Docker images deletions fails when instances share the same image ID

[rshad]

Hi all!

Steps and Info. to reproduce the problem:

• OS: Ubuntu 14.04

• Clone master branch.

• Enable Docker images deletion: remove_images: true

• Create 2 instances. kitchen create

• Delete them ktichen destroy

• Note: Tested with instances of "AmazonLinux, Ubuntu 18,04 - 16.04 and Centos7"

---

I was working with the branch master and I faced 2 errors, which I describe on the following.

• Error when creating instances

I first took the master branch, to solve the issue commented in #338, but I got an error when creating the container for each Kitchen instance, which was reported here #355 and solved here #356.

Message: Error getting internal IP of Docker container

So I then took the Fork branch of #356 and this resolved the error.

---

• Error when deleting Docker images.

After setting remove_images: true I got the following error when deleting the docker images.

Error response from daemon: conflict: unable to delete xxxx (cannot be forced) - image is being used by running container xxxxx**

After some debugging, I found that when creating instances with the same provisioning commands, Kitchen use the same image for such instances. And when running kitchen destroy, Kitchen proceed by instance:image pairs. But in this case what we have of such pairs is: (2 instances case)

• instance_1 : Image_0
• instance_2: Image_0

So both instances' containers used the same image.

When deleting, the process flow is as follows:

• Delete instance_1: ✔️

• Delete Corresponding Docker Image "Image_0": ❌
  'image ID xx can't be deleted. still being used by a running container "instance_2'

• Delete instance_2: ✔️

• Delete Corresponding Docker Image "Image_0": ✔️

kr,

Rshad