Add --restart flag to pod create

[umohnani8]

Add --restart flag to pod create to allow users to set the restart policy for the pod, which applies to all the containers in the pod. This reuses the restart policy already there for containers and has the same restart policy options. Add "never" to the restart policy options to match k8s syntax. It is a synonym for "no" and does the exact same thing where the containers are not restarted once exited.
Only the containers that have exited will be restarted based on the restart policy, running containers will not be restarted when an exited container is restarted in the same pod (same as is done in k8s).

Add Restarts column to the podman ps and podman pod ps output to show how many times a
container was restarted based on its restart policy when --format={{.Restarts}}.

Signed-off-by: Urvashi Mohnani [email protected]

Does this PR introduce a user-facing change?

Add --restart flag to pod create to be able to set the restart policy for all the containers in a pod.
Add Restarts column to the output of podman ps to show the number of times a container has been restarted based on its restart policy when --format={{.Restarts}}.
Add Restarts column to the output of podman pod ps when --format={{.Restarts}} to show the total number of container restarts in a pod.

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: umohnani8

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 [umohnani8]

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

[TomSweeneyRedHat]

After a quick skim, looks good in general. Test aren't at all happy.

[vrothberg]

I had something else in mind for pod create --restart. Currently, it does not yield a podman pod restart but instead, it'll set it for all containers. There are still some open questions in the design doc as well.

@mheon @giuseppe PTAL

[umohnani8]

If we are to follow how kubernetes does restarts for pods, it doesn't restart the whole pod based on the policy, the containers in the pod are what is restarted. So when we set the restart policy for a pod, we are setting the restart policy for all the containers in the pod and a container in the pod cannot have a different policy set for it. If the whole pod needs to be restarted, k8s removes the pod and creates a new one in place of it and that is a whole different thing that happens when using Deployments or Jobs or when the user deletes the pod and the policy was set to Always.
https://kubernetes.io/docs/concepts/workloads/pods/pod-lifecycle/#restart-policy

The design doc outlined the object of --restart for pod to apply it to the containers and have the containers restart based on the policy and there were no objections on that. We can continue the discussions in the design doc if needed.

[vrothberg]

If we are to follow how kubernetes does restarts for pods, it doesn't restart the whole pod based on the policy, the containers in the pod are what is restarted. So when we set the restart policy for a pod, we are setting the restart policy for all the containers in the pod and a container in the pod cannot have a different policy set for it. If the whole pod needs to be restarted, k8s removes the pod and creates a new one in place of it and that is a whole different thing that happens when using Deployments or Jobs or when the user deletes the pod and the policy was set to Always. https://kubernetes.io/docs/concepts/workloads/pods/pod-lifecycle/#restart-policy

I personally don't mind K8s in this context as it's outside of kube play.

The design doc outlined the object of --restart for pod to apply it to the containers and have the containers restart based on the policy and there were no objections on that. We can continue the discussions in the design doc if needed.

OK, let's do that 👍 I misinterpreted the part in the doc w.r.t "apply it to all containers". If the pod gets restarted it affects all its containers in a way.

[umohnani8]

Once this is in, I will create a follow up PR for kube gen and play to start using the pod restart policy.

[umohnani8]

@edsantiago @mheon is there a way for me to not list .RestartCount in the docs? I want the option the users know and use to be .Restarts.
This is the validation error I am getting https://github.com/containers/podman/pull/17627/checks?check_run_id=12662786279

[edsantiago]

/me groans in agony.

@Luap99 you might be able to answer that more concisely and accurately than anything I could do.

[Luap99]

The reason this shows up is because it simply speaking a valid go template option. As soon a s you add a public field or function to a type it will be included in the shell completion which the validate script compare against.
As of now there is no way within the go code to explicitly hide a field or function. I agree that we may want to hide some fields, there is #17472 which goes into more detail but is very much not even close to be implement any time soon.

For now the simple solution is to just rename the field from RestartCount to Restarts on the ListContainer type and drop the Restarts function from the psReporter type. I think this should give you your desired result.

[umohnani8]

Tests are green, @containers/podman-maintainers PTAL

[vrothberg]

LGTM other than the final nits

[rhatdan]

/lgtm

[vrothberg]

@mheon, no need to backport #18481 since the changes here did not make it into 4.5. The reports were against 4.6-dev, so we're good.