What's the difference in podman_pod's state started and restarted?

[Tronde]

Dear team,
along the possible values for containers.podman.podman_pod's state are startedandrestarted`. They somehow behave differently from what I expect. I would appreciate if you can take the time to review the following example and my question.

Example scenario

I have created a pod consisting of three containers.

]$ podman pod ps
POD ID        NAME        STATUS      CREATED        INFRA ID      # OF CONTAINERS
3e0b4e78d226  nc_pod      Running     6 minutes ago  c141f6351cec  3

]$ podman ps
CONTAINER ID  IMAGE                                    COMMAND               CREATED        STATUS        PORTS                    NAMES
c141f6351cec  localhost/podman-pause:5.1.1-1717459200                        6 minutes ago  Up 6 minutes  127.0.0.1:40671->80/tcp  3e0b4e78d226-infra
d2284c00dec3  docker.io/library/mariadb:10.11.2        mariadbd              6 minutes ago  Up 6 minutes  127.0.0.1:40671->80/tcp  nc_mariadb
7d8b7285fda7  docker.io/library/nextcloud:25-apache    apache2-foregroun...  6 minutes ago  Up 6 minutes  127.0.0.1:40671->80/tcp  nextcloud

]$ podman pod inspect -f '{{ .CreateCommand }}' nc_pod
[podman pod create --infra=True --infra-conmon-pidfile /tmp/nc_pod_infra.pid --publish 127.0.0.1:40671:80 --name nc_pod]

]$ podman inspect -f '{{ .Config.CreateCommand }}' nc_mariadb
[podman container run --name nc_mariadb --conmon-pidfile /tmp/mariadb_conmon.pid --pod nc_pod --env MYSQL_USER=<masked> --env MYSQL_PASSWORD=<masked> --env MYSQL_ROOT_PASSWORD=<masked> --env MYSQL_DATABASE=nc_db --volume mysql_data:/var/lib/mysql:Z --detach=True docker.io/library/mariadb:10.11.2]

]$ podman inspect -f '{{ .Config.CreateCommand }}' nextcloud
[podman container run --name nextcloud --conmon-pidfile /tmp/nc_conmon.pid --pod nc_pod --env MYSQL_DATABASE=nc_db --env MYSQL_USER=<masked> --env MYSQL_PASSWORD=<masked> --env MYSQL_HOST=127.0.0.1 --env NEXTCLOUD_ADMIN_USER=<masked> --env NEXTCLOUD_ADMIN_PASSWORD=<masked> --env NEXTCLOUD_TRUSTED_DOMAINS= --env SMTP_HOST=smtp.example.com --env SMTP_SECURE=tls --env SMTP_PORT=587 --env SMTP_AUTHTYPE=LOGIN --env [email protected] --env SMTP_PASSWORD=MailSecret1! --env MAIL_FROM_ADDRESS=no-reply --env MAIL_DOMAIN=example.com --env OVERWRITEPROTOCOL= --env OVERWRITECLIURL= --volume nc_html:/var/www/html:Z --volume nc_apps:/var/www/html/custom_apps:Z --volume nc_config:/var/www/html/config:Z --volume nc_data:/var/www/html/data:Z --detach=True docker.io/library/nextcloud:25-apache]

I'm able to stop and start the pod again with the following commands:

]$ podman pod stop nc_pod
3e0b4e78d22616e8d5617c9a27de0252f4d39658cf4a8761d7efc9629f435cde

]$ podman pod ps
POD ID        NAME        STATUS      CREATED         INFRA ID      # OF CONTAINERS
3e0b4e78d226  nc_pod      Exited      15 minutes ago  c141f6351cec  3

]$ podman ps
CONTAINER ID  IMAGE       COMMAND     CREATED     STATUS      PORTS       NAMES

]$ podman pod start nc_pod
3e0b4e78d22616e8d5617c9a27de0252f4d39658cf4a8761d7efc9629f435cde

