Common Setup

Project Code

The filesystem within containers is ephemeral! Any changes made there do not persist if the container is restarted. (See "Persistent Data Volume" below for a storage area that is preserved.)

In typical Docker images the code is built directly into the container at the /code path. This is a great mechanism that allows the container to be "self contained" (pun intended), immutable and not need any checkout/file system. You also know when you run a container exactly what code is in there because you generally don’t change the code unless you rebuild the image.

For development purposes, however, this is problematic because it is burdensome to rebuild an image for each code change. To solve this, we mount project code from the Host Machine into the running container, effectively overriding the files built directly into the image. The running container is then using the local project file system for the overridden paths rather than the file system built into the image. This allows a developer to use an IDE and edit code directly on the local file system of the Host Machine, but execute that code within the environment of the running container.

Note on Project Code Location

Your project code must be located somewhere within your home directory (/Users for Macs) on your local machine. This is because NFS shares your home directory into the Docker Host VM, and only files on the Docker Host VM can be referenced in volume shares.

Persistent Data Volume

DevTools maintains a data volume on the Docker Host that is mounted at /data into every container. This volume is persistent so long as you do not perform a devtools remove operation. This ensures that file access on the VM is done natively for things like databases and other things where filesystem performance matters. If you configure a container to write to this area, you should use a project and container based namespace to prevent conflicts as this is a shared resource. For example, you may want to use a namespacing method like /data/[project name]/[environment]/[service name] as a safe location to write data.

Any code from your local directories is directly shared in to the Docker Machine VM via NFS at /Users. This means that even if you destroy and re-create your Docker Host, your code will be safe since it lives on the Host Machine.


Any files that are generated or changed within a running container that you want to persist after the container is stopped should be put onto a volume that is mounted into the container from the local machine. A Docker container represents immutable infrastructure, the files on the image are able to be changed at runtime but typically do not persist. When the container is stopped and run again it "boots" the files that were built into the original image. The volume at /data (mentioned above) is a persistent volume that you can use.