Public Repository

Last pushed: 2 years ago
Short Description
ghost for the digitalgarage
Full Description


Docker image that allows running Ghost in production mode,
and is a bit more configurable than the official Ghost Docker image.

Why yet another container for Ghost?

The official container for Ghost is fine for running in development mode, but it has the wrong
permissions for running in production. That, and the config file doesn't have any easy way to tweak

This container uses the official Ghost image as it's base, has a more "environment aware"
config.js file, and uses these environment variables to tune the config.


docker run --name some-ghost -d ptimof/ghost

This will start Ghost in development mode listening on the default port of 2368.

If you'd like to be able to access the instance from the host without the
contain's IP, standard port mappings can be used:

docker run --name some-ghost -p 8080:2368 -d ptimof/ghost

Then, access it via http://localhost:8080 or http://host-ip:8080 in a browser.


There are three environment variables that can be configured:

  • GHOST_URL: the URL of your blog (e.g.,
  • MAIL_FROM: the email of the blog installation (e.g., '"Webmaster" <>')
  • MAIL_HOST: which host to send email to (e.g.,

These can either be set on the Docker command line directly, or stored in a file and passed on
the Docker command line:

sudo cp ghost.example.env /etc/default/ghost
sudo vi /etc/default/ghost
docker run --name some-ghost --env-file /etc/default/ghost -p 8080:2368 -d ptimof/ghost

If you have just pulled the Docker image with docker pull ptimof/ghost, the example
environment file looks like this:

# Ghost environment
# Place in /etc/default/ghost

MAIL_FROM='"Webmaster" <>'

Running in production

The official Ghost image places the blog content in /var/lib/ghost and exports it as a VOLUME.
This allows two main modes of operation:

Content on host filesystem

In this mode, the Ghost blog content lives on the filesystem of the host with the UID:GID of
1000:1000. If this is acceptable, create a directory somewhere, and use the -v Docker command
line option to mount it:

sudo mkdir -p /var/lib/ghost
sudo chown 1000:1000 /var/lib/ghost
docker run --name some-ghost --env-file /etc/default/ghost -p 80:2368 -v /var/lib/ghost:/var/lib/ghost -d ptimof/ghost npm start --production

Content in a data volume

This is the preferred mechanism to store the blog data. Please see the
Docker documentation
for backup, restore, and migration strategies.

docker create -v /var/lib/ghost --name some-ghost-content busybox
docker run --name some-ghost --env-file /etc/default/ghost -p 80:2368 --volumes-from some-ghost-content -d ptimof/ghost npm start --production

You should now be able to access this instance as in a browser.

Behind a reverse proxy

Of course, you should really be running Ghost behind a reverse proxy, and set things up to auto restart. For that,
a reasonable container would be:

docker create --name some-ghost -h --env-file /etc/default/ghost -p --volumes-from some-ghost-content --restart=on-failure:10 ptimof/ghost npm start --production
docker run some-ghost

Further reading

I've written a bit more about using Ghost as a Docker image:

Docker Pull Command