Last year, I switched all of my hosting from arbitrarily installed packages to Docker. This made installing and configuring incredibly simple, but updating a little less defined. Whilst Docker itself is updated through the system package manager (probably), the containers themselves aren’t. Docker container versions are known as “tags”, and…
Docker containers, and containers as a whole, are really just a regular program wrapped in some extra protections provided by the kernel (namely
cgroups etc) to create isolation, and other interesting features.
Unlike VMs, containers run closer to the host operating system, so close they use the same kernel, meaning it’s even more important to protect it. VMs are a much better understood technology, and have a lot more isolation. Most commercial container hosting offerings just run your containers in VMs, to massively reduce inter-client security issues.
A container is just a single process. Whether that be a web server like Nginx, database like PostgreSQL or an init system like
s6 to spawn and handle even more processes. But, when inside a container, which user does the process run as.
#“Who does a container run as?”
By default, containers are run as
dockerd (the docker daemon) runs as
root, and this is normal.
root is needed to configure certain container aspects needed to function correctly. There may be ways of running without
root, but it’s fine as it is.
#What’s wrong with containers running as root?”
There are so many reasons not to run all your processes as
root. Just because the process is in a container, doesn’t mean it’s completely protected, nor that these reasons don’t apply.
If there’s a vulnerability in the application, then an attacker can gain root access into the container. Inside the container, the user is root, and so can do whatever they want in the container. With this, an attacker can not only mess with the application, but potentially install additional tools to help pivot to other devices or containers.
If there’s a vulnerability in docker, or the kernel itself, allowing a process inside the container to break out, then they now have a process running on your host as
root. Game Over.
#Changing the user
Changing the user running prevents the previous issues. Not all containers will just deal with it. Many containers, if configured incorrectly, will stop functioning entirely if you try to change the user without them expecting it.
As a container user, you’re at the mercy of the container maintainers as to the quality of the support for changing user.
Docker itself supports changing the user using the
--user argument (or
user key in
docker-compose.yml). This argument takes the user id of the user to change the process to. Note that this doesn’t change anything else about the container.
LinuxServer.io, the makers of a bunch of high-quality containers, use the
$PGID environment variables to configure the user and group of the process and related files. This strategy works very well as it’ll make sure the application file permissions match the process user, so applications will always have access to their files.
Some applications, like nginx, already handle changing user as part of normal operation. Nginx requires root access to bind to port 80, but the processes handling user requests or executing further scripts (PHP etc) is configured inside nginx itself. For applications like this, changing the user is probably not what you want, and should be handled in the application configuration itself.
As a container maintainer, the
USER keyword in the
Dockerfile sets the running user. This user will need to be created manually, and any required files should be
chown-d to match this user.
An alternative approach is to change user at startup. This means the entry point script is still run as
root, but in much the same way applications like nginx change user, this isn’t for very long. Tools like
s6-overlay make this very easy to manage (this is the same tool LSIO use).
It’s not a good idea to have containers running as root, for the same reasons it’s not a good idea to run all your processes as root. Some containers still run as root, but perhaps this is for good reason (or they’re not fussed).
See a container which runs as
root? Raise an issue!
You can more about container security at docs.docker.com.
Share this page
You should back up your data, properly! If you’re not, you’re playing a dangerous game with fate. Computers are pretty reliable, but they also go wrong, often. You should always backup your files, but backing up a containerized application isn’t quite as simple. A container is 3 things:ConfigurationVolumesNetworking The point…
NFS is a great protocol for sharing files quickly and simply over the network. Whilst it’s not designed for end user use, it’s great for mounting directories from remote machines, and having them be performant. NFS’ lack of authentication is in a way a feature, honest. Not only does it…
View all →