docker_container.ubuntu: Unable to start container: API error (400):

[hashibot]

This issue was originally opened by @maks-us as hashicorp/terraform#15141. It was migrated here as part of the provider split. The original body of the issue is below.

---

Hi there,

Terraform Version

Terraform v0.9.6
Mac OS 10.12.5
Docker version 17.06.0-ce-rc1, build 7f8486a

Affected Resource(s)

resource "docker_container" "ubuntu"
resource "docker_image" "ubuntu"

It is example from this page https://www.terraform.io/docs/providers/docker/r/container.html

Debug Output

Error applying plan:
1 error(s) occurred:

• docker_container.ubuntu: 1 error(s) occurred:
• docker_container.ubuntu: Unable to start container: API error (400): {"message":"starting container with non-empty request body was deprecated since v1.10 and removed in v1.12"}

Expected Behavior

work example from the documentation website

Actual Behavior

start ubuntu container

Steps to Reproduce

1. create main.tf
2. add content:

resource "docker_container" "ubuntu" {
  name  = "foo"
  image = "${docker_image.ubuntu.latest}"
  command = ["top"]
}

resource "docker_image" "ubuntu" {
  name = "ubuntu:precise"
}

3. run:
   terraform apply

[dubo-dubon-duponey]

Apologies for the noise/previous comment 🤦‍♂️

This is just working fine (unable to reproduce), and the original bug may have been fixed (some of the logs are still confusing, but that would be a different issue).

[mavogel]

No worries @dubo-dubon-duponey :) Thank you for contributing. Which logs are still confusing? Do you have an example?

[dubo-dubon-duponey]

I'm using multiple docker-nodes in the same terraform project.

When testing the above, I mistakenly forgot to specify the provider for the docker_image resource:

resource "docker_image" "ubuntu" {
  name = "ubuntu:bionic"
}

resource "docker_container" "ubuntu" {
  provider = docker.something
  name  = "foo"
  image = docker_image.ubuntu.latest
}

... which means that the image itself was pulled on my laptop, while the container tried to start on a different node ("something").

The error message you get in that case is as follow:

Error: Unable to create container with image sha256:2ca708c1c9ccc509b070f226d6e4712604e0c48b55d7d8f5adc9be4a4d36029a: Unable to pull image sha256:2ca708c1c9ccc509b070f226d6e4712604e0c48b55d7d8f5adc9be4a4d36029a: error pulling image sha256:2ca708c1c9ccc509b070f226d6e4712604e0c48b55d7d8f5adc9be4a4d36029a: Error response from daemon: pull access denied for sha256, repository does not exist or may require 'docker login': denied: requested access to the resource is denied

This is confusing indeed, as this is typically what you get if you typo the image name or are not logged-in into the hub - which is definitely not the situation here.

[mavogel]

Hmm, TBH I don't think we can catch those configuration errors. What would you suggest @dubo-dubon-duponey ? Wait for hashicorp/terraform-plugin-sdk#61 as well?

[dubo-dubon-duponey]

Yeah. That's minor either way.

[dubo-dubon-duponey]

And as far as this ticket is concerned, I couldn't reproduce and probably got fixed some time ago.