Workaround for inconsistency container status

[shawn111]

- What I did Fix stopped container still report as running.

- How I did it When containerd restarted, check correct container running status

- How to verify it

# dockerd-> containerd -> containerd-shim -> <runc_container_process>

$ docker run xxx
$ docker ps  ## correct, shows one container is up

# simulate the heavy scenario
$ kill -STOP <dockerd_PID>

$ kill <docker-containerd-shim_PID> or kill <runc_container_process_PID>
$ kill <docker-containerd_PID>

$ kill -CONT <dockerd_PID>
$ docker ps  ## wrong, still shows one container is up

After, apply the patch, the final docker ps could be correct.

Reference issue:

• https://github.com/moby/moby/issues/38501

[rh-atomic-bot]

Can one of the admins verify this patch? I understand the following commands:

• bot, add author to whitelist
• bot, test pull request
• bot, test pull request once

[rhatdan]

bot, add author to whitelist LGTM You need to sign your commit

git commit -a --amend -s git push --force

Was this a backport from upstream?

[shawn111]

Hi @rhatdan,

Not, it is not from upstream. Maybe I should send another PR for upstream.

Thanks for your review.

[rhatdan]

Yes send it to them and get their approval.

[rhatdan]

Hopefully we don't have this inconsistency in Podman.