]$ podman pod ps
POD ID        NAME        STATUS      CREATED         INFRA ID      # OF CONTAINERS
3e0b4e78d226  nc_pod      Running     16 minutes ago  c141f6351cec  3

]$ podman ps
CONTAINER ID  IMAGE                                    COMMAND               CREATED         STATUS         PORTS                    NAMES
c141f6351cec  localhost/podman-pause:5.1.1-1717459200                        16 minutes ago  Up 11 seconds  127.0.0.1:40671->80/tcp  3e0b4e78d226-infra
d2284c00dec3  docker.io/library/mariadb:10.11.2        mariadbd              16 minutes ago  Up 11 seconds  127.0.0.1:40671->80/tcp  nc_mariadb
7d8b7285fda7  docker.io/library/nextcloud:25-apache    apache2-foregroun...  16 minutes ago  Up 11 seconds  127.0.0.1:40671->80/tcp  nextcloud

So far, so good. Now I would like to do the same using a playbook.

Playbook 1 - state: restarted

---
- name: Stop and start pod
  hosts: localhost
  connection: local
  tasks:
    - name: Stop the pod
      containers.podman.podman_pod:
        name: nc_pod
        state stopped
      tags:
        - stop
        - never

    - name: Start the pod
      containers.podman.podman_pod:
        name: nc_pod
        state restarted
      tags:
        - start
        - never

This works. Although I would have expected that a pod or container needs to be running to be able to be restarted. I'm wondering that a restart is possible from status Exited.

]$ ansible-playbook playbook_one.yml --tags stop,start
[WARNING]: provided hosts list is empty, only localhost is available. Note that the implicit localhost
does not match 'all'

PLAY Stop and start pod] *******************************

TASK [Gathering Facts] *********************************************************************************
ok: [localhost]

TASK [Stop the pod] **********************
changed: [localhost]

TASK [Start the pod] **********************************
changed: [localhost]

PLAY RECAP *********************************************************************************************
localhost                  : ok=3    changed=2    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   

]$ podman pod ps
POD ID        NAME        STATUS      CREATED         INFRA ID      # OF CONTAINERS
3e0b4e78d226  nc_pod      Running     25 minutes ago  c141f6351cec  3

Playbook 2 - state: started

Now I use the following playbook which I did expect to work as well but which failed:

---
- name: Stop and start pod
  hosts: localhost
  connection: local
  tasks:
    - name: Stop the pod
      containers.podman.podman_pod:
        name: nc_pod
        state stopped
      tags:
        - stop
        - never

    - name: Start the pod
      containers.podman.podman_pod:
        name: nc_pod
        state started
      tags:
        - start
        - never

And here is what happens:

]$ ansible-playbook playbook_one.yml --tags stop,start
[WARNING]: provided hosts list is empty, only localhost is available. Note that the implicit localhost
does not match 'all'

PLAY [Backup or Restore Nextcloud with MariaDB in a rootless Podman Pod] *******************************

TASK [Gathering Facts] *********************************************************************************
ok: [localhost]

TASK [tronde.nextcloud.backup_restore_nextcloud : Stop nc_pod for offline backup] **********************
changed: [localhost]

TASK [tronde.nextcloud.backup_restore_nextcloud : Start nc_pod again] **********************************
changed: [localhost]

PLAY RECAP *********************************************************************************************
localhost                  : ok=3    changed=2    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0

]$ podman pod ps
POD ID        NAME        STATUS      CREATED         INFRA ID      # OF CONTAINERS
8836d60999b1  nc_pod      Running     18 seconds ago  44507b446d8c  1

]$ podman ps -a
CONTAINER ID  IMAGE                                    COMMAND     CREATED         STATUS         PORTS       NAMES
44507b446d8c  localhost/podman-pause:5.1.1-1717459200              37 seconds ago  Up 38 seconds              8836d60999b1-infra

Questions

• Why are two of my containers gone?
• Can someone explain the different behavior of restarted and started to me?

Cheers,
Joerg