Octopus and Docker

Legacy Modelling
1

I have been looking to integrate docker into our octopus pipeline and am excited to see the work the octopus team has been doing to make this a reality. I read through your docker guides in octopus and have a few questions.

The guides describe how to deploy docker images, but not really how to deploy steps to those containers?

This is what I would ultimately like to do.

  1. Use a windows server image
  2. Create a container
  3. Deploy the project to the container
  4. Start the container
  5. Tear down the container.

Is this possible?

2

Hi Manny

Thanks for reaching out!

At the moment, this is not really a model we support out of the box.

We currently treat a docker image as a full artifact (ie, a runnable container), rather than creating a container and deploying to it. We expect that the container has been created as part of your CI build, rather than creating a container on the fly.

That said, this is something you could script yourself, but it wouldn’t be that simple. You would have to install a Tentacle inside the container, and then trigger a deployment to it.

We are currently doing some work to publish a docker image with the Tentacle installed, which should be available in the next few weeks.

Once this is available (or you’ve implemented your own), you could setup something similar to:

Project 1:

  • Start container with Tentacle in it
  • Trigger Project 2 using the chain deployment step template

Project 2:

  • Deploy package to Tentacle

In theory, this should work, but I’m not sure I’d treat it as a recommended approach :). Personally, I’d stick to building it on your CI server.

Hope that helps!

Regards,
Matt