correctly. By default on Windows hosts this format is set %s=%s && and on Unix, it is %s='%s'. You probably won't need to change this format, but you can see usage examples for where it is necessary below.

  • execute_command (array of strings) - The command used to execute the script. By default, on *nix systems this is:

    ["/bin/sh", "-c", "{{.Vars}} {{.Script}}"]

    While on Windows, execute_command defaults to:

    ["cmd", "/V", "/C", "{{.Vars}}", "call", "{{.Script}}"]

    This is treated as a template engine. There are several available variables: Script, which is the path to the script to run, and Vars, which is the list of environment_vars, if configured. In addition, you may access any of the variables stored in the generated data using the build template function. If you choose to set this option, make sure that the first element in the array is the shell program you want to use (for example, "sh" or "/usr/local/bin/zsh" or even "powershell.exe" although anything other than a flavor of the shell command language is not explicitly supported and may be broken by assumptions made within Packer). It's worth noting that if you choose to try to use shell-local for Powershell or other Windows commands, the environment variables will not be set properly for your environment.

    For backwards compatibility, execute_command will accept a string instead of an array of strings. If a single string or an array of strings with only one element is provided, Packer will replicate past behavior by appending your execute_command to the array of strings ["sh", "-c"]. For example, if you set "execute_command": "foo bar", the final execute_command that Packer runs will be ["sh", "-c", "foo bar"]. If you set "execute_command": ["foo", "bar"], the final execute_command will remain ["foo", "bar"].

    Again, the above is only provided as a backwards compatibility fix; we strongly recommend that you set execute_command as an array of strings.

  • inline_shebang (string) - The shebang value to use when running commands specified by inline. By default, this is /bin/sh -e. If you're not using inline, then this configuration has no effect. Important: If you customize this, be sure to include something like the -e flag, otherwise individual steps failing won't fail the provisioner.

  • keep_input_artifact (boolean) - Unlike most other post-processors, the keep_input_artifact option will have no effect for the shell-local post-processor. Packer will always retain the input artifact for shell-local, since the shell-local post-processor merely passes forward the artifact it receives. If your shell-local post-processor produces a file or files which you would like to have replace the input artifact, you may overwrite the input artifact using the artifice post-processor after your shell-local processor has run.

  • only_on (array of strings) - This is an array of runtime operating systems where shell-local will execute. This allows you to execute shell-local only on specific operating systems. By default, shell-local will always run if only_on is not set."

  • use_linux_pathing (bool) - This is only relevant to Windows hosts. If you are running Packer in a Windows environment with the Windows Subsystem for Linux feature enabled, and would like to invoke a bash script rather than invoking a Cmd script, you'll need to set this flag to true; it tells Packer to use the Linux subsystem path for your script rather than the Windows path. (e.g. /mnt/c/path/to/your/file instead of C:/path/to/your/file). Please see the example below for more guidance on how to use this feature. If you are not on a Windows host, or you do not intend to use the shell-local post-processor to run a bash script, please ignore this option. If you set this flag to true, you still need to provide the standard Windows path to the script when providing a script. This is a beta feature.

  • valid_exit_codes (list of ints) - Valid exit codes for the script. By default this is 0.

Execute Command

To many new users, the execute_command is puzzling. However, it provides an important function: customization of how the command is executed. The most common use case for this is dealing with sudo password prompts. You may also need to customize this if you use a non-POSIX shell, such as tcsh on FreeBSD.

The Windows Linux Subsystem

The shell-local post-processor was designed with the idea of allowing you to run commands in your local operating system's native shell. For Windows, we've assumed in our defaults that this is Cmd. However, it is possible to run a bash script as part of the Windows Linux Subsystem from the shell-local post-processor, by modifying the execute_command and the use_linux_pathing options in the post-processor config.

The example below is a fully functional test config.

One limitation of this offering is that "inline" and "command" options are not available to you; please limit yourself to using the "script" or "scripts" options instead.

Please note that this feature is still in beta, as the underlying WSL is also still in beta. There will be some limitations as a result. For example, it will likely not work unless both Packer and the scripts you want to run are both on the C drive.

source "null" "example" {
    communicator = "none"
}

build {
    sources = [
        "source.null.example"
    ]

    post-processor "shell-local"{
        environment_vars  = ["PROVISIONERTEST=ProvisionerTest1"]
        execute_command   = ["bash", "-c", "{{.Vars}} {{.Script}}"]
        use_linux_pathing = true
        scripts           = ["C:/Users/me/scripts/example_bash.sh"]
    }
    post-processor "shell-local"{
        environment_vars  = ["PROVISIONERTEST=ProvisionerTest2"]
        execute_command   = ["bash", "-c", "{{.Vars}} {{.Script}}"]
        use_linux_pathing = true
        script            = "C:/Users/me/scripts/example_bash.sh"
    }
}

Default Environmental Variables

In addition to being able to specify custom environmental variables using the environment_vars configuration, the provisioner automatically defines certain commonly useful environmental variables:

  • PACKER_BUILD_NAME is set to the name of the build that Packer is running. This is most useful when Packer is making multiple builds and you want to distinguish them slightly from a common provisioning script.

