Handle an already connected network in libpod API

[kubealex]

Signed-off-by: Alessandro Rossi [email protected]
The PR introduces a new "ErrNetworkConnected" error in libpod defines, to handle the specific case where a user tries to connect a container to a network that is already connected to it.

The user from the CLI will still receive an error, but since from API point of view the operation results in nothing to be done and the state is preserved, it should return 200, aligned with the behavior that docker has.

This should address #15499

Does this PR introduce a user-facing change?

None

ErrNetworkConnected was introduced for when trying to connect a container to a network it is already connected to

[baude]

/approve

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: baude, kubealex

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [baude]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[baude]

LGTM, I missed with your commit and release note message a little. Thanks for the PR

[kubealex]

@baude seems like there are issues with test image pull 😔

[n1hility]

@kubealex @baude IMO we should cover the CLI case as well, since a goal for podman is alias docker=podman. One other factor when it comes to compat, is that docker still errors on this condition once the container hits started state.

Something like this: n1hility@53d45ae

Sorry for the collision: I had self-assigned this issue, had time to work on it and then when I went to open a PR saw this referenced on the issue. Feel free to pull this in and rework.

-Jason

[kubealex]

@n1hility docker CLI preserves the error, hence my idea of creating an error that could fit better the use case.
For docker, CLI reports a strict error, stating that the container already is an endpoint for the network, while API reports 200 :)

[n1hility]

@n1hility docker CLI preserves the error, hence my idea of creating an error that could fit better the use case.
For docker, CLI reports a strict error, stating that the container already is an endpoint for the network, while API reports 200 :)

Right This only happens when started. Ex:

$ docker create alpine sleep 86400
408f8751d16a5566c506e17182fa26643513762eae2d26bf0d3e88b8ff722c0a
$ docker network connect bridge 408f8751d16a5566c506e17182fa26643513762eae2d26bf0d3e88b8ff722c0a
$ echo $?
0
$ docker start 408f8751d16a5566c506e17182fa26643513762eae2d26bf0d3e88b8ff722c0a
408f8751d16a5566c506e17182fa26643513762eae2d26bf0d3e88b8ff722c0a
$ docker network connect bridge 408f8751d16a5566c506e17182fa26643513762eae2d26bf0d3e88b8ff722c0a
Error response from daemon: endpoint with name romantic_pike already exists in network bridge

(same on the REST interface)

[kubealex]

Right, I'll focus on that tomorrow!
Thx a lot!

[kubealex]

@n1hility I just merged your effort in the commit, not sure if I did it correctly, feel free to rework it if it isn't compliant, also reach out in GChat to ease the conversation :)

[n1hility]

@kubealex cool thanks! I just made one small change with the leftover few lines on the rest endpoint, since the error is filtered early. This also allows the rest endpoint to reflect the start error conditions.

[kubealex]

Nice @n1hility 💪 thanks a lot for the integration!

[rhatdan]

/lgtm

[n1hility]

@kubealex you are most welcome @kubealex. It was great collaborating with you, thank you for digging in and submitting a PR!

[n1hility]

/cherry-pick v4.2

[openshift-cherrypick-robot]

@n1hility: new pull request created: #15554

In response to this:

/cherry-pick v4.2

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.