using docker to deliver qualification tools

Stanislav Pankevich s.pankevich at gmail.com
Tue Oct 29 13:24:45 UTC 2019


Hello,

This is a comment for Andrew from a somewhat passive reader of the RTEMS forums.

One way of having Docker while staying platform-independent is to use
a "Docker without Dockerfiles" approach like described here: [1].

The idea is very simple: you delegate all provisioning jobs to
Ansible. If your building process is all written in the Ansible
playbooks, you can run these playbooks with Docker or other
technologies such as Vagrant and virtual machines.

My use case is rather the opposite of what you seem to have: we use
Ansible playbooks and Vagrant / VirtualBox for a number of operating
systems but I also want to build the Docker containers with the same
Ansible playbooks that we use for VMs because Docker is easier and
faster when I need to jump on Linux-specific tasks from a macOS
machine.

I have created a simple example [2] that should give you an idea of
how you make Ansible build a Docker container and then create an image
from it using 'docker commit'. You can later use the same playbook to
build any other technology like Vagrant/VirtualBox.

I am not sure how much is this relevant for you but keeping build
steps in Ansible seems like a more scalable solution that potentially
unlocks your solution from being a Docker-only.

[1] https://tech.labs.oliverwyman.com/blog/2019/08/30/docker-without-dockerfiles/
[2] https://github.com/stanislaw/docker-with-ansible-example

Stanislav

On Thu, Oct 24, 2019 at 2:00 PM Joel Sherrill <joel at rtems.org> wrote:
>
>
>
> On Thu, Oct 24, 2019, 5:58 AM Andrew Butterfield <Andrew.Butterfield at scss.tcd.ie> wrote:
>>
>> Dear RTEMS users,
>>
>>  give that the RTEMS-SMP qualification project requires a lot of new
>> tooling, some very standard, some more experimental, we have a proposal
>> that we distribute these using Docker CE (Community Edition).
>>
>> This will make it easier for users to install only the parts that they
>> are interested in, and also makes it possible for us to host the
>> tools in the cloud somewhere as an on-line qualification service.
>>
>> What are the thoughts of the community regarding this?
>
>
> I don't care what ESA does after the work is delivered for their own internal use. But requiring a Docker instance for community qualification work is unacceptable to me. It is a Linux specific technology and that violates a few key tenets of the RTEMS project.
>
> Host Independence and Reproducible being the top two.
>
> Instructions on how to build the environment are key. This includes what to install with pop or with the RSB or as a host package.
>
> If a user wants a pre-configured VM or container that's their business. I personally don't think it is configuration controlled enough under the auspices of the project and even if used must be under the end user system's control to be able to pass a qualification audit.
>
> Think reproducible
>
>
>
>>
>> Best regards,
>>   Andrew Butterfield
>> --------------------------------------------------------------------
>> Andrew Butterfield     Tel: +353-1-896-2517     Fax: +353-1-677-2204
>> Lero at TCD, Head of Foundations & Methods Research Group
>> School of Computer Science and Statistics,
>> Room G.39, O'Reilly Institute, Trinity College, University of Dublin
>>                          http://www.scss.tcd.ie/Andrew.Butterfield/
>> --------------------------------------------------------------------
>>
>> _______________________________________________
>> users mailing list
>> users at rtems.org
>> http://lists.rtems.org/mailman/listinfo/users
>
> _______________________________________________
> users mailing list
> users at rtems.org
> http://lists.rtems.org/mailman/listinfo/users


More information about the users mailing list