  • PACKER_BUILDER_TYPE is the type of the builder that was used to create the machine that the script is running on. This is useful if you want to run only certain parts of the script on systems built with certain builders.

Safely Writing A Script

Whether you use the inline option, or pass it a direct script or scripts, it is important to understand a few things about how the shell-local post-processor works to run it safely and easily. This understanding will save you much time in the process.

Once Per Builder

The shell-local script(s) you pass are run once per builder. This means that if you have an amazon-ebs builder and a docker builder, your script will be run twice. If you have 3 builders, it will run 3 times, once for each builder.

Interacting with Build Artifacts

In order to interact with build artifacts, you may want to use the manifest post-processor. This will write the list of files produced by a builder to a json file after each builder is run.

For example, if you wanted to package a file from the file builder into a tarball, you might write this:

{
  "builders": [
    {
      "content": "Lorem ipsum dolor sit amet",
      "target": "dummy_artifact",
      "type": "file"
    }
  ],
  "post-processors": [
    [
      {
        "output": "manifest.json",
        "strip_path": true,
        "type": "manifest"
      },
      {
        "inline": [
          "jq \".builds[].files[].name\" manifest.json | xargs tar cfz artifacts.tgz"
        ],
        "type": "shell-local"
      }
    ]
  ]
}

This uses the jq tool to extract all of the file names from the manifest file and passes them to tar.

Always Exit Intentionally

If any post-processor fails, the packer build stops and all interim artifacts are cleaned up.

For a shell script, that means the script must exit with a zero code. You must be extra careful to exit 0 when necessary.

Usage Examples:

Windows Host

Example of running a .cmd file on Windows:

post-processor "shell-local" {
  environment_vars = ["SHELLLOCALTEST=ShellTest1"]
  scripts          = ["./scripts/test_cmd.cmd"]
}

Contents of test_cmd.cmd:

echo %SHELLLOCALTEST%

Example of running an inline command on Windows: Required customization: tempfile_extension

post-processor "shell-local" {
  environment_vars   = ["SHELLLOCALTEST=ShellTest2"],
  tempfile_extension = ".cmd",
  inline             = ["echo %SHELLLOCALTEST%"]
}

Example of running a bash command on Windows using WSL: Required customizations: use_linux_pathing and execute_command:

post-processor "shell-local" {
  environment_vars  = ["SHELLLOCALTEST=ShellTest3"],
  execute_command   = ["bash", "-c", "{{.Vars}} {{.Script}}"]
  use_linux_pathing = true
  script            = "./scripts/example_bash.sh"
}

Contents of example_bash.sh:

#!/bin/bash
echo $SHELLLOCALTEST

Example of running a PowerShell script on Windows: Required customizations: env_var_format and execute_command.

post-processor "shell-local" {
  environment_vars = ["SHELLLOCALTEST=ShellTest4"]
  execute_command  = ["powershell.exe", "{{.Vars}} {{.Script}}"]
  env_var_format   = "$env:%s=\"%s\"; "
  script           = "./scripts/example_ps.ps1"
}

Example of running a PowerShell script on Windows as "inline": Required customizations: env_var_format, tempfile_extension, and execute_command

post-processor "shell-local" {
  tempfile_extension = ".ps1"
  environment_vars   = ["SHELLLOCALTEST=ShellTest5"]
  execute_command    = ["powershell.exe", "{{.Vars}} {{.Script}}"]
  env_var_format     = "$env:%s=\"%s\"; "
  inline             = ["write-output $env:SHELLLOCALTEST"]
}

Unix Host

Example of running a Shell script on Unix:

post-processor "shell-local" {
  environment_vars = ["PROVISIONERTEST=ProvisionerTest1"]
  scripts = ["./scripts/example_bash.sh"]
}

Example of running a bash "inline" on Unix:

post-processor "shell-local" {
  environment_vars = ["PROVISIONERTEST=ProvisionerTest2"]
  inline           = ["echo hello", "echo $PROVISIONERTEST"]
}

Example of running a Python script on Unix:

post-processor "shell-local" {
  script           = "hello.py"
  environment_vars = ["HELLO_USER=packeruser"]
  execute_command  = [
    "/bin/sh",
    "-c",
    "{{.Vars}} /usr/local/bin/python {{.Script}}"
  ]
}


Where "hello.py" contains:

    import os

    print('Hello, %s!' % os.getenv("HELLO_USER"))
Edit this page on GitHub