How do I run a Docker container that uses SystemD from the latest version of Ubuntu (18.10)?

Question

I'm trying to execute a Docker image built using 'ubuntu:latest' and I keep getting SystemD error messages when I run the container:

System has not been booted with systemd as init system (PID 1). Can't operate.

If I try this solution and spawn the container using docker run -it -e container=docker your-image-name /sbin/init, I get the following error:

Failed to mount tmpfs at /run: Operation not permitted
Failed to mount tmpfs at /run/lock: Operation not permitted
[!!!!!!] Failed to mount API filesystems, freezing.
Freezing execution.

What should I try differently?

Nick

You generally don't want to run a process manager like systemd (or anything else) inside a container. What exactly are you trying to do? – Crus

I'm trying to startup an Nginx and SSHD container. How do you avoid SystemD if all the Linux distros have already moved to it? seems like this is an issue that should have gotten ironed out already? – Nick

It has been ironed out. You don't use a processor manager in your container. You start the service you want, and that's it. Take a look at any of the official containers for things like nginx, httpd, mysql, etc. You don't run ssh inside your container; that's what docker exec is for. – Crus

Hmm okay, thanks. I need SSH because I use Ansible to send it config files. :) – Nick

It sounds like you're managing your container like a VM. I strongly recommend against that. The most you should be doing with ansible is building an image and running the container. – Livery

Not so much. The config files are going into the container to be hosted by the Nginx server. It's pretty basic actually. – Nick

You typically provide that configuration before the container starts up, maybe injecting them using a docker run -v option. Running an ssh daemon in a container IME is pretty unusual (and presents some tricky management problems). – Stereobate

These are config files that get hosted by Nginx. They're not used by the Docker container itself. I found that if my host OS is running Bionic or later, the same error messages disappear using the /sbin/init method from above. – Nick

Answer 1

You can solve this using a different version, like 16.04:

docker run -d \
  -h ubuntu \
  --name ubuntu \
  --privileged \
  docker.io/library/ubuntu:16.04 /sbin/init

After run, you can accessed using the follow command:

docker exec -it ubuntu /bin/bash

This version uses systemd.

Edithe

Answer 2

In order to run ubuntu 18 systemd inside docker, you can use this:

Run container:

docker run -d --name stack_overflow_53750952 --privileged -v /sys/fs/cgroup:/sys/fs/cgroup:ro jrei/systemd-ubuntu:18.04

Enter in container:

docker exec -it stack_overflow_53750952 bash

Then as you can see: screenshot of systemd running inside the container

Fermata

Answer 3

Upgrade to Ubuntu Bionic and try /sbin/init method

Nick

My question was for Ubuntu, not CentOS – Nick