vagrant/website/docs/source/v2/docker/configuration.html.md

95 lines
3.7 KiB
Markdown

---
page_title: "Configuration- Docker Provider"
sidebar_current: "docker-configuration"
---
# Docker Configuration
The Docker provider has some provider-specific configuration options
you may set. A complete reference is shown below.
### Required
* `build_dir` (string) - The path to a directory containing a Dockerfile.
One of this or `image` is required.
* `image` (string) - The image to launch, specified by the image ID or a name
such as `ubuntu:12.04`. One of this or `build_dir` is required.
### Optional
General settings:
* `build_args` (array of strings) - Extra arguments to pass to
`docker build` when `build_dir` is in use.
* `cmd` (array of strings) - Custom command to run on the container.
Example: `["ls", "/app"]`.
* `create_args` (array of strings) - Additional arguments to pass to
`docker run` when the container is started. This can be used to set
parameters that aren't exposed via the Vagrantfile.
* `env` (hash) - Environmental variables to expose into the container.
* `expose` (array of integers) - Ports to expose from the container
but not to the host machine. Useful for links.
* `link` (method, string argument) - Link this container to another
by name. The argument should be in the format of `(name:alias)`.
Example: `docker.link("db:db")`. Note, if you're linking to
another container in the same Vagrantfile, make sure you call
`vagrant up` with the `--no-parallel` flag.
* `force_host_vm` (boolean) - If true, then a host VM will be spun up
even if the computer running Vagrant supports Linux containers. This
is useful to enforce a consistent environment to run Docker.
* `has_ssh` (boolean) - If true, then Vagrant will support SSH with
the container. This allows `vagrant ssh` to work, provisioners, etc.
This defaults to false.
* `host_vm_build_dir_options` (hash) - Synced folder options for the
`build_dir`, since the build directory is synced using a synced folder
if a host VM is in use.
* `name` (string) - Name of the container. Note that this has to be unique
across all containers on the host VM. By default Vagrant will generate
some random name.
* `ports` (array of strings) - Ports to expose from the container to the
host. These should be in the format of `host:container`.
* `remains_running` (boolean) - If true, Vagrant expects this container
to remain running and will make sure that it does for a certain amount
of time. If false, then Vagrant expects that this container will
automatically stop at some point, and won't error if it sees it do that.
* `stop_timeout` (integer) - The amount of time to wait when stopping
a container before sending a SIGTERM to the process.
* `vagrant_machine` (string) - The name of the Vagrant machine in the
`vagrant_vagrantfile` to use as the host machine. This defaults to
"default".
* `vagrant_vagrantfile` (string) - Path to a Vagrantfile that contains
the `vagrant_machine` to use as the host VM if needed.
* `volumes` (array of strings) - List of directories to mount as
volumes into the container. These directories must exist in the
host where Docker is running. If you want to sync folders from the
host Vagrant is running, just use synced folders.
Below, we have settings related to auth. If these are set, then Vagrant
will `docker login` prior to starting containers, allowing you to pull
images from private repositories.
* `email` (string) - Email address for logging in.
* `username` (string) - Username for logging in.
* `password` (string) - Password for logging in.
* `auth_server` (string) - The server to use for authentication. If not
set, the Docker Hub will be used.