Containers created with --restart=always do not restart when started through a pod

[nivekuil]

Issue Description

containers created with --restart=always do not restart when they exit, when started through the parent pod

Steps to reproduce the issue

Steps to reproduce the issue

1. podman pod create --name test-pod
2. podman create --restart=always --pod test-pod docker.io/bash exit 0
3. podman pod start test-pod
4. podman ps -a

Describe the results you received

See that the container Exited (127) N seconds ago and is not restarting

Describe the results you expected

I think --restart=always should always restart the container if it exits regardless of how it's started

podman info output

4.6.0

Podman in a container

No

Privileged Or Rootless

None

Upstream Latest Release

Yes

Additional environment details

Additional environment details

Additional information

Additional information like issue happens only occasionally or issue happens with a particular architecture or on a particular setting

[nivekuil]

@rhatdan thanks for putting up the patch. Are you able to put out a rush release? I see that there was discussion about this in #17627 where reviewers didn't want to break this exact situation, but unfortunately the line in question would actually have prevented the current situation where all such workloads on 4.6 are at very high risk of unexpected downtime.

[rhatdan]

@mheon WDYT?

[mheon]

We are planning a 4.6.2 soon, aiming to finalize plans for when on Monday.