packer-cn/website/source/docs/provisioners/shell.html.md

10 KiB

description layout page_title sidebar_current
The shell Packer provisioner provisions machines built by Packer using shell scripts. Shell provisioning is the easiest way to get software installed and configured on a machine. docs Shell - Provisioners docs-provisioners-shell-remote

Shell Provisioner

Type: shell

The shell Packer provisioner provisions machines built by Packer using shell scripts. Shell provisioning is the easiest way to get software installed and configured on a machine.

-> Building Windows images? You probably want to use the PowerShell or Windows Shell provisioners.

Basic Example

The example below is fully functional.

{
  "type": "shell",
  "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:

  • 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 upload and execute in the machine. 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 uploaded and 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:

  • binary (boolean) - If true, specifies that the script(s) are binary files, and Packer should therefore not convert Windows line endings to Unix line endings (if there are any). By default this is false.

  • 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.

  • execute_command (string) - The command to use to execute the script. By default this is chmod +x {{ .Path }}; {{ .Vars }} {{ .Path }}. The value of this is treated as configuration template. There are two available variables: Path, which is the path to the script to run, and Vars, which is the list of environment_vars, if configured.

  • expect_disconnect (bool) - Defaults to true. Whether to error if the server disconnects us. A disconnect might happen if you restart the ssh server or reboot the host. May default to false in the future.

  • 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.

  • remote_folder (string) - The folder where the uploaded script will reside on the machine. This defaults to '/tmp'.

  • remote_file (string) - The filename the uploaded script will have on the machine. This defaults to 'script_nnn.sh'.

  • remote_path (string) - The full path to the uploaded script will have on the machine. By default this is remote_folder/remote_file, if set this option will override both remote_folder and remote_file.

  • skip_clean (boolean) - If true, specifies that the helper scripts uploaded to the system will not be removed by Packer. This defaults to false (clean scripts from the system).

  • start_retry_timeout (string) - The amount of time to attempt to start the remote process. By default this is 5m or 5 minutes. This setting exists in order to deal with times when SSH may restart, such as a system reboot. Set this to a higher value if reboots take a longer amount of time.

Execute Command Example

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.

Sudo Example

Some operating systems default to a non-root user. For example if you login as ubuntu and can sudo using the password packer, then you'll want to change execute_command to be:

"echo 'packer' | sudo -S sh -c '{{ .Vars }} {{ .Path }}'"

The -S flag tells sudo to read the password from stdin, which in this case is being piped in with the value of packer.

By setting the execute_command to this, your script(s) can run with root privileges without worrying about password prompts.

FreeBSD Example

FreeBSD's default shell is tcsh, which deviates from POSIX semantics. In order for packer to pass environment variables you will need to change the execute_command to:

chmod +x {{ .Path }}; env {{ .Vars }} {{ .Path }}

Note the addition of env before {{ .Vars }}.

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.

  • PACKER_HTTP_ADDR If using a builder that provides an http server for file transfer (such as hyperv, parallels, qemu, virtualbox, and vmware), this will be set to the address. You can use this address in your provisioner to download large files over http. This may be useful if you're experiencing slower speeds using the default file provisioner. A file provisioner using the winrm communicator may experience these types of difficulties.

Handling Reboots

Provisioning sometimes involves restarts, usually when updating the operating system. Packer is able to tolerate restarts via the shell provisioner.

Packer handles this by retrying to start scripts for a period of time before failing. This allows time for the machine to start up and be ready to run scripts. The amount of time the provisioner will wait is configured using start_retry_timeout, which defaults to a few minutes.

Sometimes, when executing a command like reboot, the shell script will return and Packer will start executing the next one before SSH actually quits and the machine restarts. For this, put use "pause_before" to make Packer wait before executing the next script:

{
  "type": "shell",
  "script": "script.sh",
  "pause_before": "10s"
}

Some OS configurations don't properly kill all network connections on reboot, causing the provisioner to hang despite a reboot occurring. In this case, make sure you shut down the network interfaces on reboot or in your shell script. For example, on Gentoo:

/etc/init.d/net.eth0 stop

SSH Agent Forwarding

Some provisioning requires connecting to remote SSH servers from within the packer instance. The below example is for pulling code from a private git repository utilizing openssh on the client. Make sure you are running ssh-agent and add your git repo ssh keys into it using ssh-add /path/to/key. When the packer instance needs access to the ssh keys the agent will forward the request back to your ssh-agent.

Note: when provisioning via git you should add the git server keys into the ~/.ssh/known_hosts file otherwise the git command could hang awaiting input. This can be done by copying the file in via the file provisioner (more secure) or using ssh-keyscan to populate the file (less secure). An example of the latter accessing github would be:

{
  "type": "shell",
  "inline": [
    "sudo apt-get install -y git",
    "ssh-keyscan github.com >> ~/.ssh/known_hosts",
    "git clone git@github.com:exampleorg/myprivaterepo.git"
  ]
}

Troubleshooting

My shell script doesn't work correctly on Ubuntu

  • On Ubuntu, the /bin/sh shell is dash. If your script has bash-specific commands in it, then put #!/bin/bash -e at the top of your script. Differences between dash and bash can be found on the DashAsBinSh Ubuntu wiki page.

My shell works when I login but fails with the shell provisioner

  • See the above tip. More than likely, your login shell is using /bin/bash while the provisioner is using /bin/sh.

My installs hang when using apt-get or yum

  • Make sure you add a -y to the command to prevent it from requiring user input before proceeding.

How do I tell what my shell script is doing?

  • Adding a -x flag to the shebang at the top of the script (#!/bin/sh -x) will echo the script statements as it is executing.

My builds don't always work the same

  • Some distributions start the SSH daemon before other core services which can create race conditions. Your first provisioner can tell the machine to wait until it completely boots.
{
  "type": "shell",
  "inline": [ "sleep 10" ]
}